HTA Core Model® Online | View Application

HTA Core Model Application for Screening Technologies 1.0

Iris Pasternack, Sunya-Lee Antoine, Katrine Bjørnebek Frønsdal, Kristian Lampe, Alessandra Lo Scalzo, Mirella Marlow, Suvi Mäklin, Ulla Saalasti-Koskinen, Petra Schnell-Inderst, Ingrid Wilbacher

A. About the HTA Core Model®

Any health technology assessment (HTA) contains a vast amount of information. The content, focus, quality and reporting of HTAs vary a lot; this makes finding and transferring the information into local context difficult. The HTA Core Model tackles particularly this problem. The Model defines the content elements to be considered in an HTA and enables standardized reporting. The aim is to improve the applicability of an HTA in other national HTA projects, and enable actual collaboration between HTA agencies by providing a common framework for HTA production.

The HTA Core Model divides HTA information into standardized pieces referred to as assessment elements. An assessment element defines a piece of information that is relevant for the HTA. .The elements that are most likely to be useful for international sharing of information are defined as core elements. Each assessment element contains a question that one should consider including and answering for a specific technology. The Model provides methodological guidance to assist the answering of these questions, and a reporting tool (Fig 1). There is also a storage function for the question-answer pairs referred to as pool of structured HTA information.

The HTA Core Model, and the electronic tool supporting it the HTA Core Model Online, is used to produce the structured Core HTA information. A Core HTA is a project which provides the answers for all relevant core elements for a specific technology , considers the findings per domain in "domain discussions", and summarizes the most important findings. The model serves also those who wish to pick a free selection of elements to be answered. E.g. one could consider sharing certain pieces of information from a national HTA project by sharing them in the pool of structured HTA information with other European HTA agencies.

The HTA Core Model builds on earlier work of projects EUR-ASSESS1, HTA Europe (ref) and ECHTA/ECAHI as well as on other theoretical guidance (refs). It is loyal to the definitions of HTA that emphasize the multidisciplinary nature of assessments. It employs the nine domains that were originally identified in the EUR-ASSESS project (Table 1). Two first applications of the HTA Core Model, one for medical and surgical interventions (EUnetHTA 2008d) and the other for diagnostic technologies (EUnetHTA 2008c), were created during the EUnetHTA Project 2006-08.

11818.Intro1

Figure 1

Table 1. Domains of an HTA

  1. Health problem and current use of the technology
  2. Description and technical characteristics of technology
  3. Safety
  4. Clinical effectiveness
  5. Costs and economic evaluation
  6. Ethical aspects
  7. Organisational aspects
  8. Social aspects
  9. Legal aspects

Assessment element structure – The ontology

In philosophy, ontology has traditionally been a theory of being or existence, i.e. a description of what types of things exist. In recent times, the term has been increasingly used in contexts where the aim has been to assign meanings to information and to describe the relations between concepts. Ontologies typically make it easier for both humans and computers to understand information and its context. Within HTA increased standardisation of the way of searching, handling, and presenting of information may lead to better use of information. The use of other HTAs essentially requires extraction of data from foreign reports and appraisal of its usability in local settings. When data extraction is made easier through well-defined structure and when meanings of each piece of information are clear, the application of foreign data is likely to be less complicated than before.

The HTA Core Model structures the information of an HTA first by dividing it into nine Domains (Table 1). Each Domain is divided into three or more Topics, and further, each Topic is divided into several Issues. The Issues are the generic questions that should be considered when doing a Core HTA. The combination of Domain, Topic and Issue defines an Assessment element (Fig 2).

Assessment elements are the standardized pieces of a Core HTA. Each assessment element is connected to an "element card", which provides tangible information on the element and its relations to other elements. A card may provide advice on how to answer the question defined by the element. Two characteristics within a card (importance and transferability) define whether an element is a "core element" or "non-core element". The answers to questions defined by the element cards are recorded as structured pieces of information in respective "result cards". These are associated with relevant metadata to enable their effective use in the database of HTA information that is being built within EUnetHTA Joint Action WP4.

11818.Intro2

Figure 2. An assessment element

Being in Core

The inclusion of an element in the core is a function of two basic characteristics of the element: its importance and transferability. If the information is fully or partly transferable, it may provide valuable input beyond its original production location. Transferability is low for information that is very specific to a particular context (e.g. region, country, health care system) and is most likely not useful as such in other settings. On the other hand even non-transferable information may be useful; e.g. Italian incidence data on cardiovascular mortality is applicable to all Italian HTAs assessing cardiovascular technologies or,  Swedish data on current use of the technology may suggestover-or underuse of the technology in ones own country.

Importance is included in the consideration to ensure that the core is robust enough, i.e. that it contains information that is really significant from the viewpoint of HTA. The importance considered here is not equal to relevance of information for a particular policy question. It is assumed, however, that issues perceived important from the viewpoint of HTA are often useful when making decisions on health care policy.

The inclusion in the core is defined according to the following core matrix:

CORE MATRIX

Importance

1 Optional

2 Important

3 Critical

Transferability

3 Complete

Not core

Core

Core

2 Partially

Not core

Core

Core

1 Not

Not core

Not core

Core

It should be emphasized that the inclusion or exclusion of an element into or from the Core is driven by usability of the information across national borders of other contexts. Not belonging to the core does not mean that an element would be unimportant, insignificant or not worth considering in an HTA. On the contrary, important assessment elements (that are not transferable) are excluded from the Core by definition (see Core matrix above). Such elements are likely to provide useful or even critical information to guide decision-making and need to be addressed locally by individual HTA agencies.

In the current version of this document the importance and transferability of each element - and hence their status regarding the Core - has not always been considered enough. Therefore any judgements should be regarded as tentative. Further piloting will provide more accurate values.

Generic Model and its applications

Different types of technology - such as drugs, devices or procedures - may require different kinds of assessment. Therefore it has been decided that within one HTA Core Model there are different applications for the assessment of different types of technologies. There are two earlier HTA Core Model applications created during the EUnetHTA Project 2006-2008: one for medical and surgical interventions and the other for diagnostic technologies. During the EUnetHTA Joint Action 2010-12 there will be two further Model applications designed: one for screening technologies (which is described in this document) and the other for the relative effectiveness assessment of pharmaceuticals (WP5). Additionally there will be a subset of the pharmaceutical Model, the so-called "rapid model".

When creating a new HTA Core Model application, or updating an existing one, the aim is to keep most of the Model generic, i.e. identical across various applications. Additional to the generic main part, the applications contain assessment elements and methodological guidance which are specific for thy type of technologies the application covers. When updating the HTA Core Model, all changes in the generic part of the Model will be transformed automatically to all applications. Application-specific amendments need to be updated separately.

HTA Core Model Online Tool and Handbook

A test version of the HTA Core Model Online, at http://www.corehta.info, was opened to restricted use for EUnetHTA JA partners in June 2010. An improved version will be available in March 2011. The tool contains a Handbook which guides the users of the tool in five phases. The project and its participants are defined in phase 1 and the assessment protocol designed in phases 2 and 3. After finding answers to the study questions the process continues in submitting the results of the research in the online database (phase 4) and publishing the results (phase 5). An editorial process precedes the final publication. The HTA Core Model is subject to Terms of Use, available through www.eunethta.net.

Concepts

from HTA Core Model Handbook, available at http://www.corehta.info

HTA Core Model: A structured manner of creating and presenting HTA information as assessment elements. Some elements are prioritized over others to support European collaboration through defining them as "core elements".

Assessment element: The basic unit of the model. Defines a piece of information that describes the technology or the consequences of implications of its use, or any other implication that is relevant for the assessment, such as the patients and the disease for which it is applied. Each assessment element contains an "issue", which is a question that should be answered in an HTA. Not all issues, however, are relevant to all technologies/settings, and hence their relevance is considered separately for each assessment. Elements are defined through a combination of domain, topic and issue.

Domain: A wide framework within which technology is considered. It provides an angle of viewing the use, consequences and implications of technology. A standard set of domains is used in the HTA Core Model.

Topic: A more specific area of consideration within the domains. One domain is divided into several topics. Similar topics may be addressed within more than one domain.

Issue: An even more specific area of consideration within any of the topics. One topic typically consists of several issues, but it may also contain only one issue. An issue is always expressed as a question that can be answered through answering one or more research questions.

Application of the HTA Core Model: Different kinds of technologies (e.g. surgical interventions or pharmaceuticals) may require different questions to be asked in an assessment and the answers to the questions may require different research methods. An application of the HTA Core Model is built for assessing a specific kind of health technology. Different applications all draw from the same pool of assessment elements, but not all elements are used in all applications. Currently there are two existing applications, one for medical and surgical interventions and another for diagnostic technologies. This document presents the draft version of the third application: HTA Core Model for screening technologies.

Element card: Each assessment element is connected to an element card, which provides tangible information on the element and its relations to other elements. A card may provide advice on how to answer the question defined by the element. Two characteristics within a card (importance and transferability) define whether an element is a "core element" or "non-core element". While assessment elements are generic (i.e. one element can belong to several applications of the HTA Core Model), element cards are application-specific (i.e. the cards describing an element within different applications may be different).

Structured HTA information: Information on any aspect of health technology that has been created through answering the issues defined in the assessment elements of the HTA Core Model.

Core HTA information: Any information on a technology that has been produced through answering the issue defined in a core element, or a collection of such information. This information is very likely to be useful in the European context (i.e. also in another country) due to its importance and/or transferability.

Core HTA: An actual assessment that a) has been conducted using the HTA Core Model and b) has considered all core elements of all 9 domains. (Note: through this consideration some elements may be defined as irrelevant, but that should be documented).  A Core HTA contains a chapter that draws together key findings of various domains, but does not make recommendations regarding the use of technology. Through the wide scope, focus on core elements and the summary chapter, a Core HTA gives an overview of a technology that is likely to be useful in the European context. A Core HTA can be used as a basis for producing local HTA reports that take into account local circumstances (e.g. epidemiology, organisation, resources, values).

B. HTA Core Model on Screening Technologies

Work process

The HTA Core Model was built by several working groups called Domain teams (see page 3). Each team focused on one domain. The roles were divided into investigators and reviewers. The work of investigators within each domain team was coordinated by a primary investigator. The investigators used the existing two HTA Core Model applications (EUnetHTA 2008 and b) as base text and had as task to update the generic Model and to consider adding specific elements and methodological guidance relevant to screening technologies. Reviewers commented on the draft versions of the investigators' work. The primary investigators from each domain formed the Coordination and editing team (CET). CET had the task to consider and prepare documents with common interest across domains: e.g. the communication protocol (Box 1) and definition of screening (see below). There was a primary coordinator for the whole project who also led the CET.

The task was divided into three sections:

–        Updating the Domain description,

–        updating the Assessment elements table, and

–        updating the Domain methodologies.

1) Updating the Domain description

The investigators’ task was to modify the base text so that it remains generic, i.e. is applicable to all types of technologies; medical & surgical interventions, diagnostic, and screening technologies. If there was a need to amend information that is specific for screening technology only, it should be marked in italics or placed under separate subheading.

The domain descriptions in the earlier Model applications were heterogeneous across domains; they were of different length and different in content. In some domains, there were actually two quite different domain descriptions from the earlier model applications, and the current investigators had to start with combining these two into one generic domain description. Another task of the investigators was to harmonize the texts within and across the domains, and shorten and clarify the text for  the online tool. Four subheadings were proposed to structure the content:

  1. What is this domain about? (including concepts)
  2. Why is it important?
  3. Relations to other domains
  4. Specific features in finding, interpreting or implementing information for this domain
  5. Issues specific for screening technologies

The fifth subheading, Issues specific for screening technologies, was optional. It should be used only when there is a need to add domain specific information regarding screening technologies. All the general screening specific issues (e.g. the fact that screening is a process and it deals with large numbers of healthy people) will be explained in the "Definition for screening technology" chapter in the Introduction section (see below).

2) Updating the Assessment elements table

The investigators were encouraged to go through the topics and issues in the table, one by one, asking the following questions:

  1. Does this issue apply/ Should we keep this issue in the Core model on screening?

–        If no, suggest removing it from the screening model

–        If yes, ask yourself if  the issue needs some modification

  • If yes, please modify. Notice that the issues should be generic questions. You should formulate them to be screening specific.
  1. Are there issues missing that are relevant to screening?

–        If yes, add new element (row)

  1. Is the topic-issue hierarchy logical and useful

–        If not, suggest changes

  1. Suggest amendments or changes to the Clarification, Information sources and Reference fields
  2. Consider the Importance and Transferability of each element. Suggest changes to the existing numbers.
  3. Suggest amendments or changes to the Relations fields. Identify elements with relevance for which the order of answering may be important in doing a core HTA (Do certain elements need to be answered before one can continue working with this element?) or with possible overlapping (Are there elements which require the same original studies, is there a possible duplication of work?).

3) Updating the Domain Methodologies section

The task and the problems were here the same as in domain description. The investigators should combine and harmonize the original texts that were lengthy and heterogeneous. The task of keeping generic items separate from screening technology specific methodology items was a challenge. The authors were also encouraged to identify those parts of the text that were not directly specific for their particular domain. The methodological guidance that is applicable in several, or even in all domains, was moved to the "Common methodologies" section in Appendix 3. 

It was made explicit that the style should not be a text book, neither a methodological article.  Instead of lengthy descriptions, the investigators were encouraged to write brief sentences and use lists and links to useful sources and tools.

In order to enhance common structure for Methodology sections, a set of subheadings was proposed. Not all domains could use this set of headings and more freedom was given to the investigators in formulating their text. Suggested subheadings were:

Where to find information?

–                    Databases and search strategies

–                    Useful other sources and links

What kind of information is required?

–                    Study types: including design, outcome measures

–                    Quality assessment

How to collect information?

–                    Systematically vs other

–                    Data extraction

–                    Own research/evidence generation

Analysing and synthesizing evidence

–                    Biases, confounding factors, level of evidence

–                    Evidence tables

–                    Meta-analysis

–                    Qualitative synthesis

Reporting and interpreting

Communication

 “Restricted authorship”

We agreed that the authorship in this project is not as it is normally when preparing a scientific article in a team of authors. This is because there are two earlier versions of the Core Model, and the task of the investigators was to do only the modifications that are necessary to adjust the Model for screening technologies. It is not desired to rewrite the text anew. The investigators need to accept that they deal with others’ text and give their contribution on top of others' work. It is process writing, similar to Wikipedia. Careful consideration and separate instructions will be needed if someone wants to publish an article about the work done in the domain team during this project.

External communication and feedback

Communicating about the project is in general encouraged. Anyhow, it was agreed that all project participants keep the project coordinator informed about any occasion where the aims or results of the project are presented; be it interview, poster, speech or article. In order to gather success stories (or failures), and all kind of feedback of the HTA Core Model and its screening application, all participants were encouraged to inform the coordinator of any comments and feedback they have encountered. EUnetHTA Joint Action secretariat were be kept informed about the external communication an feedback issues.

Definition for screening technology

Depending on background and training, people give different meaning to the word "screening". The following observations and definitions were agreed for this project.

Why do we need a dedicated Model application for screening technologies?

Screening involves testing to identify people at high risk of having a specific disease (diagnosis). As there is already a HTA Core Model application for diagnostic technologies that covers testing procedures, why do we need additional application for screening? The following properties of screening were identified that justify the need of a dedicated application of the HTA Core Model.

–        As preventive or early diagnostic intervention, screening is targeted to a large number of healthy or asymptomatic people – in contrast to diagnostics where people typically already have some symptoms or signs of illness.

–        Screening tests are usually applied in a population with low disease prevalence; mostly healthy people. Therefore, the diagnostic tools often perform very differently from clinical settings (i.e. very low positive predictive value). The same technology has different performance when used in diagnosis than in screening.

–        Effectiveness depends on participation rate of the target population.

–        Screening usually requires careful ethical and legal considerations, due to the risk of false positives and false negatives, the consequences related to theunder-or over-diagnosis and -treatment, and earlier diagnosis in cases where  prognosis improvement is neglible. Equity of access is always an issue in screening programs.

–        There are several organizational issues specific for screening as it

  • involves active contact of the target population by the health service
  • is multidisciplinary and involves multiple providers
  • requires quality control and a continuous monitoring system.

–        There are many specific characteristics and methodological issues which have to be taken into account when evaluating economic impact of a screening program. For example, most of the costs of a screening program are incurred within a relatively short time period and the benefits (e.g. life years gained) further in the future. This means that decisions about whether to discount the future costs and effects or not, and which discount rate(s) to use, need to be carefully considered.

Multitude of definitions for screening

There are two main streams of considering screening as a public health intervention.

–        The first, mostly adopted in Europe, considers screening as a program in which

  • the target population and adequate screening interval are determined in advance;
  • all individuals in a certain category (e.g. all women of a certain age) are involved;
  • the health services contact systematically and actively the target population; and
  • a standard process is determined for further diagnostic examinations subsequent to the screening test, as well as for treatming those with the diagnosed condition.
  • This approach is also called universal screening, mass screening, population screening, or community screening.

–        The second stream, mostly adopted in the USA, considers screening to be spontaneous, or so called opportunistic screening, in which the practitioners recommend the test to their (asymptomatic) patients more or less systematically and according to their attitudes and knowledge. This kind of screening lacks systematic identification and contacting of the target population. Instead it is dependent on the activity of the individuals themselves, their health service providers, and funding arrangements (health insurance package).The process for further examinations and treatment is not standardized.

There are additional uses of the word screening in medicine

–        "Screening" may be performed during a regular patient visit, on an asymptomatic patient, to exclude or confirm diagnosis (e.g. bone density measurement).

–        Surveillance screening involves testing of a sample of the population to survey the prevalence of a disease or an exposure, without the aim of improving prognosis in diseased individuals.

–        Toxicological screening involves testing of environmental or clinical samples to identify toxic substances.

–        Molecular screening is a phase in the selection of active molecules in pharmacology.

More related concepts

–        Case finding: Involves a smaller group of people based on the presence of risk factors (e.g. when a family member has been diagnosed with a hereditary or communicable disease). "Case finding" is also used in the context of screening a single patient who consults the doctor on a problem not directly related to the disease being screened . An example of this is cervical cancer screening during a consultation for other gynecological problem.

–        Routine safety checks (e.g. related to anaesthesia)

–        Baseline value assessment (e.g. liver enzymes before medication)

–        Check-up, periodic health examinations often involve a number of screening elements

What is meant by 'screening technology' in the context of this Core Model application?

The producers of a core HTA should be aware of the multitude of uses of the word 'screening', and the fact the 'HTA Core Model on screening technologies' is not applicable to assessing everything that is called screening. The primary target is the full population screening program with the following components:

–        It involves a test or an examination or a series of tests or examinations, AND

–        is provided either systematically to the whole target population (i.e. in a screening program) , or unsystematically for asymptomatic people,  e.g. in the form of locally provided health promotion or case finding programs, AND

–        is done in order to make a statement regarding the possibility of having a certain disease or risk factor, AND

–        aims at improved prognosis, or an improvement of the management or coping with the disease (excludes technologies which aim at surveying the prevalence or spread of a certain disease, risk factor, or exposure only).

Sometimes it is necessary to assess only a certain part of the program; e.g. the effects of replacing the conventional mammography device with a digital one in a breast cancer screening program. In this case a relevant subset of the HTA Core Model of screening technologies is applicable.

The HTA Core Model on Screening is not suitable for use when the aim of the HTA is assessing

–        the accuracy of a single test to determine exposure/risk factor or disease or

–        effectiveness of opportunistic screening practices.

References

Commission on Chronic Illness. Chronic illness in the United States. Vol I. Prevention of chronic illness. Cambridge, MA: Harvard University Press, 1957;1:45.

EUnetHTA. 2008a. Work Package 4. Core HTA on drug eluting stents (DES), [updated December 31, 2008]. http://www.eunethta.net/Public/EUnetHTA_Deliverables_project_2006-2008/.

EUnetHTA. 2008b. Work Package 4. Core HTA on multislice computed tomography (MSCT) coronary angiography,[updated December 31, 2008]. Available at: http://www.eunethta.net/Public/EUnetHTA_Deliverables_project_2006-2008/.

EUnetHTA. 2008c. Work Package 4. HTA Core Model for diagnostic technologies v 1.0r. Available at: http://www.eunethta.net/Public/EUnetHTA_Deliverables_project_2006-2008/.

 EUnetHTA. 2008d. Work Package 4. HTA Core Model for medical and surgical interventions v 1.0r. Available at: http://www.eunethta.net/Public/EUnetHTA_Deliverables_project_2006-2008/

UK National Screening Committee. Criteria for appraising the viability, effectiveness and appropriateness of a screening Committee. Criteria for appraising the viability, effectiveness and appropriateness of a screening programme. 2009.  http://www.screening.nhs.uk/criteria  Last accessed April 5 2011.

Wilson JMG, Jungner G. Principles and practice of screening for disease. Public Health Paper Number 34. Geneva: WHO, 1968.

Cochrane AL. Holland WW. Validation of screening procedures. Br Med Bull. 1971, 27, 3.

Sackett DL, Holland WW. Controversy in the detection of disease. Lancet 1975;2:357-9.

Wald NJ (Editor). Antenatal and Neonatal screening. Oxford University Press, 1984.

Holland WW, Stewart S. Screening in Healthcare. The Nuffield Provincial Hospitals Trust, 1990.

Gray JAM. Dimensions and definitions of screening. Milton Keynes: NHS Executive Anglia and Oxford, Research and Development

Morrison AS. Screening for chronic disease. Oxford University Press, Oxford, New York, 1985

Fletcher RH, Fletcher SW, Wagner EH. Clinical epidemiology, the essentials. 2.ª ed. Baltimore: Williams and Wilkins, 1988.

Sackett DL, Strauss S, Haynes RB, Richardson WS. Evidence-based Medicine. How to practice and teach EBM, 2nd ed. Edinburgh: Churchill Livingstone, 2000)

Health Problem and Current Use of the Technology

Sunya-Lee Antoine, Paolo Giorgi Rossi, Chris Lawinski, Leonor Varela Lema

Description

What is this domain about?

This domain describes the target conditions, target groups and the availability and patterns of use of the technology in question. Some of the topics considered relevant for this domain have generally been called “Background Information” in previous European projects or recommendations for conducting assessments(Burls 2000, Busse 2002, Liberati 1997).

The qualitative description of the target condition, including the underlying mechanism (pathophysiology), natural history (i.e. course of disease), diagnosis and prognosis, the risk population and risk factors for acquiring the condition as well as available treatments are described in this domain. A description of subgroups or special indications should be included especially when the technology does not target the whole population.

Current management patterns are described, and whether the technology is intended to replace, add or triage another technology in the management chain. Anticipated problems in the use, e.g. inappropriate extension of indications, participation rate, over-diagnosis, misuse, and acceptability by the population, are to be discussed, as well as the alternatives to the technology and agreed policies on whom to treat as patients or target group.

Why is this domain important?

The information produced in this domain provides baseline knowledge which is needed when the results from other domains of the assessment are put into context in a particular geographical or organisational setting. If health problem and the target population can not be clearly defined, the appropriate use of the technology may be rightfully challenged. If the current management practice is not in accordance with evidence-based guidelines, the public might get the impression that a need for a new technology exists. A new technology could be costly and not necessarily more effective than existing ones. In that case it could be more appropriate to improve the compliance to guidelines than to add a new technology with a similar effectiveness and/or higher costs.

National decision-makers are interested in the extent of utilization of technology in their own country, and if there is regional variation. On the other hand, international benchmarking may have a great impact on decision-making process(Zentner 2004). Particularly important it may be when the estimation of the harm-benefit-costs equation is inconclusive. It might be important to be aware of the variation in the management patterns and current use of the technology in Europe, because this may reveal under- or overuse of the technology in one's own country.  In Europe, great variation in approval status of  technologies is seldom expected; therefore it may be of interest to compare the status with non-European countries.

Relations to other domains

The issues in this domain should be considered at an early stage of a Core HTA, because they may help in refining the research questions and formulating the methodological approach in e.g. effectiveness, costs and organisational aspects domains. The life cycle of the technology, its regulatory (approval and coverage) status and manufacturer information are of joint interest with other domains (description and technical characteristics, organisational, social, ethical, and legal aspects domains).

Some issues in this domain will necessarily overlap with issues in the effectiveness and costs domains (e.g. issues of consequences and alternative interventions), organizational domain (e.g. utilisation issues), description and the technical characteristics domain (e.g. life-cycle), social domain (coverage and access issues), legal and ethical domains as well as safety domain (e.g. over-diagnosis, false positive and false negative test results). It is important to coordinate the work with these issues, and determine who answers them within a particular Core HTA.

Issues specific for screening technologies

Usually a technology is proposed for screening after a long experience in clinical diagnostic use. This means that assessing a screening technology is usually assessing the features of the technology in a new application context. Screening as context means that the assessment should include the whole management chain, from the screening test, through the subsequent diagnostic tests to treatments. It is therefore important to distinguish if the proposed assessment topic includes a new screening technology, that only slightly modifies the existing screening pathway, or if it is an assessment of a completely new screening pathway. Regulatory processes hardly ever distinguish between these two uses of a technology: clinical or screening setting.

Knowledge on the following aspects is essential for the construction of decision analytic models for screening technologies:

  1. Natural course of the health problem,
  2. Diagnosis of the health problem, 
  3. Effect of available treatments on the course and prognosis,
  4. Burden of disease, incidence, mortality, survival,
  5. Current guidelines and existing screening flow charts
  6. Effects of the screening technology on the epidemiology (incidence, prevalence, overdiagnosis) of the health problem

Methodology

Where to find information?

Databases and search strategies

–        The EUnetHTA pool of structured HTA information will be a pertinent source of information on e.g. disease incidence.

–        HTAs, systematic reviews and original research can be found in reference databases: e.g. CLIB, CRD DARE,Medline, Embase, Cinahl, PsychInfo.

–        Evidence based guidelines can be found in reference databases, guidelines producers’ web sites and in Guidelines international network's (GIN) web site.

–        Textbooks are valuable source of descriptive information, for example for information on disease mechanism.

Useful other sources

–        Registers and statistics

  • Technology and procedure registers ( [100] in Appendix 1)
  • Disease registers ([105] in Appendix 1)
  • Routinely collected statistics and administrative data (e.g. DRG, discharge databases, reimbursement claims databases)

–        Horizon scanning databases and web sites

–        Ongoing research databases

–        Scientific specialist associations' web sites

–        Patient associations' web sites

–        Market approval and other regulatory institutions' web sites ([109] in Appendix 1)

–        National health services' web sites

–        Regional/local governments' health departments' web sites

–        Benefits and sickness funds' web sites

–        Technology developers and manufacturers web sites

–        Various sources through using internet search engines

–        There are some issues, e.g. the coverage status of a technology (inclusion in the benefit catalogue, levels of co-payment, etc.), where information is not easy to retrieve. It requires local knowledge of the health-care system  to identify adequate and usable information sources(Velasco-Garrido 2006).

Own research and evidence generation

–        Own qualitative research might be the only way to assess real practice use and misuse. However, these studies are not frequently undertaken since they are resource consuming.

–        Discussions with experts or officials

–        Expert surveys or interviews

–        Own register based research

What kind of information is required?

Study types, design, outcome measures

There is no single methodological approach which can be applied to all issues in this domain (See Table 1). The epidemiology of the target health condition and its consequences are usually described in terms of prevalence and incidence (e.g. mortality, disability, sickness leave, retirement).

Specific for screening technologies

It is difficult to obtain information on misuse or overuse of a screening technology, or the spontaneous diffusion o using a test in the healthy population before the implementation of a screening programme. Consequently, this information needs to be collected from indirect sources. A case report that describes routine use of a screening test in all cases admitted for a certain disease or health problem in a certain hospital gives reliable information on the use of the screening technology, although the clinical results of this study would not be reliable.

Table 1. Types of information required in this domain

Research question

Study type

Quality assessment

Systematic data retrieval needed?

Synthesis

Disease mechanisms

Descriptive

No established way to assess the quality of narrative reviews and text books.

No. Updating existing information is sufficient.

Narrative

Natural course of condition

Observational

STROBE check list

No. Updating existing context relevant information is sufficient.

Narrative

Prevalence and incidence of the condition

Observational

STROBE check list

No. Updating existing context relevant information is sufficient.

Data may be meta-analysed, but often there is no opportunity to do that.

Risk factors and consequences

Observational

Newcastle-Ottawa scale

Yes

Meta-analysis per subgroups if possible.

Prognosis

Prognostic

Newcastle-Ottawa scale

Yes

Data may be meta-analysed

Technology utilisation

Narrative reviews, surveys, observational and qualitative research, register analysis

Relevant at least for quantitative studies.

Not necessarily, in particular in Google or other non-scientific sources.

Narrative

Current practise in the management of the condition, practise variation

Guidelines, consensus statements,  observational and qualitative research

Not needed

Not necessarily, information from internet or or other non-scientific sources may be useful.

Narrative

Tools for critical appraisal

The validity of the information may differ considerably depending on the source and type of information requested (see Table 1). For example, it might be difficult to find up-to-date information on the approval status of a technology by doing a review of published literature. Even if there are scientific publications on the issue (i.e. policy studies) they are likely to be rapidly outdated. The information obtained by directly inquiring (e.g. via telephone query) the relevant approval agencies will be more reliable and practical. Quality assessment of the information retrieved may be difficult, as there is often no standard way of doing it.

Quality assessment of epidemiologic studies

Newcastle Ottawa scale (see Appendix 3) may not be appropriate in the quality assessment of studies examining disease prevalence or burden of disease. It is more appropriate for studies assessing the link between diseases and risk factors. STROBE check list can be used as a check list for study quality, although it is an instrument meant for assessing the quality of reporting (see Appendix 3).

Quality assessment of registers and statistics

Several national and international sources of statistics exist which can be used to assess the incidence, prevalence, mortality, or burden of disease. These statistics are usually available in aggregated form and increasingly through the internet. The use of these sources has several limitations. The reliability of the diagnosis varies and usually it is not possible to differentiate between different stages of the disease. Even the validity of the coding of causes of death may be variable, and in some countries it is known to be very limited. However, when there is a quality assured register, as in the case of many organized screening programs, the information can be highly reliable.

The relevance and quality of registers should be appraised carefully considering the following questions:

–        How representative is the register? (European, national, regional, local?)

–        What kind of information is coded?

–        What are the inclusion/exclusion criteria for data entered?

–        What is the quality of information?

–        How complete is the coverage?

Data access is an important aspect when working with registers. It may be impossible for institutions other than the ones managing the register to analyze the raw data. However some registers conduct customized analyses.

Quality assessment of routinely collected statistics and administrative data

Routinely collected administrative data (e.g. DRGs, discharge databases, reimbursement claims databases) can be useful too, when available. For example sickness funds collect great amounts of information which could be used to analyse utilisation of technology. By definition, these data have been collected for other purposes than research and they can not be used to answer scientific questions without previous processing. Analysis of this kind of data might be very time consuming, since data need to be “prepared” before analysis. This might not be feasible in the context of an HTA project, due to resource constraints.

The use of routinely collected statistics has several limitations. The reliability of the diagnosis varies and usually it is not possible to differentiate between different stages of the disease. Even the validity of the coding of causes of death may be variable, and in some countries it is known to be very limited. Own analysis of administrative data often requires authorization from the data owner, which in some countries might be difficult to obtain due to issues of privacy protection and confidentiality.

Quality assessment of manufacturer data

The information provided by manufacturers might be limited by issues of confidentiality and marketing. This source can be useful in order to answer questions concerning the requirements for use of the technology, development status or forthcoming innovations of the technology. Manufacturers may also provide information on ongoing research and on scientific literature which has not been published yet. Potential bias in the information provided by manufacturers needs to be carefully assessed.

Analysing and synthesizing evidence

There are several issues, particularly in this domain, where systematic data retrieval is not necessary (see Table 1). Unsystematic gathering of information from books, introduction sections of reviews and articles, registers and internet until saturation is reached, may be enough.

Reporting and interpreting

Transparency in information retrieval is crucial when reporting a Core HTA; the sources and methods of retrieval, systematic or not, and quality assessment criteria (also when missing) should be explicitly stated for each issue.

A reader of a Core HTA might be interested to know the incidence of the condition and the extent of use of the technology in other countries, particularly when there is no information available from own country. Therefore, both aggregate data and national data can be of importance, and can be reported. Tables, graphs and figures make abundant numerical information, e.g. trends in epidemiology, more digestible.

Overview of guidelines synthesizing the main recommendations on management practises would be illustrative. Flowchart of the current management pathway is particularly illustrative in diagnostic technologies. It helps the reader to understand what is the intended role of the new technology in the current management chain (add-on, replacement or triage).

Assessment elements

A0001 Assessment element card

Issue: Which disease/health problem/potential health problem will the technology be used for?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Definition (naming) of the condition, health problem, disease for which the technology is intended.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Medical literature, narrative reviews, book chapters

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Sequential relations

A0002 Assessment element card

Issue: What, if any, is the precise definition/ characterization of the target disease? Which diagnosis is given to the condition and according to which classification system (e.g. ICD-10)?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Characteristics of the condition which allows a precise diagnostic and differentiation of the indication for the use of the technology.Subgroups or indications are considered under the Domain Clinical Effectiveness

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

WHO

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness Domain

Sequential relations

A0003 Assessment element card

Issue: Which are the known risk factors for acquiring the condition?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The prevalence of different risk factors might be different in different geographic areas and among different groups of population. This element clarifies the identification of alternative (also preventive) management approaches.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Narrative and systematic reviews, book chapters

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness Domain

Sequential relations

A0004 Assessment element card

Issue: What is the natural course of the condition?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Description of underlying mechanisms or pathophysiology. Possible relation between early diagnosis and better prognosis?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Registries

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness and Costs Domains

Sequential relations

A0005 Assessment element card

Issue: What are the symptoms at different stages of the disease?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Symptoms by stage might give an idea of possible improvements, and provide proxy outcomes for effectiveness assessment.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Registries, quality of life studies, narrative and systematic reviews, book chapters

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness Domain

Sequential relations

A0006 Assessment element card

Issue: What is the burden of the condition?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Prevalence or incidence of disease specific mortality, life years lost, disability

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Registries and national statistics

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Social and Costs Domains

Sequential relations

A0009 Assessment element card

Issue: What aspects of the burden of disease are targeted by the technology?

Topic: Target Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The technology can affect only some aspects (e.g. mortality) and leave other aspects (e.g. quality of life) untouched. Screening may increase disease incidence due to early diagnosis and over diagnosis.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Deductive models (based on the natural history of the disease, test target and treatment target; epidemiological studies (if sufficient testing has been done)

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Social and Costs Domains

Sequential relations

A0007 Assessment element card

Issue: What is the target population of the technology?

Topic: Target Population

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The technology may be used for all patients having the condition, or only those in early stages, or certain severity level, or people with moderate risk of having the condition. In screening and other preventive interventions the target population represent a defined subgroup of healthy or asymptomatic individuals. Who have defined the selected subgroup(s) and for which reasons?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Medical literature, narrative reviews, commentaries, editorials of scientific associations, guidelines, recommendations

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness Domain

Sequential relations

A0023 Assessment element card

Issue: How many people belong to the target population?

Topic: Target Population

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

National registries, statistics, systematic reviews

References
Content relations
Sequential relations

A0011 Assessment element card

Issue: How much is the technology being used?

Topic: Utilisation

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

What is the current rate of screening adherence?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

National statistics, surveys, disease management studies, manufacturer sales data

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Costs and Organisational Domains

Sequential relations

A0012 Assessment element card

Issue: What kind of variations in use are there across countries/regions/settings?

Topic: Utilisation

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

National statistics, surveys, disease management studies, manufacturer sales data

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Sequential relations

A0013 Assessment element card

Issue: How is the disease/health condition currently diagnosed or screened?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Properties of diagnostic or screening tests affect patient spectrum and thus the effectiveness of subsequent interventions. Different tests are applied by different professional groups. This information is needed e.g. in cost-effectiveness models.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Surveys, utilisation reviews. If such information is lacking: Expert surveys / expert interviews, web search

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Costs and Organisational Domains

Sequential relations

A0014 Assessment element card

Issue: How should the condition be diagnosed or screened according to published algorithms/guidelines?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Costs and Organisational Domains

Sequential relations

A0015 Assessment element card

Issue: How is the condition currently managed?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Deviation from eb-guidelines may hint over/under use of the technology and it may increase the burden of disease.Identification of practice variations may point out differences in the quality of health care.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Surveys, utilisation reviews. If such information is lacking: Expert surveys / expert interviews, audits

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Costs and Organisational Domains

Sequential relations

A0016 Assessment element card

Issue: How should the condition be managed according to published algorithms/guidelines?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

An assessment of the main differences between guidelines and actual practice allows conclusions to be drawn on how optimal the current management is.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Review of clinical guidelines, recommendation. If such information is lacking: Expert surveys / expert interviews, textbooks

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Costs and Organisational Domains

Sequential relations

A0017 Assessment element card

Issue: What are the differences in the management for different stages of disease?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Surveys, utilisation reviews, clinical guidelines, recommendations. If such information is lacking: expert surveys / expert interviews

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Social Domains

Sequential relations

A0018 Assessment element card

Issue: What are the other evidence-based alternatives to the current technology?

Topic: Current Management of the Condition

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical guidelines, recommendations, systematic reviews

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical Effectiveness, Costs and Organisational Domains

Sequential relations

A0019 Assessment element card

Issue: In which phase is the development of the technology?

Topic: Life-Cycle

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It can be experimental, emerging, or routine use? Usually a new test for primary screening needs studies with very long follow up. Consequently new screening putative tests are usually very old tests. On the other hand a new triage test, therapy for positive individuals may be experimental. Finally we may have a new version of an old primary test, in this case it may be experimental.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Horizon scanning databases, ongoing research databases, information from manufacturers.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety, Clinical Effectiveness, Ethical, Social and Legal Domains

Sequential relations

A0020 Assessment element card

Issue: Which market authorization status has the technology in other countries, or international authorities?

Topic: Regulatory Status

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Imaging devices may require approval. Substances needed for obtaining images may require additional approval (e.g. radiotracers). In some cases the approval for primary screening is different to that for clinical use (FDA recently licensed tests explicitly for screening), but in most cases approval is obtained for diagnostic use and the test is proposed for screening without any other formal approval.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

e.g. CE-Approval, EMEA, national authorities. Manufacturers should be contacted in order to identify which steps have they taken/ are they planning to take concerning market approval

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Legal Domain

Sequential relations

A0021 Assessment element card

Issue: What is the reimbursement status of the technology across countries?

Topic: Regulatory Status

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The technology may be differently included in / excluded from the benefit basket of the countries. Reimbursement status may be different for different purposes: e.g treatment vs prevention, diagnosing vs screening or monitoring. Information of full-coverage, co-payments, coverage under special circumstances/conditional coverage is useful.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Lists of benefits / services of the national health services / sickness funds, inquiry of technical officers from MoH. Manufacturers. Literature on benefit basket (Comparative policy studies)

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Burls 2000, Busse 2002, Liberati 1997 Imaz-Iglesia 1998, Kristensen 2009

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Legal Domains

Sequential relations

References

Burls A, Cummins C, Fry-Smith A, et al. West Midlands Development and Evaluation Service Handbook. West Midlands DES (2000).

Busse R, Orvain J, Velasco M, Perleth M, Drummond M, et al. Best practice in undertaking and reporting health technology assessments. Working group 4 report. International Journal of Technology Assessment in Health Care 2002;18(2):361-422.

Evidence Based Medicine Working Group. User’s guides to the medical literature. Series published in JAMA between 1993 and 2000.

Liberati A, Sheldon TA, Banta HD. EUR-ASSESS Project Subgroup report on Methodology. Methodological guidance for the conduct of health technology assessment. International Journal of Technology Assessment in Health Care 1997;13(2):186-219.

Velasco-Garrido M, Schreyögg J, Stargardt T, Busse R. Identification of health baskets in nine EU countries. Revue francaise des Affaires Sociales 2006; 60(2-3): 63-88.

Zentner A, Busse R. Das Ausland in aller Munde. Gesundheits- und Sozialpolitik 2004; (9-10): 24-3

Description and technical characteristics of technology

Katrine Bjørnebek Frønsdal, Sunya-Lee Antoine, Patricia Harrington, Antonio Migliore

Description

What is this domain about?

The information given in this domain  describes the technology (or a sequence of technologies), when was it developed and introduced, for what purpose(s), who will use the technology, in what manner, for what condition(s), and at what level of health care. The material requirements for premises, equipment and staff are described, as well as any specific training and information requirements. The regulatory status of the technology should be listed, where applicable.

The issues in this domain need to be described in sufficient detail to differentiate the technology from its comparators. Such terms and concepts should be used that allow those unfamiliar with the technology to get an overall understanding of how it functions. It is important to distinguish between scientifically proven versus suspected mechanisms of action. Important terms should be defined, and a glossary or a list of product names provided. The section may include pictures, diagrams, videos, or other visual material, in order to facilitate understanding.

Why is this domain important?

A careful description of the technical characteristics and special requirements of the technology, and the rationale for its use may help with translating policy questions into research questions in other domains. Different generations or versions of a technology may have different indications, performance characteristics and applicability. A good description of the technology is particularly important in a fast developing field where even minor changes or improvements in a technology can have variable effects on the measures of benefit.

Relations to other domains

There is a considerable overlap with the current use, organisational and legal Domains. The authors should co-operate with the authors of those domains to avoid duplication of work.

Methodology

Where to find information?

The source of information will depend on the location of a technology within its product life cycle. Review articles and textbooks can be helpful when searching for information about the history and characteristics of established technology. For prototypes and innovative technologies published peer reviewed literature may be limited. It may need to be supplemented by grey literature (includes non-peer reviewed and non-published literature, as well as confidential commercial information) as well as anecdotal information from general web-searches. The use of a systematic search is usually not necessary when gathering information on the descriptive and technical characteristics of a technology.

Databases and search strategies

Published literature may be obtained by searching bibliographic databases such as Pubmed, EMBASE, EBSCO, the Cochrane Library and the Centre for Reviews and Dissemination (CRD). Establishing regular notifications for new results using the alert function on these databases will facilitate easy updating of the literature review to ensure that it is current at the time of completion of the HTA. Electronic searches can be supplemented by hand-searching the reference lists of key papers.

Useful other sources and links

Grey literature (e.g.working papers from research groups or committees, white papers, or preprints), hand-searching of reference lists, as well as conference proceedings may be identified by searching the websites of HTA and related agencies, professional associations. Key information may also be extracted from the life sciences database BIOSIS (http://science.thomsonreuters.com/training/biosis), which includes patents, journals, conferences, books, review articles etc. While selection of the most relevant of these sources to search will largely depend on the technology in question, compilations of potentially relevant sources of information, such as the HTAi IRG Vortal (http://www.htai.org) and Institute of Health Economics (IHE) ‘Health technology assessment on the net’ report (http://www.ahfmr.ab.ca) can provide a useful starting point (see also other sources in [111] in Appendix 1).

If the technology has obtained regulatory approval then the information that has been submitted as part of the approval process could be used as a source of data on the description and technical characteristics of the technology. This may be available from the major EU or US regulatory bodies as well as regulatory bodies in those countries where the technology has been approved for use (see [109] in Appendix 1). Further information (e.g. description of the technology, expected performances, and intended use) can be obtained from the manufacturers of the technology by direct request or through their website.

There may be also relevant user information on clinicians', nurses', paramedics' and patients' web sites. Published information may be supplemented through contacts or interviews with appropriate experts and agencies.

Reporting and interpreting

The users of HTA require sufficient information on the design and function of the technology to understand the technology’s mode of action, its technical requirements and possible problems and alternatives, its staffing requirements, its applicability range, its variants, and its possible direct risks. For medical devices it may be helpful to include drawings or schematics for the technology that illustrate the components, dimensions and materials of construction of the device.

For diagnostic and monitoring technologies (laboratory tests, imaging, questionnaires etc), it is important to include sufficient information about the technical precision of the technology. This information, which is different from the accuracy data presented in the clinical effectiveness domain, should be reported in this domain.

For management processes (such as screening programs) the position and interaction of the technology within the broader healthcare sequence should be described. This also may require listing alternative technologies.

Assessment elements

B0001 Assessment element card

Issue: What is this technology?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Type of device, operation, imaging, etc. Biological rationale and mechanism of action of the technology. Technology may include a single device, a questionnaire, imaging or sequence of technologies. The HTA may address one or several similar technologies. Minor modifications between manufacturers/products need to be accounted for as these may affect performance.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, introduction sections of research articles.

References
Content relations
Sequential relations

B0002 Assessment element card

Issue: Why is this technology used?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Describe the aim of using the technology: How is it expected to be an improvement over previous / existing technologies used for the same health problem?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, introduction sections of research articles, grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0009, A0018, D1019, C0008

Sequential relations

B0004 Assessment element card

Issue: Who will apply this technology?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Which professionals (nurses, doctors, other professionals) use the technology?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Current Use

Sequential relations

B0016 Assessment element card

Issue: To what population(s) will this technology be used on?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The technology might behave differently in different patient groups. Define as many subgroups as possible. The technology might behave differently in different patient groups. Define as many relevant subgroups as possible (e.g., ‘optimal’ age group versus optional age groups). Are there specific populations that should not be recipients of the technology because of technical difficulties, inaccuracy, inconclusive results or because of safety issues? Does the population need to use the technology more than once? In that case how many times, and how frequently?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0007, C0005

Sequential relations

B0003 Assessment element card

Issue: What is the phase of the technology?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Is the technology an innovation? When was it developed? Is the technology only partially innovative (i.e. a modification of an existing technology), and in that case, is it possible to specify the degree of innovation the technology may represent? When was the technology introduced into healthcare? Is the technology an already established one, but now used in a different way, for instance for a new indication? Most technologies will be introduced at approximately the same time in several countries. The evidence base (published trials etc) may change rapidly for technologies that are at an earlier stage in their development.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, introduction sections of research articles, grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0019, A0020, F0001

Sequential relations

B0017 Assessment element card

Issue: Is this technology field changing rapidly?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

For end users it is useful to know if new versions or adaptations of the technology are expected in the near future.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, clinical trial sites, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Sequential relations

B0006 Assessment element card

Issue: Are there any special features relevant to this technology?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

How does this technology differ from its predecessors (other technologies used for similar purposes)? Are there new aspects that may need to be considered when applying it? Is there evidence that the technology works (or is used) outside its current indication area or produces incidental findings that can have consequences relevant to effectiveness, safety, organisational, social and ethical domains.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, introduction sections of research articles, interviews with specialists, grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0018, C0007, C0060, D0022

Sequential relations

B0005 Assessment element card

Issue: In what place and context is the technology intended to be used?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It can be primary care, secondary care or self care. Its role in the management pathway can be as a replacement, an add-on or for triage.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Current Use, D1007, G001, G0005

Sequential relations

B0018 Assessment element card

Issue: Are the reference values or cut-off points clearly established?

Topic: Features of the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Are conflicting /varying definitions of an abnormal finding likely to affect the interpretation of the results?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Sequential relations

B0007 Assessment element card

Issue: What material investments are needed to use the technology?

Topic: Investments and tools required to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Devices, machinery, computer programs, etc. Those parts of the technology that need to be purchased (and often installed) by an organisation in order to use the technology. Includes need for back-up investment to cover for breakdowns in use.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, published literature including reviews, textbooks, handbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

E0001, E0002, G0006

Sequential relations

B0008 Assessment element card

Issue: What kind of special premises are needed to use the technology?

Topic: Investments and tools required to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Many technologies require purpose-built premises within organisations, such as radiation-secured areas, Faraday cages, etc. Typical premises in primary or secondary care may differ markedly from country to country. A clear description of necessary facilities is needed instead of general statement (e.g. to be used in hospitals only)

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, approving authority, published literature including reviews, handbooks, textbooks, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational domain

Sequential relations

B0009 Assessment element card

Issue: What equipment and supplies are needed to use the technology?

Topic: Investments and tools required to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Syringes, needles, medicines, fluids, bandages etc. All disposable items necessary for using the technology

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, including published literature such as reviews, introduction sections of research articles, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

E0001, E002

Sequential relations

B0010 Assessment element card

Issue: What kind of data and records are needed to monitor the use the technology?

Topic: Investments and tools required to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

What kind of data needs to be collected about the use of this technology regarding care processes, professionals involved, patients and their health outcomes? How is this collected?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

HTA-reports, local authorities

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

G0008

Sequential relations

B0011 Assessment element card

Issue: What kind of registers are needed to monitor the use the technology?

Topic: Investments and tools required to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Are there existing registries that could be used, or should a registry be established to collect the necessary data?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

HTA-reports, local authorities

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

G0008

Sequential relations

B0012 Assessment element card

Issue: What kind of qualification, training and quality assurance processes are needed for the use or maintenance of the technology?

Topic: Training and information needed to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

We need to differentiate between the users who are. 1. applying the technology (could be different from those interpreting results) 2. interpreting the results and make treatment decisions. 3. taking care of service and maintenance. Training materials: writing and/or translation, other adaptation? Personal training: individual and/or group sessions, number and length of sessions, number and qualifications of trainers. Are regular or frequent standardisation or quality checks required? E.g. CME points.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturers´ sites, approving authority, published literature including handbooks, textbooks, reviews, HTA-reports, interviews with specialists, as well as grey literature, hand-searches and conference proceedings.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

G0003, C0020, C0062, C0063

Sequential relations

B0013 Assessment element card

Issue: What kind of training is needed for the personnel treating or investigating patients using this technology?

Topic: Training and information needed to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Training materials: writing and/or translation, other adaptation? Personal training: individual and/or group sessions, number and length of sessions, number and qualifications of trainers. If the technology requires a specific skill that is developed over a period of time using the technology (learning curve), an estimate should be provided of the number of patients a professional needs to treat (as a basis or per year) in order to reach an acceptable minimum standard

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturer, effectiveness studies, observational studies, applicability studies, clinical experts, user information. National or local judgement.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0062, C0063, D1008, G0003

Sequential relations

B0014 Assessment element card

Issue: What kind of training and information should be provided for the patient who uses the technology, or for his family/carer?

Topic: Training and information needed to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Training materials: writing and/or translation, other adaptation? Personal training: individual and/or group sessions, number and length of sessions, number and qualifications of trainers Informed consent regarding the risk / benefits of participation.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturer data, effectiveness studies, observational studies, applicability studies, clinical experts, user information, HTA-reports.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0001, C0003, C0005, C0007, C0062, F0004, F006 G0004, H0003, H0007, H0008, I0002

Sequential relations

B0015 Assessment element card

Issue: What information of the technology should be provided for patients outside the target group and the general public?

Topic: Training and information needed to use the technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Information materials: writing and/or translation, other adaptation? Informed consent for participating?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

HTA-reports, manufacturers' sites, interviews, as well as grey literature, hand-searches and conference proceedings..

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

F0005, F0011, G0004, H0002, H0007, H0008, I0002, I0008

Sequential relations

Safety

Iris Pasternack, Lidia Becla, Tom Jefferson, Aurora Llanos, Leonor Varela Lema

Description

What is this domain about?

Safety is an umbrella term for any unwanted or harmful effects caused by using a health technology. Safety information, balanced with the effectiveness data, forms the basis for further assessments of the technology on e.g. costs and organisational aspects.

There are several ways to categorize harms:

  • A technology may have direct harm; mortality, morbidity or disability due to e.g. radiation, toxicity or invasiveness; or it can indirectly cause harm due to e.g. insufficient training, experience, maintenance of equipment, or inappropriate patient selection.
  • Indirect harms can further be categorized into operator or setting dependent and patient dependent risks. The former can be modified by changing practices or affecting users´ knowledge, skills and behaviour. Latter means that there are vulnerable patient groups in whom protection is especially required.
  • Harms are usually classified according to their fatality or intensity into mild, moderate, and serious or severe (Higgins 2008). ‘Serious’ refers to adverse effects that have significant medical consequences, e.g. lead to death, permanent disability or prolonged hospitalisation. In contrast, ‘severe’ refers to the intensity of a particular adverse effect. For example, a non-serious adverse effect, such as headache, may be severe in intensity (as opposed to mild or moderate).  The term ‘risk’ includes both the seriousness and the probability of the harm. Thus, moderate but very rare harm results in low to moderate risk, whereas even a mild harm with high occurrence is seen as a high risk.  
  • They can be classified according to their dose-relatedness or time-relatedness
  • Harms do not occur only in patients or individuals using the technology. Their family and close ones, other patients, health care professionals, public, and the environment can be affected also.

The definitions and the terminology of safety used in HTA have not been standardised. Frequently used terms include: side-effects, adverse events or adverse effects, complications, harms, risks and hazards, safety, tolerability and toxicity. It has been suggested that the term ‘harms’ should replace the use of the word safety in randomized trials (Ioannidis 2004). ‘Harm’ defines something once it has occurred, whereas ‘risk’ includes both the seriousness and probability of the harm. Thus, a moderate harmful effect . The Cochrane Handbook proposes some definitions for safety related terms (Higgins 2008). A number of initiatives aim to harmonise safety terms. Examples include the National Cancer Institute severity grading system http://ctep.cancer.gov/reporting/CTC-3.html and the WHO system-organ class categories http://www.umc-products.com/graphics/3149.pdf.  Some researchers have found that the standard ‘preferred terms’ can distort descriptions in the original reports of adverse events and blur distinctions between them (Medawar 2003).

. Why is this domain important?

Assessment of safety issues is especially needed when

  • The technology has major risk of harm
  • The margin between benefit and harm is narrow
  • Several technologies with similar effectiveness can be used for the condition, and they have different safety profiles
  • The rate of false positive in a diagnostic test is high and patients may end up with unnecessary potentially harmful investigations or treatments
  • Adverse effects or poor tolerability threatens the acceptability and use of the technology (modified from  Loke 2007) (Loke 2007).

Relations to other domains

Work in the safety domain should be carefully coordinated with the effectiveness domain. Benefit-risk balance is an essential issue in the effectiveness domain. It is worthwhile to discuss how to avoid duplicate work in finding information for that. Safety domain may require information from health problem and current use, description and technical characteristics, and ethical domains. Information provided by safety domain is of relevance to at least organisational, costs and economic evaluation, ethical and possibly also legal domains.

Specific features in finding, interpreting or implementing information

Systematic assessment of all safety issues of a technology can be time consuming. Authors of a Core HTA may need to limit themselves to the safety issues that are significant for patients, or most likely to be important in guiding the decision of health care providers and policy makers (Busse 2002). Severe and serious harms should always be reported. Mild harms should be considered if they can be accumulated or if they influence acceptability or are of importance for patients.

Issues specific for screening technologies

While screening technology is used for large number of healthy persons, the tolerance threshold for risks should be very low (Kristensen 2007). Indirect harms specific to screening technologies are:

–        False positive results, which may cause stress, anxiety, and lead to unnecessary, possibly harmful further investigations or treatments.

–        False negative results of screening test may have the potential to delay the detection of the illness. The false negative results may have medical, psychological, economic, and legal consequences.

–        True negative test result may reduce normal alertness to symptoms of disease and lead to false sense of security.

–        Overdiagnosis and overtreatment can be a problem if screening tends to detect and lead to treatment of conditions with good prognosis, even if left untreated. The same occurs if screening detects other conditions than the one it is aimed to detect.  

Methodology

Where to find information?

Databases and search strategies

–        IAEA: Safety standards for diagnostic radiology http://www-pub.iaea.org/MTCD/publications/PDF/Pub1206_web.pdf

–        IAEA: Radiological protection of patients http://rpop.iaea.org/RPoP/RPoP/Content/index.htm

–        ICRP: Publications of International Comission of Radiological Protection http://www.icrp.org/

–        TGA (Therapeutic Goods Administration), http://www.tga.gov.au/index.htm

–        The Medical Devices section of the UK Medicines and Healthcare Products Regulatory Agency http://devices.mhra.gov.uk/

–        Canada Vigilance Adverse Reaction Online Database http://www.hc-sc.gc.ca/dhp-mps/medeff/databasdon/index-eng.php

–        Manufacturers product data sheets or applications for a product license

Searches do not detect all relevant studies while indexing terms for adverse effects are not always assigned in original studies, and the authors do not mention adverse effects in the title or abstract (Derry 2001). To improve the sensitivity of the search, terms for specified adverse effects have to be defined and looked up in each database thesaurus to identify the relevant subject headings to be added in the search strategy (Golder 2006). New, previously unrecognised remain therefore easily undetected (Golder 2006b).  Several study types should be considered for inclusion in the search.

There is no optimal search strategy for specifically identifying reports of adverse effects. There are several highly sensitive (97-100%) search strategies, but the problem is their low specificity 0.9- 2.8%). This precision means that for each retrieved relevant article on adverse effects one have to screen 36-125 other records (Golder 2009).

Following approaches can be used to complement the search strategy with key elements derived from study population and the technology in question:

  • Index terms (thesaurus terms, e.g. MeSH in Medline)
    • For specified adverse effects: e.g gastrointestinal hemorrhage, lymphedema, pain, nausea, lethargy, fatigue
    • For risk in general: e.g. Adverse Effects (subheading), safety, toxicity, drug toxicity, complications
  • Subheadings/qualifiers either attached to technology name indexing terms or "floated", i.e. searched without being attached to an indexing term (floating subheadings) 
  • Text words (terms used by the original authors in title and abstract), also taking into account different conventions in spelling and variations in the endings of the terms.
    • For specified adverse effects: nausea, pain, anxiety, tiredness, lethargy, malaise, fatigue
    • For risk in general: side-effect, adverse effect/event/reaction, complications, poisoning, drug effects, safety management.
  • Index terms and text words to capture certain study design, such as cohort studies or case reports.

There are suggested search strategies for MEDLINE and EMBASE (Golder 2010 and 2006) and other sources (The InterTASC Information Specialists' Sub-Group, http://www.york.ac.uk/inst/crd/intertasc/adverse.htm).

The methodological approach used in identifying literature on harms can vary and the approach adopted will lead to different estimates of risk (McIntosh 2004). Therefore, the search strategies for electronic reference databases and study inclusion criteria should be clearly reported. This applies also for information retrieved elsewhere.

Search issues specific for screening technologies

Suggested index terms:

  • Primary Prevention [Mesh] or Mass Screening [Mesh] or Public Health Practice [Mesh]
  • Medicalisation, false positive, false negative, over-diagnosis, over-treatment

Example: Suggested search strategy in CURRENT CONTENTS.

  • #1 «Primary Prevention» [Mesh] or «Preventive Health Services» [Mesh] or «Mass Screening» [Mesh] or «Public Health Practice» [Mesh]
  • #2 «screening»
  • #3 «preventive drug»
  • #4 «preventive drugs»
  • #5 #1 or #2 or #3 or #4
  • #6 «Safety Management» [Mesh] or «adverse effects» [Subheading]
  • #7 «safety»
  • #8 «adverse events»
  • #9 «medicalization»
  • #10 #6 or #7 or #8 or #9
  • #11 #5 and #10
  • #1 «screening»
  • #2 «false positive»
  • #3 «false positives»
  • #4 «false negative»
  • #5 «false negatives»
  • #6 #2 or #3 or #4 or #5
  • #7 #1 and #6

 |}}

Useful other sources
  • Drug monographs
  • Industry submissions
  • Bulletins
  • Conference proceedings
  • Reference checking
  • Hand searching
  • Personal communication
  • Manufacturers Periodic Safety Update Reports (PSURs)
  • National or international safety monitoring systems (databases) of adverse events which may be managed by a national statutory body or by a supra-national body ([110] in Appendix 1).
  • Disease ([105] in Appendix 1) or technology registries ([104] in Appendix 1) of patients receiving treatment which may be organised at an international, national or regional level and managed by a government agency, professional body or the manufacturer.
  • In some cases routine statistics from hospital, primary care or health system funders may be available and provide suitable information
    • Specific enquiries to manufacturers, regulators or professional bodies may help identify additional sources of information.
    • Information from patient associations may provide valuable patient experiences especially in emerging technologies (Cross 2005).
    • Internet discussion forums may provide valuable, but probably unreliable, additional information.

Inclusion of unpublished studies can provide additional adverse effects information and more precise risk estimates. However, there is insufficient evidence to indicate whether inclusion of unpublished studies has a major influence on the pooled risk estimates in meta-analyses of adverse effects (Golder 2010b).

What kind of information is required?

Study types, design, outcome measures

Randomised controlled trials, observational studies and case reports provide evidence on the frequencies of harms. Randomised trials are methodologically most solid, and may alone be an appropriate source of evidence for some review questions about harm. However, safety reporting in randomized trials is heterogeneous and often inadequate (Pitrou 2009, Ioannidis 2001). Rare adverse effects are not usually detected in randomised trials, and even relatively frequent harms with a longer latency period cannot be quantified easily.  Information about new, serious, rare or long-term adverse effects are thus typically found in observational studies (cohort, case-control, nested case-control, and cross-sectional studies).

Besides published research, routinely collected data can be used. Often these databases are generic and may not contain enough information. However, their advantages are bigger size or coverage over long periods of time (Busse 2002). Their information is especially relevant in the assessment of e.g. public preventive programs.

Spontaneous reporting of adverse drug reactions is a standard method to identify safety signals for marketed drugs. Its primary purpose is to provide early warnings of adverse drug reactions not recognized prior to the marketing. Once a signal has been identified, other methods will be used to quantify the potential risk in order to avoid unnecessary alarms.

The risks are sometimes quantified as a quality-adjusted life years (QALYs) or disability-adjusted life years  (DALYs).  QALYs are non-disease specific measurement of outcomes that incorporates both quality and duration of life, defined as years of healthy life lived (Drummond 2005). DALYs are defined as years of healthy life lost. DALYs and QALYs are complementary concepts and both approaches multiply the number of years by the quality of those years. In order to reflect the burden of the same states QALYs use “utility” weights of health states, whereas DALYs use “disability weights”. QALYs and DALYs simultaneously capture both positive and negative changes in morbidity and mortality associated with treatment-related benefits and risks, and translate outcomes from different disease states into a comparable common metric that is useful for subsequent quantitative benefit–risk analysis (Arnesen 1999, Rehm 2010)

Results from trials are usually presented as information on the frequency of occurrence, relative risk RR, risk difference (RD), odds ratio (OR), or number needed to harm (NNH). Estimates of risk from case-control studies are presented in exposure odds ratio of cases compared with controls. The unintuitive odds ratios have been used to calculate the additional absolute risk of an adverse event NNTH (number of patients needed to be treated for on additional patient to be harmed) (Bjerre 2000).

For meta-analysis risk ratio (RR) is the most common summary statistic, followed by Peto odds ratio. Risk difference (RD) is rarely used in meta-analyses although it is the most interpretable statistics and is particularly appropriate in examining rare event data (Deeks 2002).

Analysing data based on NNH can be dangerous since this measure can be very sensitive if the point estimate is close to zero (i.e. close to 1 for an OR or RR and close to 0 for a RD) (Vandermeer 2009).

Issues specific for screening technologies

Diagnostic accuracy studies are essential in the assessment of screening technologies in order to assess sensitivity and specificity of the test itself, and the rates of false negative and false positive results and their consequences.

A basic diagnostic accuracy study consists of a group of patients in whom the target disease is suspected. All of them undergo the test under consideration (index test) and the best possible test to verify the diagnosis (reference standard, gold standard). The results of the index test(s) are then compared to the results of the reference standard. Positive and negative results from both tests are shown in a 2x2 table or a variation thereof, depending on the number of cut-off points chosen.

If there is no appropriate reference test it is possible to construct a reference diagnosis by using a predefined rule for a set of other tests, consensus among experts, or a statistical model based on actual data (Rutjes 2007). Another possibility is to investigate the probability of disease presence as a function of all diagnostic variables simultaneously with multivariable modelling (Moons 1999). Problems may arise from the spectrum (patient characteristics, patient selection and setting), the non-optimal reference standard, partial verification (not all patients receive the reference test) or differential verification (patients receive different reference tests).

Tools for critical appraisal

There is often a trade-off between the comprehensiveness and quality of the risks data to be included in an assessment. Including evidence that is likely to be biased, even if no better evidence exists, may lead to biased conclusion. All included data should be critically appraised. There is a lack of a relevant quality assessment tool to risk analysis (Loke 2007). Any available tool should be used cautiously. Comparing evidence from randomised trials and observational studies is useful.

The authors of a Core-HTA-report should consider at least some important aspects:

  • How rigorous were the methods used to detect adverse effects? Were the methods used for monitoring reported?
  • Was follow up sufficiently long to assess the risk for serious longer term safety issues? 
  • How complete is the reporting? Did the investigators report all important or serious harms? Did the report give numerical data by group?
  • How were data collected: prospective/routine monitoring, spontaneous reporting, patient checklist/questionnaire/diary; systematic survey of patients
  • Were any patients excluded from the risks analysis

Different methods of monitoring risks yield different results, which make comparisons between studies meaningless. Active surveillance and use of checklists yield higher harm frequencies than passive or less-focused methods (Loke 2007). Authors in the original studies may report only some outcome categories although they measured several, or the intervention groups may be combined (e.g. X participants withdrew from the study), or the statements are unclear or too generic (e.g. no unexpected adverse effects were seen).

Systematic reviews of adverse effects have often used inadequate searches to identify studies (Golder 2008).

Trials

Adverse events are variably and sometimes poorly reported in randomised trials (Pitrou 2009), and in systematic reviews of trials (Ernst 2001, Golder 2006b). The definition of a particular risk may vary between studies, as can definitions of severity. They can be measured in different ways and different thresholds can be used. An extension of the CONSORT Statement (Consolidated Standards for reporting Trials) is made for better reporting of harms in randomised trials (Ioannidis 2004).

Basic requirements for the data are: it should be presented in numbers; the severity of adverse effects should be stated (at minimum the frequency of severe events should be provided per study arm); and the data should be given separately for each type of adverse effect (MacMahon 2001). The analysis of zero events ("no serious adverse effects were seen") needs careful consideration. Before concluding that no adverse effect occurred, reviewers should ask themselves how thorough were the methods used to detect adverse effects in the original studies and how many patients were studied and for how long (Loke 2007)?

Even in cases where adverse events are examined and reported adequately, there is often insufficient evidence for conclusion since most trials are tailored towards optimizing efficacy estimates (Vandermeer 2009).

Many trials are too small for reliable estimates and they are usually not designed to collect information of adverse events, at least not as their primary outcomes. This may lead to partial or inadequate reporting of harms: lumping adverse effects of varying seriousness or severity into one number, or giving only generic statements like "few patients had adverse effects". Note, that no mention of harms in an original study does not necessarily mean that no harms occurred. Authors must choose whether to exclude the study from the risk analysis or, exceptionally, to include it on the assumption that the incidence was zero (Loke 2007).

Caution is needed when interpreting withdrawal or drop-out data as surrogate measures for safety or tolerability. The reason of withdrawal can be anything from mild side effects to serious toxicity or lack of efficacy or non-medical reason (Ioannidis 2004). Patients in trials and investigators may be more (or less) willing than generally to continue in trial although there are some side effects (Loke 2007).

Observational studies

Trials may report small, fragmented pieces of evidence of risks that are not primary outcomes, whereas observational studies may be primarily devoted to assessing specific risks. Nested case control studies, full cohort analysis, and survival analysis methodologies are the study designs frequently used for risk assessment. Major sources of bias in observational studies are confounding by factors associated with both treatment and outcome, bias due to differential recall of exposure, and bias due to differential detection of outcomes (MacMahon 2001). A brief summary of the strengths and weaknesses of different study designs that may be included in a systematic review of harms is given by Jefferson and Demicheli (Jefferson 2003). Newcastle Ottawa scale is a tool to assess observational studies, available at http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm . STROBE-Statement provides a checklist of items that should be addressed in reports of observational studies (vonElm 2007).

Case reports of suspected adverse events are widely published in scientific journals and few of these reports have been subsequently investigated or confirmed to be valid (Loke 2006). Some spontaneous reporting systems are inevitably erroneous (Loke 2007).

Issues specific for screening technologies

Aspects of study quality of diagnostic accuracy studies include the selection of a clinical relevant cohort, the consistent use of a single good reference standard, and the blinding of results of experimental and reference test (Deeks 2001). QUADAS tool is a checklist to assess quality of diagnostic accuracy studies.

Quality assessment of diagnostic accuracy studies is subjective and hampered by poor reporting. Incorporation of quality in overall assessment is difficult due to limited studies. Relation between quality items and bias are not as straightforward as it is for interventions. 

There are many different tools to assess the quality of diagnostic accuracy studies. Cochrane handbook (Higgins 2008) recommends QUADAS tool with its 11 mandatory and more than 10 facultative items. HTA-authors create an own selection of relevant items presented in the tool. Two assessors are recommended. Background of assessors should be reported, and the way they resolved disagreements. Results of the quality assessment of the original studies should be presented in a table or graphically. Individual quality items should be investigated as a potential source of heterogeneity. See more about Quadas tool in Appendix 3.

Analysing and synthesizing evidence

A systematic approach is required in the assessment of safety (risks). Core HTA authors, who are not aware of any specific safety problem, usually start with a broad overview of the whole range of adverse effects associated with the use of the technology. They may be confronted with an unstructured mix of lists and texts covering many diverse outcomes due to lack of consistency of reporting harms. A predefined classification of adverse effects could help the authors to approach the data (Loke 2007).

The aim is not necessarily to cover all known and previously unrecognised risks of a technology.  Rather, Core HTA preparers should focus their review and predefine the safety issues and outcome measures they wish to work in their assessment (Higgins 2008). The demographic characteristics of the population in which the technology is to be used should be defined for later comparison against the populations in which safety data has been identified.

Core HTA authors may choose to narrow down into some of the following areas:

  • the five to ten most frequent adverse effects
  • all adverse effects that either the patient or the clinician considers to be serious
  • the most common adverse effects that lead the patient to stop using the intervention;
  • By category, for example:
    • diagnosed by clinician (e.g. gastrointestinal haemorrhage)
    • diagnosed by lab results (e.g. hypokalaemia)
    • patient-reported symptoms (e.g. pain).
    • biomarkers that may be early indicators of possible adverse effects (for example, abnormal liver enzymes); offering a means of collecting relevant information even from short-term studies.

This is not a comprehensive list, but the use of any of the above strategies should help authors approach the adverse effects analysis in a systematic, manageable and clinically useful fashion (Higgins 2008).

Biases, confounding factors, level of evidence

Harms are frequently insufficiently reported (Pitrou 2009). Poor safety reporting of the original research can lead to misinterpretation and inadequate conclusion of the technology assessed.

Reported harm frequencies may differ greatly in different study types. A study comparing harms reported in randomised and observational studies found that observational studies yield lower estimates of absolute risk of harm (Papanikolaou 2006).

Randomized trials have frequently restrictive inclusion/exclusion criteria which can underestimate harm. Most preliminary trials exclude specific sensitive subgroups because of ethical concerns, or include them in insufficient sample size.

Individual measurements of late onset harms (e.g number of radiation induced cancers) can usually not be seen in research publications. Frequency of such stochastic harms is always an estimate, and based on analogies and presumptions from epidemiological risk research.

Adverse effects data is usually as well reported in industry funded than in non-industry funded studies. However, there is a risk that interpretations and conclusions of industry funded authors carry potential bias (Golder 2008b).

Evidence tables

A table of included evidence might be a helpful way to make overall assessment for each assessment element. The table could contain following information for each included piece of evidence.

  • Reference: article/ book/report/ web/ database reference
  • Source: name of reference database, agency, discussion forum, other, e.g. Medline, IAEA.
  • Study/information type: e.g. prospective cohort study, trial, systematic review, HTA report, manufacturer report, register data, consensus
  • Which harm?
  • Intensity: 1=mild, 2=moderate, 3=serious/severe
  • Other classification: self reported/objective measure, immediate/delayed etc.
  • Number of harm events per study arm
  • Quality of information: how was data collected etc
  • Comments on generalisability of the evidence
Meta-analysis

Safety estimates usually require larger sample size to detect differences in patient groups in trials. Safety events are usually rare (incidence <5%). Exact methods seem to be superior to the asymptotic Mantel-Haenzel method for rare event data, and to the Peto method when trials are balanced (Bradburn 2007). Asymptotic approximations are known to be imprecise with rare events; still majority of systematic reviews use them.

While asymptotic approximations in dichotomous data require a non-zero event rate, most reviewers add 0.5 to each cell in stead of zero. This approach is inappropriate if the event is rare. Exact methods do not provide a point estimate in a situation where no events are observed in one arm, which is intuitively acceptable too.

Read more about meta-analysis of diagnostic accuracy studies in Appendix 3.

Qualitative synthesis

At this stage authors of a Core-HTA-report should check, that the data extracted is relevant to the research questions, and that analysing and synthesizing the data is still answering the question. Often the evidence available is not quite as useful as hoped, and in that case it should be made explicit how well it answers the original research question.

In many circumstances it is not possible to calculate frequencies, and information about risks is best presented in a qualitative or descriptive manner. Data derived from different study designs, different populations or different data collection methods cannot be combined. Anticipated risks can be reported congruently, whereas unanticipated risks that are detected during a trial might be reported in a markedly different ways by different investigators (Papanikolaou 2004).

There is no consensus on how to synthesise information about quality from a range of study designs within a systematic review. Special techniques have been tried (Jefferson 2003, Wald 2003).

Reporting and interpreting

The interpretation of evidence should clearly state qualitative and quantitative limitations of the sources, searches, data and methods used for the analysis. Presentation through tables is transparent and may be helpful in summarising different data (Busse 2002). The sources of information should be clearly stated.

When discussing the safety of a technology, the way harms were caused should be described. Harm may be device dependent or related to the application of the technology. Occurrence of adverse effects may be also operator or setting dependent (e.g. learning curve). Timing and severity of adverse effects should be considered too and the differences in risk among different groups of patients.

It is recommended that whenever possible the overall effect of the harms needs to be quantified, as a QALY or DALY as well as information on the frequency of occurrence, relative risk or number needed to harm (NNH). NNH is perceived as the most understandable summary statistic for adverse events, A small absolute risk is still clinically important if an adverse event is serious or severe, or if the absolute benefit of the intervention is also small (Papanikolaou 2004). Comment should be made about the generalisability of the findings to the population for whom the HTA may be used.

In RCTs presenting adverse event rates, non-statistically significant differences are associated with low statistical power. A high probability of type II error may lead to erroneous inferences (Ioannidis 2001).

Assessment elements

C0001 Assessment element card

Issue: What kind of harms can use of the technology cause to the patient; what are the incidence, severity and duration of harms?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Observational research, safety monitoring databases, registers, statistics

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Loke Y 2006, 2007, Ioannidis 2001, 2004, Higgins 2006, Papanikolaou 2006Busse 2002Golder 2006, Mac Mahon 2001

Content relations
Sequential relations

C0002 Assessment element card

Issue: What is the dose relatedness of the harms to patients?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Here one should consider also the accumulated harm due to repeated dosage or testing

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Aronson 2003

Content relations
Sequential relations

C0003 Assessment element card

Issue: What is the timing of onset of harms to patients: immediate, early or late?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Aronson 2003

Content relations
Sequential relations

C0004 Assessment element card

Issue: Is the incidence of the harms to patients likely to change over time?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

For some technologies the occurrence of harms may change over time and be dependant on the experience or training of the operator?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Medical literature/ grey literature/ professional societies/ registries

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Current use, effectiveness, costs domains

Sequential relations

C0005 Assessment element card

Issue: Are there susceptible patient groups that are more likely to be harmed through use of the technology?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Aronson 2003

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical, F0005

Sequential relations

C0006 Assessment element card

Issue: What are the consequences of false positive, false negative and incidental findings brought about using the technology to the patients from the viewpoint of patient safety?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Effectiveness, Social, Costs, Ethical and Legal domains

Sequential relations

C0029 Assessment element card

Issue: Does the existence of harms influence tolerability or acceptability of the technology?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Qualitative research articles, patient associations’ web sites, Internet discussion forums

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Effectiveness, Social, Ethical and Legal domains

Sequential relations

C0007 Assessment element card

Issue: What are the special features in using (applying/interpreting/maintaining) the technology that may increase the risk of harmful events?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Is there evidence for operator dependent harms? Is there a learning curve and what is its consequence? Is there is a big intra- or inter-observer variation in the reading of test results, what is its consequence?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Description and technical characteristics and Organisational domains

Sequential relations

C0008 Assessment element card

Issue: What is the safety of the technology in comparison to alternative technologies used for the same purpose?

Topic: Patient safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Current use, Clinical Effectiveness and Ethical domains

Sequential relations

C0020 Assessment element card

Issue: What kind of occupational harms can occur when using the technology?

Topic: Occupational safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical and Social domains

Sequential relations

C0040 Assessment element card

Issue: What kind of risks for public and environment may occur when using the technology?

Topic: Environmental safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical and Social domains

Sequential relations

C0060 Assessment element card

Issue: How does the safety profile of the technology vary between different generations, approved versions or products?

Topic: Safety risk management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Description and Technical Characteristics

Sequential relations

C0061 Assessment element card

Issue: Is there evidence that harms increase or decrease in different organizational settings?

Topic: Safety risk management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy and effectiveness research, epidemiological risk research

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Current use, Effectiveness, Organisational

Sequential relations

C0062 Assessment element card

Issue: How can one reduce safety risks for patients (including technology-, user-, and patient-dependent aspects)?

Topic: Safety risk management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technical means, protective equipment, education etc. Including information on what kind of risk communication is needed for patients, citizens and decision makers

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets, safety monitoring databases

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical F0006, Description and technical characteristics B0012, B0014, B0015

Sequential relations

C0063 Assessment element card

Issue: How can one reduce safety risks for professionals (including technology-, user-, and patient-dependent aspects)?

Topic: Safety risk management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technical means, protective equipment, education etc. Including information on what kind of risk communication is needed for patients, citizens and decision makers

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research in occupational health and safety

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Social Domains

Sequential relations

C0064 Assessment element card

Issue: How can one reduce safety risks for environment (including technology-, user-, and patient-dependent aspects)?

Topic: Safety risk management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technical means, protective equipment, education etc. Including information on what kind of risk communication is needed for patients, citizens and decision makers

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Research articles, manufacturers' product data sheets.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social Domain

Sequential relations

References

Arnesen T, Nord E. The value of DALY life: problems with ethics and validity of disability adjusted life years. BMJ 1999;319:1423–5

Bjerre LM, LeLorier J. Expressing the magnitude of adverse effects in case-control studies: "the number of patients needed to be treated for one additional patient to be harmed". BMJ. 2000 Feb 19;320(7233):503-6. Review.

Bradburn MJ, Deeks JJ, Berlin JA, Localio AR. Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events. Statistics in Medicine.2007 January; 26(1):53-77

Busse R, Orvain J, Velasco M, Perleth M, Drummond M, Gurtner F, et al. Best practice in undertaking and reporting health technology assesments. International Journal of Technology Assessment in Health Care 2002;18(2):361-422.

Cross M. Drug companies monitor web chat for adverse reaction reports. BMJ 2005 June 11;330(7504):1347.

Deeks JJ. Issues in the selection of a summary statistic for meta-analysis of clinical trials with binary outcomes. Stat Med. 2002 Jun 15;21(11):1575-600.

Deeks JJ. Systematic reviews in health care: Systematic reviews of evaluations of diagnostic and screening tests. BMJ 2001 Jul 21;323(7305):157-162.

Derry S, Kong Loke Y, Aronson JK. Incomplete evidence: the inadequacy of databases in tracing published adverse drug reactions in clinical trials. BMC Med.Res.Methodol. 2001;1:7.

Deville W, Buntinx F, Bouter L, Montori V, de Vet H, van der Windt, Danielle, et al. Conducting systematic reviews of diagnostic studies: didactic guidelines. BMC Medical Research Methodology 2002;2(1):9.

Drummond M, Sculpher M, Torrance G, O’Brien B, Stoddart G. Methods for economic evaluation of health programs (3rd edn). Oxford University Press: New York, NY, 2005

Ernst E, Pittler MH. Assessment of therapeutic safety in systematic reviews: literature review. BMJ. 2001 Sep 8;323(7312):546.

Golder S, Loke YK. Sources of information on adverse effects: a systematic review. Health Information & Libraries Journal 2010;27(3):176-190.

Golder S, Loke YK, Bland M. Unpublished data can be of value in systematic reviews of adverse effects: methodological overview. J.Clin.Epidemiol. 2010 b Oct;63(10):1071-1081.

Golder S, Loke Y. Search strategies to identify information on adverse effects: a systematic review. J.Med.Libr.Assoc. 2009 Apr;97(2):84-92.

Golder S, Loke Y, McIntosh HM. Poor reporting and inadequate searches were apparent in systematic reviews of adverse effects. J.Clin.Epidemiol. 2008 5;61(5):440-448.

Golder S, Loke YK. Is there evidence for biased reporting of published adverse effects data in pharmaceutical industry-funded studies? Br J Clin Pharmacol. 2008 b Dec;66(6):767-73. Epub 2008 Jul 31.

Golder S, McIntosh HM, Duffy S, Glanville J, Centre for Reviews and Dissemination and UK Cochrane Centre Search Filters Design Group. Developing efficient search strategies to identify reports of adverse effects in MEDLINE and EMBASE. Health.Info Libr.J. 2006 Mar;23(1):3-12.

Golder S, Loke Y, McIntosh H. Room for improvement? A survey of the methods used in systematic reviews of adverse effects. BMC Medical Research Methodology 2006 b;6(1):3.

Harbord RM, Deeks JJ, Egger M, Whiting P, Sterne JAC. A unification of models for meta-analysis of diagnostic accuracy studies. Biostat 2007 April 1;8(2):239-251.

Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.1 [updated September 2008]. Chapter 14:  Adverse effects. Authors: Yoon K Loke, Deirdre Price and Andrew Herxheimer on behalf of the Cochrane Adverse Effects Methods Group.

Ioannidis JP, Evans SJ, Gotzsche PC, O'Neill RT, Altman DG, Schulz K, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Ann.Intern.Med. 2004 Nov 16;141(10):781-788.

Ioannidis JPA, Lau J. Completeness of Safety Reporting in Randomized Trials: An Evaluation of 7 Medical Areas. JAMA 2001 January 24;285(4):437-443.

Irwig L, Macaskill P, Glasziou P, Fahey M. Meta-analytic methods for diagnostic test accuracy. Journal of Clinical Epidemiology, 1995 1;48(1):119-130.

Jefferson T, Demicheli V. Balancing benefits and harms in health care: observational data on harm are already included in systematic reviews. BMJ 2003 Sep 27;327(7417):750.

Kester AD, Buntinx F. Meta-analysis of ROC curves. Med.Decis.Making 2000 Oct-Dec;20(4):430-439.

Kristensen FB, Sigmunt H; Health Technology Assessment Handbook; Copenhagen: Danish Centre for Health Technology Assessment, National Board of Health, 2007

Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat.Med. 2002 Jun 15;21(11):1525-1537.

Loke Y, Price D, Herxheimer A, the Cochrane Adverse Effects Methods Group. Systematic reviews of adverse effects: framework for a structured approach. BMC Medical Research Methodology 2007;7(1):32.

Loke YK, Price D, Derry S, Aronson JK. Case reports of suspected adverse drug reactions--systematic literature survey of follow-up. BMJ 2006 February 11;332(7537):335-339.

MacMahon S, Collins R. Reliable assessment of the effects of treatment on mortality and major morbidity, II: observational studies. Lancet 2001 Feb 10;357(9254):455-462.

McIntosh H, Woolacott N, Bagnall A. Assessing harmful effects in systematic Reviews. BMC Medical Research Methodology 2004;4(1):19.

Medawar C, Herxheimer A. A comparison of adverse drug reaction reports from professionals and users, relating to risk of dependence and suicidal behaviour with paroxetine. International Journal of Risk and Safety in Medicine 2003; 16: 5-19.

Medical Services Advisory Committee. Guidelines for the assessment of diagnostic technologies. August 2005.

Moons KG, van Es GA, Michel BC, Buller HR, Habbema JD, Grobbee DE. Redundancy of single diagnostic test evaluation. Epidemiology 1999 May;10(3):276-281.

Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat.Med. 1993 Jul 30;12(14):1293-1316.

Papanikolaou PN, Christidi GD, Ioannidis JPA. Comparison of evidence on harms of medical interventions in randomized and nonrandomized studies. CMAJ 2006 February 28;174(5):635-641.

Petticrew MP, Sowden AJ, Lister-Sharp D. False-negative results in screening programs. Medical, psychological, and other implications. Int J Techn Assess Health Care. 2001;17 (2):164-70

Pitrou I, Boutron I, Ahmad N, Ravaud P. Reporting of Safety Results in Published Reports of Randomized Controlled Trials. Arch.Intern.Med. 2009 October 26;169(19):1756-1761.

Reitsma JB, Glas AS, Rutjes AWS, Scholten RJPM, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. Journal of Clinical Epidemiology, 2005 10;58(10):982-990.

Rehm J, Frick U. Valuation of health states in the US study to establish disability weights: lessons from the literature. Int. J. Methods Psychiatr. Res. 2010, 19(1): 18–33

Rutjes AW, Reitsma JB, Coomarasamy A, Khan KS, Bossuyt PM. Evaluation of diagnostic tests when there is no gold standard. A review of methods. Health Technol.Assess. 2007 Dec;11(50):iii, ix-51.

Vandermeer B, Bialy L, Hooton N, Hartling L, Klassen TP, Johnston BC, et al. Meta-analyses of safety data: a comparison of exact versus asymptotic methods. Statistical Methods in Medical Research 2009 August 01;18(4):421-432.

Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med.Res.Methodol. 2003 Nov 10;3:25.

von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007 Oct 20;370(9596):1453-145

Clinical Effectiveness

Petra Schnell-Inderst, Lidia Becla, Tom Jefferson, Marjetka Jelenc, Jaana Leipälä, Chris Lawinski, Claus Løvschall, Heike Raatz, Uwe Siebert

Description

What is this domain about?

The effectivenessdomain in a health technology assessment considers two questions: Can this technology work, and does this technology work in practice? Efficacy is the extent to which an intervention does more good than harm under ideal circumstances. Effectiveness assesses whether an intervention does more good than harm when provided under usual circumstances of health care practice (Haynes 1999). The research questions defined within this domain aim at answering these questions, with emphasis on the second question.

Two or more alternative methods to prevent, diagnose, treat and monitor a clinical condition or to improve the delivery of care are compared in comparative clinical effectiveness research. The two key elements are that effective interventions should be directly compared and studied in patients who are typical of day-to-day health care settings (Sox 2009). The focus is in determining the magnitude of health benefits and harms, and the net benefit (benefits minus harms) that is caused by an intervention, and present the certainty of the evidence (Sawaya 2007). The generally accepted standard for proving the evidence of a causal relationship between intervention and health outcomes is an appropriately designed and conducted randomised controlled trial (RCT), even without a need for a deeper biological theory as to why the intervention works or not (Ashcroft 2002).

The assessment of health benefits should primarily consider patient relevant outcomes such as mortality, morbidity, and quality of life. Intermediate outcomes such as biochemical or physiological markers, or the proportion of early detected cases may be useful and necessary in order to understand how interventions work or as quality assurance benchmarks for health care programmes. If long term clinically important outcomes are not available, surrogate endpoints may be used to indicate or predict clinically important outcomes. To be valid the surrogate must have been shown to correlate with and accurately predict the outcome of interest (CRD 2009).

New diagnostic technologies frequently enter into clinical practice without evidence of improved patient outcomes. Randomised trials of test-and-treatment strategies are not routinely performed, and they are not required for marketing approval. Accuracy studies are far more frequent, but relying on accuracy information only when deciding whether to adopt a new diagnostic test is usually insufficient (Tatsioni 2005).

Why is this domain important?

In health policy, the insurer, agency or government providing care as well as users, citizens and consumers require primarily information on the effectiveness and safety of a technology. It is of no interest to examine the other aspects such as the costs of a technology if the technology is not effective.

Relations to other domains

–        Effectiveness domain requires information from health problem and current use domain, as well as safety domain in order to specify the appropriate populations, interventions, comparisons and outcomes for the research questions.

–        There is a possibility of overlapping with safety domain, so co-operation is needed in the protocol phase.

–        The costs and economic evaluation domain requires information from the effectiveness domain in order to determine the incremental health benefit part of the incremental cost-effectiveness ratio

–        Depending on the technology the ethical domain may be important for the setting of the framework of the effectiveness analysis. For example value judgements in how patient relevant outcomes are defined may be important. (Strech 2008)

–        Effectiveness may sometimes strongly depend on organisational aspects.

–        Effectiveness may also be related to the legal domain, e. g. when there is legal support to a public health programme (mandatory vaccination or mass screening)

Specific features in finding, interpreting or implementing information for this domain

If all trials concerning a technology have been performed under ideal conditions one will have to make assumptions about the magnitude of effectiveness based on the available efficacy data. The challenge is then to examine the reasons why the technology works or wouldn't work in specific circumstances. Long term surveillance information from observational studies usually becomes relevant.

Issues specific for screening technologies

For population based screening programmes the most important determinants of effectiveness are a reduction in disease specific mortality and morbidity and a gain in health related quality of life.

The overall effectiveness of a screening programme is determined by a combination of several factors:

–        the prevalence and incidence of a disease

–        the natural history of disease and the proportion of subclinical or reversible cases that would not become clinically relevant (potential for overdiagnosis and overtreatment)

–        the participation rate as the number of participants divided by the number of eligible individuals in the target screening population

–        the screening interval

–        the accuracy of the screening test

–        the proportion of subjects with positive screening test results which have a diagnostic follow-up

–        the test accuracy of the tests used in the diagnostic follow-up

–        the impact of the test results on treatment decisions and quality of life

–        the effectiveness of the therapies for the cases identified by screening

The evaluation of a screening technology must comprise the whole chain from the screening test with true and false test results, the possibility of adverse effects from the test, the accuracy and potential for adverse effects of the subsequent confirmatory diagnostics, the losses to follow up before the therapeutic intervention is provided, and the effectiveness and adverse events of the therapeutic intervention. (Sawaya et al 2007).

Large randomised controlled trials in a representative asymptomatic population comparing a group invited to screening with a group not invited to screening with a follow-up until all patient relevant outcomes can be analysed are rarely available, especially when the development of the disease takes a long time as, for example, in the case of cancer. Therefore often indirect evidence from different study types has to be linked.

Additionally, it is probable that the effectiveness will fall during the early stages of a new screening programme. This occurs as a larger number of cases (both early stage and late stage disease) are likely to be picked up in the first screening round when compared to later rounds.  Thus it is desirable to analyse the results of several screening intervals in order to estimate the effectiveness of a screening programme. 

Methodology

Methodology

The specification of the research question using the PICO scheme (Appendix 3) is the first step in performing the evaluation of the clinical effectiveness of a technology. The choice of target population, comparisons and outcomes usually has a strong influence on the results on clinical effectiveness. How to do a systematic search of clinical effectiveness, safety and cost-effectiveness is described elsewhere. The clinical effectiveness results are especially sensitive to flaws in the literature search and study selection when the outcomes of interest are quantitatively pooled in a meta-analysis. Results may be substantially biased if relevant studies are not found or not properly selected.

Specific to screening technologies

Starting with the publication of Wilson and Jungner in 1968 different lists of criteria were developed stating under which conditions the introduction of a screening programme might be useful. (Wilson 1968, NSC 2009, Andermann 2008). Many of these criteria directly relate to the clinical effectiveness of the screening test, diagnostic workup and treatment and stress the linkage between them. They are integrated in the following parts.

Where to find information?

Many different sources of information should be searched, including published and grey literature, searching of journals, contacting experts as well as scanning reference lists of relevant papers.

Databases and search strategies

General medical databases such as

  • Medline,
  • Medline in Process,
  • Embase,
  • Cochrane Controlled Trials register

Specialised databases for specific questions such as

  • CINAHL,
  • PSYCINFO,
  • ASSIA, (Applied Social Sciences Index and Abstracts)
  • SOCIOLOGICAL ABSTRACTS
  • Social Services Abstracts,
  • Social Care on line/Caredata and SocINDEX,

Administrative studies: General science publishers’databases such as

  • Emerald Library,
  • Science Direct and Ebsco Academic Search Elite,
  • Pub Med Central (PMC),
  • Bio Med Central (BMC),
  • ProQuest Health Management

Trial registers such as

For searching systematic reviews databases such as:

  • HTA,
  • DARE,
  • NHS EED,
  • CDSR,
  • Cochrane CENTRAL.

GIN guidelines

Educational database:

  • ERIC

Sources and search strategies for test accuracy information

Inadequate and inconsistent reporting of diagnostic accuracy studies and their indexing in medical reference databases make their identification particularly challenging. Unpublished and ongoing studies of diagnostic accuracy would be valuable but not as easily detected as trials. Reviewers are likely to retrieve thousands of records to scan for potentially relevant studies. Routine use of methodological search terms is not generally recommended because relevant records may be lost with no significant reduction in the number needed to read (Leeflang 2006, Ritchie 2007).

Over 20% of studies included in diagnostic accuracy reviews were not found in MEDLINE and 6 % were not found by the electronic searches (Whiting 2008). The majority of the studies that were not found in databases were identified by scanning reference lists of included articles.

More information on diagnostic search filters and information on their performance can be found at:

Useful other sources
  • Hand searching of journals and abstract books, and the so-called “grey literature” can be performed if information is scarce (Dissertational Abstracts, Scirus - Reports of hospital studies and doctoral thesis, OAIster).
  • Additional information can be collected also from contacts with manufacturers and consultation with domestic and foreign experts and agencies (Handbooks).
  • Performing an additional SCI-search of the included articles is a valuable complementary approach. Add information about other sources and links specific to clinical effectiveness.
  • Other sources: Conference proceedings (Web of Science Database), national and regional guidelines, expert opinions, International, national and regional routinely collected statistics (Health Information Database DRG)
Own research and evidence generation

If the data retrieved from the current body of evidence through a systematic review does not provide enough adequate information on the effectiveness of a technology, new primary research may be warranted in the form of e.g. a randomised controlled trial.

What kind of information is required?

Study types, design, outcome measures

With a bit of luck one may identify a systematic review on the topic of interest, which is sufficiently comprehensive, satisfies the requirements on methodological quality, and meets the research questions. If the report is judged to be transferable to one's own health care system and the local setting, then the work might end right here. Following the hierarchy of study designs (Guyatt 2006), reviews on efficacy / effectiveness are generally limited to randomised designs. To assess the generalisability to routine clinical practice it might be relevant to distinguish between efficacy (explanatory) and effectiveness (pragmatic) RCT. A set of criteria has been suggested to differentiate between them (Gartlehner 2006). In addition registry data reflecting clinical routine care help judging whether study populations, interventions and outcomes in RCT are comparable to clinical practice. It may be necessary to broaden the inclusion to other designs, if data from randomised trials are not available or are insufficient (see Appendix 3).

A number of effect measures are in use for describing the treatment effect. For binary data, common measures are relative effect measures such as risk ratio (= relative risk), odds ratio, and relative risk reduction, or absolute effect measures such as risk difference (= absolute risk reduction), often converted into number needed to treat (NNT) or events per thousand patients to allow for a comparison across studies. Since both relative and absolute effect measures carry important complementary information, recent approaches such as the GRADE profiler {www.gradeworkinggroup.org} encourage a presentation of both measures.

Continuous data are often more difficult to summarize. Commonly used effect measures that allow the summary of treatment effects are “standardised mean difference” or “weighted mean difference”. Unfortunately, both measures are difficult to interpret in a clinical context.

A more recent statistic, the ratio of means, reports the percentage reduction for continuous data such as proteinuria. This measure allows a meaningful interpretation to clinicians (Friedrich 2005). For more details about effect measures and their calculations, we refer to the comprehensive, user-friendly description of common measures in the Cochrane handbook.

If there are different outcome measures for benefits and harms it may be difficult to calculate the net benefit quantitatively. For example in prostate cancer screening the benefit might be a reduction in disease specific mortality, on the other hand, surgery may cause sexual dysfunction and incontinence. Therefore summary measures like the QALY or DALY or other multi-criteria models where health states are weighted according to their desirability could be used to create a common measure (EMA 2010).

Study types for the assessment of the effectiveness of diagnostic and screening technologies*

The following chapters marked with * were provided for the Core Model on diagnostic technologies. As the effectiveness of diagnostic technologies requires evaluation of the whole chain from test accuracy, through change in management to treatment effectiveness, the text is at least partly applicable to the assessment of the effectiveness of screening technologies as well.

Randomised controlled trials (RCTs) are the ideal study design to provide direct evidence of effectiveness of a diagnostic technology. However these studies are rarely available. Furthermore, they are not always feasible or even necessary to determine the effectiveness of the technology. When direct trial evidence is not available other study types, that provide evidence about test safety, accuracy, impact on management and the effectiveness of the treatment, are relevant to the assessment of effectiveness. Evidence from these studies can be linked to yield an estimate of effectiveness of the diagnostic technology (linked evidence). When linking evidence across studies, it is essential to assess whether the patient spectrum in the studies is similar (does the test detect the same disease for which the treatment is effective?).

Direct trial evidence*

The diagnostic RCT is the most reliable study design. The point in the test-treatment chain at which patients are randomized can vary depending on the study question or other constraints, the most simple design randomizing subjects to receive the new test (strategy) or the routine test (strategy) (Lijmer 2009). RCTs measure the difference in health outcomes when patients from the same source population are allocated to different diagnostic pathways. The only difference between groups is due to the selection of the diagnostic pathway and in subsequent treatment decisions. Other comparative study designs like cohort and case-control studies have greater potential for bias.

Linked evidence*

When direct trial evidence on test effectiveness is not available, we need to consider other study types evaluating one or more outcomes in the diagnostic pathway.

Study type

Optimal study design

Safety research

All study designs including case series, surveillance registers

Diagnostic accuracy research

Cohort studies of diagnostic accuracy

Change-in-patient-management studies

Diagnostic before-after studies and time series

Treatment effectiveness studies

Treatment RCTs

Evidence of accuracy can be used to infer effectiveness of the technology when the spectrum of patients, disease, technologies and other conditions are similar enough in diagnostic accuracy and treatment effectiveness studies. The transferability must be reasonably justified. Sometimes evidence from accuracy studies is alone sufficient to infer effectiveness of the technology. This happens when the technology is a cheaper, safer or more accurate replacement for an existing diagnostic strategy.

Change-in-management, or therapeutic-impact, or diagnostic before-after-studies measure how often treatment is started, stopped or modified before and after the incorporation of the new diagnostic technology in the management pathway compared to the management pathway without the new diagnostic technology (Guyatt 1986b). Physicians in change-in-management studies are provided with test results from a new diagnostic technology and the researchers then compare their pre-test management plan to post-test management plan. The study type is usually applied to add-on type technologies.

In replacement-type new technologies we usually assume that the behavioural pattern from test result to management decisions remains unchanged. Especially if there is a well established standard treatment for the condition detected. In other cases, change-in-management studies may be required to demonstrate that the test results are sufficient to alter the clinician's threshold for changing management (MSAC 2005).

Change-in-management studies are required if other factors than the test result, like individual patient characteristics or patient preference, influence treatment decision. They are also valuable when the impact of test information is uncertain, as it is when the test is used to distinguish between multiple differential diagnoseis, or when accuracy studies are conducted in patients with different prevalence or severity of disease than the intended patient population or usual practice.

When there is a trade-off between benefits and harms, e.g. when the better safety of a less invasive but less specific new test needs to be assessed against the harms arising from additional false-positive results, decision analytic modelling can be used. Decision analysis allows also the comparison of the test effectiveness in those with a different prevalence of the disease and of multiple test-and-treat strategies of existing tests in clinical practice where it is unfeasible to directly compare all strategies in clinical trials. In fast developing fields completed clinical trials may not be applicable to current practice standards. Modelling can help to assess the trade-offs of a newer test and could also consider potential shifts in the disease spectrum. Modelling can explicitly account for uncertainty in key parameters and assumptions (Trikalinos 2010). Decision analysis is appropriate when the evidence of test accuracy can be linked to the evidence of treatment effect. If this linkage is uncertain, we need randomised trials. In these situations, trials investigating the effect of treatment in patients who have positive results on the new test and negative results on the old test may be more efficient and more clinically relevant than trials conducted in all patients who are new-test-positive (Bossuyt 2000).

Study types for test accuracy studies*

A systematic review and critical appraisal of existing research literature and other data is the basic method of finding answers to research questions on diagnostic accuracy. Regarding some issues, e.g. when asking "what are the requirements for accuracy in the specific context?" or "what is the optimal threshold value?" published research findings may need to be complemented with expert interviews or own reasoning.

The design of a basic diagnostic accuracy study is that of a group of patients with the suspected target disease undergoes the test (strategy) under consideration (index test) and the best possible test (strategy) to verify the diagnosis (reference standard, gold standard). Positive and negative results from both tests are shown in a 2x2 table or a variation thereof, depending on the number of cut-off points chosen.

If there is no appropriate reference test it is possible to construct a reference diagnosis by using a predefined rule for a set of other tests, consensus among experts, or a statistical model based on actual data (Rutjes 2007). Another possibility is to investigate the probability of disease presence as a function of all diagnostic variables simultaneously with multivariable modelling (Moons 1999). Problems may arise for example from the patient spectrum (patient characteristics, patient selection and setting), the non-optimal reference standard, incorporation bias (the index test is part of the reference standard), partial verification (not all patients receive the reference test) or differential verification (patients receive different reference tests).

If a new technology can replace an existing one, the accuracy of the new test (index test) and the routine test (comparator test) has to be compared in comparable groups or preferably in the same patients (Irwig 2002). This can be done indirectly by looking at studies where test A has been compared with a reference standard, and other studies where test B has been compared with the same reference standard. Studies that do the index test, the comparator test and the reference test to all patients are preferred (paired study). If not all patients had verification with the reference standard test, then the sensitivity and specificity of the two technologies cannot be calculated, but relative true and false positive rates can still be estimated, which allows the accuracy of the two tests to be compared against a common reference standard.

Another option is a randomised controlled trial where patients are randomly allocated to receive either new or existing test, after which all patients undergo the reference standard testing. Randomised trials are preferred if the new test is too invasive to be done to all patients or if the tests interfere with each other (Bossuyt 2006). For further options see Lijmer 2009.

In triage, the new technology is used before the existing technology and only the patient with a particular result of the test continues the diagnostic pathway. Triage technology may be less accurate than the existing ones and are therefore not meant to replace them. Instead, it is simpler or cheaper. If the triage technology can reliably rule out the target condition, it can safely reduce the number of patients who need to be sent further to invasive, cumbersome or expensive testing.

Several designs can be used to compare the accuracy of the triage pathway to the existing pathway. In a paired study design all patient undergo the triage technology, the existing technology and the reference standard. Limited verification can be used here as well, but is a source of bias.

An add-on technology is positioned after the existing diagnostic technology. This is the case when the new technology is more accurate, but too expensive or invasive or poorly available to be used for every patient. The use of the new diagnostic technology may then be reserved for only those patients in whom the existing technologies failed to identify the disease. Add-on technology can increase the sensitivity of the existing diagnostic pathway, usually at the expense of specificity. Or, add-on technology may be used to limit the number of false positives (increase specificity) after the existing pathway.

Fully paired or randomised methods are preferred but not always needed in researching add-on tests. Limited designs can be more efficient. E.g. limiting the study to patients who are negative after existing diagnostic pathway, with verification by reference standard only those who test positive on new technology, allows still us to calculate the number of extra true positives and false positives from using the new add-on technology (Bossuyt 2006).

In screening processes subjects are typically first tested with a triage technology, then with a more accurate test, and sometimes finally with an add-on technology. The various stages need to be evaluated both separately and as an entity.

Outcome measures for test accuracy studies*

Diagnostic test results are often reported as a numeric quantity on a continuous scale which is then divided by a threshold value above which the test is positive and below which it is negative. Results may then be summarized in a 2x2 table to reflect the agreement between the "true" disease state and the test result.

 DiseasedNo disease
Test positiveTPFP
Test negativeFNTN

Figure 2x2 table

The numbers in the table state the number of true-positive, false-positive, true-negative and false-negative results. Changing the threshold, changes these figures and thus the sensitivities and specificities and other summary measures calculated out of the numbers in the 2x2 table.

Measures of test performance (Tatsioni 2005)

Metric

Definition

Advantages

Disadvantages

Accuracy

(TP+TN)/N

Intuitive

Depends on prevalence

Sensitivity

TP/(TP+FN)

Does not depend on prevalence

 

Specificity

TN/(TN+FP)

Does not depend on prevalence

 

Positive predictive value

TP/(TP+FP)

Clinical relevance

Depends on prevalence

Negative predictive value

TN/(TN+FN)

Clinical relevance

Depends on prevalence

Positive likelihood ratio

 

Does not depend on prevalence

Applies only to positive test

Negative likelihood ratio

 

Does not depend on prevalence

Applies only to negative test

Diagnostic Odds ratio

TP x TN/FN x FP    =Lr+/Lr-

Does not depend on prevalence; combines sensitivity and specificity. Invariant to test positivity threshold.

Values FP and FN errors equally; not intuitive

Area under curve

Area under ROC curve

Does not depend on prevalence; combines sensitivity and specificity

Lack of clinical interpretation

TP = true-positive, TN = true-negative, FP = false-positive, FN = false-negative, N = sample size, ROC = receiver-operating-characteristic

Primary measures of diagnostic accuracy are sensitivity and specificity. They are always considered together as a combined measure of accuracy. They are not directly influenced by the prevalence of the disease and thus the results from one study may be applicable to different populations. Paired data with 95 % confidence intervals can be graphically presented and pooled.

Sensitivity and specificity depend highly on the test threshold. Increasing the threshold increases the specificity but decreases sensitivity. The inverse relationship between sensitivity and specificity is often best illustrated using a graph (ROC curve) where pairs of sensitivity and specificity are plotted for different thresholds.

There are explicit thresholds like laboratory test values, although different laboratory kits provide numbers that are not necessarily comparable. Then there are implicit differences in threshold caused by case-mix and factors affecting test reading. Especially in imaging tests it is the eye of the reader that determines test positivity, and different readers may draw different conclusions on test positivity.

A likelihood ratio (LR) describes how many times a person with a disease is more likely to receive a particular test result than a person without disease. It can be calculated for all different levels of the test result. It is therefore useful measure of test accuracy when test results can be reported in more than two categories. It can be combined with the estimated prevalence of the disease to calculate the post test probability of the disease. It can be treated as a risk ratio for data pooling and presented graphically with 95% confidence intervals (CI) in systematic reviews. Data can be pooled only if there is no variability in the test threshold used (MSAC 2005).

A diagnostic odds ratio (DOR=Lr+/Lr-) provides a single summary estimate of test accuracy that combines sensitivity and specificity. It does not usually vary by the test threshold and is not dependent on the prevalence of the disorder (although it may vary with disease severity). It can be used for indirect comparisons between two tests. It can be calculated with 95% CI and presented in a forest plot. DOR from different studies can be pooled to calculate a summary DOR using standard meta-analytic methods, if no heterogeneity is present. Every single point in a symmetric (symmetry around the diagonal where sensitivity = specificity) ROC curve has the same DOR. An important disadvantage is that DOR as a single number leaves out information on sensitivity and specificity (the same DOR could result from tests with very different sensitivities or specificities). Furthermore, it cannot be used to summarise multi-level test results.

A ROC curve demonstrates the trade-offs between the sensitivity and specificity of the test. A horizontal line would mean constant sensitivity, vertical line constant specificity. Constant likelihood ratio is seen as linear relationship of sensitivity and specificity. A diagonal line from lower left to upper right corner would mean that the test is not informative at all. Usually there is a curvilinear relationship with the plots. The point in the curve that is closest to the upper left corner gives the test threshold with best accuracy.

If the distribution of possible test values in healthy and sick persons is different, e.g. the distribution of PSA-measures in healthy is quite narrow and in prostate cancer patients broad, then the ROC curve becomes asymmetric and high and low DORs occur in different parts of the curve.

The area under the ROC curve (AUC) provides a measure of the overall accuracy of the test. AUC can be interpreted as the probability of correctly identifying the disease on a pair of subjects, when one of them has the disease and the other has not. Values for AUC can range from 0 to 1. If the sensitivity and specificity of the test is 100% at each threshold, then AUC is 1.0 and the test is perfect. If AUC is 0.5, the test does not discriminate between the presence and absence of the disease. And, if it is below 0.5 the test is misleading because it systematically misclassifies diseased and healthy people, but by a swap of the classification of diseased and healthy it would discriminate better than chance (AUC>0.5). From AUC data alone it is not possible to derive false positive and false negative rates. Because the consequences of false positive and false negative test results may be weighted differently in clinical practice a summary measure like the AUC might be misleading.

Study types for the assessment of the effectiveness of screening technologies

The most reliable evidence whether screening does more good than harm are well conducted RCTs with a study population representative of those eligible for, and invite to or informed of the screening programme. The control group would those who are not informed of the screening programme. Otherwise the probability of a cross–over of the control group to screening group would increase and this could result in an underestimation of the screening effect. Time trend studies which analyse changes in disease frequency such as incidence, the distribution of different severity of disease stages and death can be valuable. But there are many sources of bias such as changes in ascertainment and diagnostic practice or other influences on outcomes such as advances in treatment, or reduction in co-morbidities. Case-control studies can be useful for a comparison of different screening policies but cannot give a reliable estimate of the difference between screening and no screening because their confounding factors can not be controlled (Raffle 2007).

Modelling studies are especially useful in comparing many different screening options varying in test combinations, screening intervals and treatment options incorporating alternative eligible populations, whereas clinical trials can compare only a limited number of screening options over a short time horizon. When high quality primary data is available, decision analytic modelling can synthesize information from a wide range of sources. Sensitivity analysis can help to show areas in which further research is likely to be most useful (Karnon 2007, Trikalinos 2009)

Tools for critical appraisal

Sources of bias in studies designed to evaluate the effectiveness of an intervention, or diagnostic test and subsequent interventions,canrelate to differences in patients assigned to intervention and control group, including differences in the selection process (selection bias); the unbalanced provision of care (performance bias); the methods of measuring or interpreting the outcomes (detection bias); or imbalances in patient drop-out (attrition bias) (Moher 1998, Schulz 1995).

A thorough assessment of the methodological quality of the included studies is crucial to any systematic review. In randomised controlled trials, concealed treatment allocation, blinding of health care provider, patient and outcome assessor to the allocated intervention (experimental or control) and a sufficient rate of follow-up are the minimum items that need to be looked at when assessing the potential for bias of individual studies. Depending on the research question, however, it might be warranted to look at additional features where bias could enter the study design, or where the results might get distorted.

Quality assessment of single studies

The body of checklists for assessing the methodological quality of randomised controlled trials is considerable, most of them are variations (e.g. vanTulder 2003) of the structure suggested in the User’s Guides to the Medical Literature (Guyatt 2007), the CONSORT Statement (Altman 2001, Moher 2001, Rennie 1996, Schulz 2010) or the criteria suggested in the Cochrane Handbook.

Agreement on the methodological criteria for non-randomised trials and observational studies are considerably less well developed. However, a methodological HTA-report by John Deeks provides a good overview of available instruments to assess non-randomised intervention studies (Deeks 2003, MacMahon 2001, Radford 2001, Stroup 2000, Equator web site).

Modelling studies

The validity of the results of modelling studies are highly dependent on the model structure, the model assumptions, the validity of the data used as inputs to models, and model validation. There are several checklists for quality assessment for modelling studies available (Weinstein 2003, Philips 2006, Karnon 2007).

Overall quality of the whole body of evidence

Having reviewed the methodological quality of the individual studies, researchers attempt to capture the overall quality of the body of evidence. The concept of the GRADE Working Group captures the currently most comprehensive approach (Atkins 2004, Guyatt 2006). Besides looking at the quality of the individual studies, they also include the consistency or heterogeneity of the results of all included studies and the directness of the comparisons (i.e. how directly does the identified literature address the questions of our HTA-report regarding the population, the intervention and comparators and the selected endpoints, they comment on imprecision of the available data (number of total events and width of the confidence interval) and provide an estimate about the likelihood of the presence of reporting bias. Deficiencies in any of those considerations can lower the methodological quality of the entire body of evidence. On the other hand, the overall judgement about the methodological quality of the evidence can be raised in the presence of strong and plausible associations between intervention and outcome or an obvious dose-response gradient.

Quality assessment of the effectiveness of diagnostic technologies*

Direct trial evidence

A diagnostic technology may appear to be effective because of a careless or incomplete pre-test work-up. This occurs when the technology becomes an alternative to careful history, physical examination, and a set of less invasive or less expensive procedures. Therefore it is worthwhile to carefully consider the pre-test examination scheme in the studies.

Linked evidence

The strengths and limitations of other study types than RCT need to be considered. There are quality check lists for studies of effectiveness in MSAC page 69 (MSAC 2005 page 69).

Change-in-patient-management studies can be appraised using the same criteria as case series (see list of criteria MSAC page 70) (MSAC 2005). Potential bias is common and it is related to the selection of patients, the objective execution of the diagnostic test, and measurement of the results in all eligible patients. One of their limitations is that stated plans may differ in the study setting compared to real life situations where the technology is not available. Physicians' subconscious bias may also occur. Change of management is only relevant when it results in a benefit in patient relevant outcomes. Otherwise it can be held only as an surrogate end-point.

Quality assessment of test accuracy studies*

Quality assessment of diagnostic accuracy studies is not as straightforward as it is for interventions. It is hampered by poor reporting and the fact that so far there is less methodological and empirical evidence on the importance of the different potential sources of bias. There are many different tools to assess the quality of diagnostic accuracy studies. The Cochrane handbook recommends QUADAS tool with its 11 mandatory and additional items.

QUADAS quality assessment tool (Whiting 2003), QUADAS 2 is in development

Mandatory items (as in Cochrane handbook)

  1. Was the spectrum of patients representative of the patients who will receive the test in practice?
  2. Is the reference test likely to correctly classify the target condition?
  3. Is the time period between reference test and index test short enough to be reasonably sure that the target condition did not change between the two tests?
  4. Did the whole sample, or random selection of the sample, receive verification using a reference standard of diagnosis (reference test)?
  5. Did patients receive the same reference test regardless of the index test result?
  6. Was the reference test independent of the index test i.e. the index test did not form part of the reference test?
  7. Were the index test results interpreted without knowledge of the results of the reference test?
  8. Were the reference test results interpreted without knowledge of the results of the index test?
  9. Were the same clinical data available when the test results were interpreted as would be available when the test is used in practice?
  10. Were uninterpretable / intermediate test results reported?
  11. Were withdrawals from the study explained?

Additional items

  1. If a cut-off value has been used, was it established before the study was started (pre-specified cut-off value)?
  2. Is the technology of the index test likely to have changed since the study was carried out?
  3. Did the study provide a clear definition of what was considered to be a "positive" test result?
  4. Was treatment started after the index test was carried out but before the reference test was performed?
  5. Was treatment started after the reference test was carried out but before the index test was performed?
  6. Were data on observer variation reported?
  7. Were data on instrument variation reported?
  8. Were data presented for appropriate patient sub-groups?
  9. Was an appropriate sample size included?
  10. Were objectives pre-specified?

HTA-authors can adapt QUADAS by dropping irrelevant items. Two assessors are recommended. Background of assessors should be reported, and the way they resolved disagreements. Results of the quality assessment of the original studies should be presented in a table or graphically. Individual quality items should be investigated as a potential source of heterogeneity.

Issues specific for screening technologies

There are three main sources of bias which are specific to the evaluation of screening:

–        People taking part in screening are usually healthier than those who do not (healthy screenee bias.

–        Less aggressive cases of disease have a longer asymptomatic period and are therefore more likely to be detected by screening. Consequently patients detected by a screening programme tend to have a better prognosis even without therapy (length time bias).

–        Survival falsly appears to be longer after diagnosis by screening not because the patients actually live longer but because the diagnosis is known earlier and therefore for a longer period of time (lead-time bias) (Raffle 2007, Gates 2001.)

If a high proportion of participants in the control group (no screening) cross over to screening the effects of screening will be underestimated.

Analyzing and synthesizing information

Ideally systematic reviews on randomized controlled trials (RCTs) are the basis of knowledge of effectiveness of an intervention (Khan 2002). The principles on how to conduct a systematic review are nowadays widely agreed upon and most of the methodologies published by different organisations vary only in details (See Appendix 3).

Evidence Tables

A meaningful presentation of the study results is essential for an informative and transparent HTA report. A high degree of reliability and transparency are reqired for the transfer of HTA reports from one setting to another. Comprehensive and informative evidence tables about the methodology and content of the individual studies are the best guarantor for transparency and reliability. They should allow a judgement of the similarities and differences of the included studies and should provide the basis for the conclusions of the review.

The majority of HTA organisations produce tabulated evidence summaries that follow the PICO structure (ideally with an additional cell for comments on issues not captured by the PICO cells but that could have an impact on the results). Although the items reported in each cell will always be driven by the questions of the review, they should follow some core considerations (Malmivaara 2006). A description of the data extraction process including the number of reviewers involved assures objectivity and reliability of the results.

Meta-analysis

Studies on the same topic can report their findings in very different ways which hinders meaningful comparisons across studies and a fair and appropriate interpretation of the body of evidence. Reviewers are encouraged to convert (re-calculate) the results to a joint effect measure and attempt a meta-analysis when the data allow a summary of the results. However, sufficient clinical homogeneity of the studies is a prerequisite for a meta-analysis.

Although the nature of the data can prevent pooling for a summary estimate and researchers can provide only a descriptive summary of the data, it can nevertheless be very helpful to display the results in a forest plot, but omitting the summary.

Presenting a measure of precision for the estimate of the treatment effect (confidence interval) is needed for the interpretation of the data and must not be omitted. Researchers need to report if the primary studies lack this essential information.

Further exploration of the data: Homogeneity and heterogeneity, sensitivity analysis and publication bias

Reviewers need to provide statements about clinical homogeneity or heterogeneity of the studies and their results. While homo-/ heterogeneity in the clinical data is often a matter of judgement, there are statistical tests available to help assessing the presence of statistical heterogeneity (Higgins 2003) which should then be further explored and considered in the discussion. Pre-specified sensitivity analyses based on clinical or methodological issues allow further exploration of the stability of the data. Researchers should always consider publication and reporting bias and explore these either graphically using a funnel plot (provided the number of included studies is large enough) or make a plausible judgement about the likelihood of these biases.

Data extraction from test accuracy studies*

Included studies table columns

  • Participants, prevalence of target condition
  • Prior tests
  • Index test, cut-off point
  • Reference test
  • Test results (2x2 data)
  • Sensitivity/specificity + 95% CI
  • Other accuracy metrics
  • Study quality
Pooling and meta-analyzing test accuracy studies*

No heterogeneity present

A forest plot of sensitivity versus specificity with 95 % confidence intervals can be used whenever the results from two or more comparable studies are included in the review. The forest plot illustrates the range of results, enables the reader to assess heterogeneity, and possible trade-off between sensitivity and specificity, and may show the summary estimate where pooling is appropriate.

Another option is to plot pairs of sensitivity and 1 - specificity from original studies on a ROC plane. If sensitivity or specificity is constant or if there is linear relationship between them, simple summary measures for sensitivity, specificity, or likelihood are adequate.

When pooling pairs of sensitivity and specificity, the statistical model used depends on the studies selected. A fixed effect model assumes the studies to represent a random sample of one large common study. The differences between study outcomes are considered to be the result of random error. The model weights individual studies based on the inverse variance of the accuracy or the number of participants. Random effects model assumes the differences between studies to be due to real differences between the study populations and procedures. A more complex mathematical model is used to weight studies. Separate estimates of mean sensitivity and specificity underestimate test accuracy.

Heterogeneity present

When forest plot or heterogeneity testing shows that there is significant heterogeneity in sensitivities and specificities across studies, it is not appropriate to report pooled values of sensitivity and specificity as a summary estimate. Instead, further analysis of the heterogeneity detected is needed, and it starts with examining of threshold effect. Threshold effect can be seen in forest plot if there is an inverse relationship between sensitivity and specificity. If this is not apparent the results should be plotted to a ROC plane to examine the data further.

Threshold effect only

If there is symmetry in the SROC curve, DOR is constant regardless of the diagnostic threshold, and any variability in the paired sensitivity and specificity between different studies is due to differences in the test threshold. In this case, SROC curve represents the most informative synthesis of evidence about test accuracy and the pooled DOR is a useful single summary measure.

SROC curve does not provide one summary estimate of sensitivity and specificity but it allows assessment of their interdependence. Summary DOR (SDOR) of the test and a comparator test can be presented with 95 % CIs to compare differences in diagnostic performance. The area under SROC curve and its 95% confidence interval provides a global summary of overall test accuracy. The point on the curve where sensitivity equals specificity, the Q* statistics, can also be used as a summary measure of the accuracy of the test. These summary measures can also be used to compare the accuracy of two test strategies. Software for diagnostic meta-analysis include Meta-Test, Meta-Disc, Stata and SAS.

Heterogeneity that is more than just threshold effect

If the slope b(the estimated regression coefficient) in the SROC model is statistically significant, the SROC will be asymmetrical and the DOR changes along the threshold. In such cases advanced methods for fitting the SROC is used. Advanced methods to pool are indicated if heterogeneity in the results can be attributed to known sources of variation (see above Chapter Assessing heterogeneity). Otherwise the interpretation of the summary estimate is not possible (Lijmer 2002).

Advanced models enable incorporation of covariates, e.g. population subgroup in the meta-regression analysis. Poor reporting of primary studies may though lead to biased estimates. The two main advanced models are hierarchical SROC and bivariate meta-regression, which are mathematically identical (Harbord 2007). Syntax to run these models in SAS, STATA, WINBUGS, S-PLUS and R is  or will be available. Hierarchical SROC (HSROC) produces informative summary measures with confidence ellipses (Reitsma 2005).  Model is infrequently used, probably due to the complex fitting.

More reading: Deeks 2001, Deville 2002, Kester 2000, Irwig 1995

The problem of imperfect reference standard in test accuracy studies*

If there is an acceptable reference standard test but for various reasons not all patients in the study received it, the researches either impute or adjust for the missing data (Rutjes 2007). If the fraction of patients verified with the reference standard is small, or if the patterns of replacing the missing values are not determined in the study design, the authors of a Core HTA should be careful with the results.

Sometimes the reference standard is known to be imperfect: i.e. it does not distinguish the diseased from healthy quite correctly. Then it is possible that the researchers have adjusted the estimates of accuracy of the index test (Rutjes 2007). These correction methods can be useful if there is evidence from previous studies about the extent of imperfection of the reference standard and about the correlation of the errors between the index test and the reference standard. Another way to deal with the problem of imperfect reference standard is a sensitivity analysis to demonstrate the effect of imperfect reference test to the accuracy of the index test.

Assessing heterogeneity across test accuracy studies*

Heterogeneity in test accuracy across studies is very common. Any differences in the results of studies that address the same research question should be clearly identified and interpreted in the diagnostic Core-HTA report. Simple methods of pooling sensitivities and specificities are contraindicated if heterogeneity exists.

Sources of heterogeneity are

  1. Chance
  2. Different test threshold
  3. Different study designs, methods, biases: different reference standard, different versions of the technology
  4. Variation by clinical subgroups in terms of age, severity or stage of disease, prevalence of the target condition, differential diagnoses, and setting
  5. Unexplained heterogeneity

If differences in the results can not be attributed to these known sources of heterogeneity, then pooling of results to one summary estimate should not be attempted, because its interpretation will be impossible (Lijmer 2002).

Methods to test for heterogeneity (MSAC 2005):

  1. Plot the sensitivity and specificity from each study with their 95% confidence interval in a table and/or forest plot to illustrate the range of estimates and identify outliers.
  2. If sufficient data are available, plot the paired sensitivity and 1-specificity results for each study on the ROC plane to detect heterogeneity and identify outliers. A small number of studies will limit the power of regression to detect heterogeneity.
  3. Use a chi-square test for heterogeneity (Cochran's Q test) or Fischer's exact test for small studies to test the hypothesis that there is no statistically significant difference in the sensitivity and specificity reported.
Assessing threshold effect in test accuracy studies*

Paired estimates of sensitivity and 1 - specificity in original studies are plotted in a ROC plane. Regression model is used to fit the SROC curve (Moses 1993). If the SROC curve is symmetrical around the line where sensitivity equals specificity, the studies share one common DOR, and any variability is due to differences in the test threshold. In statistical terms, if in the model the slope b (estimated regression coefficient) is not statistically significant and approaches zero, The SROC will be symmetrical.

Issues specific for screening technologies

For diagnostic and treatment interventions in patients already showing symptoms or being ill there is a trade-off between benefits and harms of diagnostics and treatment on the individual level. Because screening is usually done in asymptomatic people there is an additional trade-off on the population level between healthy people who will not benefit from screening but can be harmed by a loss in quality of life by false positive screening results, potential over-diagnosis and over-treatmen,t and people who will benefit by an early detection of the disease. Decision analytical modelling is an explicit and quantitative method which can be used to analyse these trade-offs.

The accuracy of the screening/ diagnostic test can be highly dependant on the competence (qualifications, training and experience) of the staff/personnel using the device and analysing the test results.

Reporting and interpreting

Which steps are required?

  • Rating the body of evidence as being of high / moderate / low quality (following the GRADE methodology) clarifying (e.g. in footnotes) the reasons for up-/downgrading. 
  • Interpreting the clinical relevance of the findings:
    • Considering the importance of the outcomes for clinical decision making (distinguishing between a critical and an important outcome as done when formulating the question)
      • Deciding what would be the minimal clinically important effect size for each outcome (independent of its statistical significance):
      • Identifying knowledge gaps by comparing the research questions (including the predefined outcome) with the available evidence.

It is possible to make only a preliminary interpretation of the results based on effectiveness data only. A global and balanced interpretation of the benefits and harms of a technology requires also the results of other relevant domains. Evidence about benefits and harms can be combined using e.g. decision analytic methods (Trikalinos 2009).

Interpreting and reporting test accuracy studies*

Pair of sensitivity and specificity is a general measure of test performance. The numbers (0.0–1.0)  per se are not very informative in determining whether the test performs well. The intended use of the technology determines the requirements for the test accuracy. If sensitivity is sufficiently high, a negative test result rules out the disease. High sensitivity is particularly important if the penalty for missing a disease is high. Sufficiently high specificity rules in the disease. High specificity is particularly important if a false positive result can harm the patient. Positive and negative predictive values are clinically informative measures of the accuracy of a diagnostic test, but must be considered in relation to the prevalence of the disease.  

Summary likelihood ratios can be estimated from the pooled estimates of sensitivity and specificity. Likelihood ratio tells how many times more likely the disease is in patients with that test result compared to those without the disease. A likelihood ratio 1 indicates that the test does not provide any useful diagnostic information. Positive likelihood ratios more than 10 and negative likelihood ratios less than 0.1 can provide convincing diagnostic information. Some guidelines suggest that positive likelihood ratios more than 5, and negative likelihood ratios less than 0.2 can provide strong diagnostic evidence. However, the interpretation depends on the context and prevalence of the condition. Likelihood ratios usually have to be more than10 for a test to be useful (MSAC 2005), although this is very seldom the case.

Diagnostic odds ratio shows the association between a dichotomous test result and the diagnosis. If the diagnostic odds ratio (DOR) is 1 then the test does not provide any useful information. The size of the DOR greater than 1 reflects the strength of the test to discriminate between the presence and absence of disease. A DOR of 100 provides convincing evidence of the presence or absence of disease and correspond to a positive likelihood ratio of 10 and a negative LR of 0.1.  It is often 50-90 but can be even thousand, and it should be over 80 in a good test. A DOR less than 1 indicates that the test identifies more positives among the non diseased than the diseased. Diagnostic odds ratio is useful summary measure for meta-analysis but it does not provide information that can be directly applied to clinical decisions. (MSAC 2005).

Variation in results by cut-off points, prevalence or any other covariate and characteristics of the SROC curve should be explained. Area under SROC curve can be used to compare accuracy of two test strategies. The test whose SROC curve encloses the largest area is the most accurate.

Additional methods of expressing test accuracy beyond sensitivity and specificity, e.g. likelihood ratios or diagnostic odds ratios, are preferred. Explaining how many patients will be missed (false negative rate) and how many treated unnecessarily (false positive rate) using certain cut-off point in a population with certain disease prevalence, may be illustrative.

Assessment elements

D0001 Assessment element card

Issue: What is the effect of the intervention on overall mortality?

Topic: Mortality

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In screening the technology is seen as the combination of screening test, subsequent diagnostic work-up and treatment.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Systematic reviews of RCTs (Randomised controlled trials) or CTs (controlled trials); if not available RCTs or CTs itself. If these not available, non-controlled studies and respective systematic reviews. Health care register data. Modelling studies.

References
Content relations
Sequential relations

D0002 Assessment element card

Issue: What is the effect of the intervention on the mortality caused by the target disease?

Topic: Mortality

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

A screening test can lead to an earlier diagnosis, thus earlier treatment which might reduce the mortality.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Systematic reviews of RCTs (Randomised controlled trials) or CTs (controlled trials), if not available RCTs or CTs itself. If these not available, non-controlled studies and respective systematic reviews. Health care register data. Modelling studies.

References
Content relations
Sequential relations

D0003 Assessment element card

Issue: What is the effect of the intervention on the mortality due to other causes than the target disease?

Topic: Mortality

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This may be due to e.g. side effects, accidents, or consequences of interventions after false positive or incidental findings.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Systematic reviews of RCTs (Randomised controlled trials) or CTs (controlled trials), if not available RCTs or CTs itself. If these not available, non-controlled studies and respective systematic reviews. Health care register data. Modelling studies.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0001, C0006

Sequential relations

D0004 Assessment element card

Issue: What is the mortality related to the diagnostic test?

Topic: Mortality

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In diagnostic and screening technologies it is worthwhile distinguishing the possible mortality risk of the test itself from the mortality outcomes of the whole diagnostic or screening process (D0001-D0003). Inappropriate use of the technology or errors may contribute to this issue.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Observational research, RCTs, safety monitoring databases, registers, statistics

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0001

Sequential relations

D0005 Assessment element card

Issue: How does the use of the technology modify the symptoms and findings of the target condition?

Topic: Morbidity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Severity, frequency and recurrence of symptoms and findings.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D0006 Assessment element card

Issue: How does the technology modify the progression of the target condition?

Topic: Morbidity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

E.g. complete cure, alleviation, delay of the onset of the next stage of the disease.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials, prognostic studies

References
Content relations
Sequential relations

D0026 Assessment element card

Issue: How does the technology modify the effectiveness of subsequent interventions?

Topic: Morbidity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Different tests may detect slightly different subpopulations as test positive. Results from further diagnostic testing and the effectiveness of subsequent interventions can be different in test A positive compared to test B positive. E.g. treatment may work differently in screening-identified cases than in cases that are diagnosed at regular physician's appointment.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials, observational studies, accuracy studies

References
Content relations
Sequential relations

D0008 Assessment element card

Issue: What is the morbidity directly related to the technology?

Topic: Morbidity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In diagnostic and screening technologies it is worthwhile distinguishing the possible morbidity caused by the test itself from the morbidity outcomes of the whole diagnostic or screening process (D0005-D0006). Inappropriate use of the technology or errors may contribute to this issue.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials reporting adverse events.. Observational studies. Registries

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0003 to C0005

Sequential relations

D0020 Assessment element card

Issue: Does use of the test lead to improved detection of the condition?

Topic: Change-in management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Although the test is reliable, the information it provides does not necessarily affect clinical decision making. If it does not change sufficiently the pre-test probability the added value of the information may be low. E.g there may be routine preoperative lab tests that nobody uses in decision making. Moreover, users´ ability to make a correct diagnosis may depend on their knowledge and ability to interpret the results.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, accuracy studies, before-after studies, interrupted time series, change-in management studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational domain

Sequential relations

D0021 Assessment element card

Issue: How does the use of the test change physicians' management decisions?

Topic: Change-in management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

There may be technology-related or non-related factors that might influence the physicians' perceptions, ability and attitude to decision making. Management decisions mean both testing and treatment decisions.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Change-in-management studies, qualitative research

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational domain

Sequential relations

D0024 Assessment element card

Issue: Is there an effective treatment for the condition the test is detecting?

Topic: Change-in management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical domain

Sequential relations

D0022 Assessment element card

Issue: Does the test detect other potential health conditions that can impact the subsequent management decisions?

Topic: Change-in management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Management decisions mean both testing and treatment decisions.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials, Descriptive literature

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0006

Sequential relations

D0023 Assessment element card

Issue: How does the technology modify the need for other technologies and use of resources?

Topic: Change-in management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Some treatments require ongoing monitoring and healthcare visits including hospitalisation. Screening tests may cause further diagnostic testing and different treatment due to detection of disease at an earlier stage.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies, statistics

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Costs, organisational aspects domain

Sequential relations

D0011 Assessment element card

Issue: What is the effect of the intervention on global function?

Topic: Function

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D0014 Assessment element card

Issue: What is the effect of the technology on return to work?

Topic: Function

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Sick leave, retirement and various outcomes describing working ability are relevant outcomes to this issue.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Trials and other studies with return-to-work or work ability outcomes reported.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social and costs domain

Sequential relations

D0015 Assessment element card

Issue: What is the effect of the technology on return to previous living conditions?

Topic: Function

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Testing may affect the ability to return to previous living conditions. It may have implications for family members / carers too.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D0016 Assessment element card

Issue: How does use of the technology affect activities of daily living?

Topic: Function

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D0012 Assessment element card

Issue: What is the effect of the technology on generic health-related quality of life?

Topic: Quality of life

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Costs, social domain

Sequential relations

D0013 Assessment element card

Issue: What is the effect of the technology on disease specific quality of life?

Topic: Quality of life

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Costs domain

Sequential relations

D0030 Assessment element card

Issue: Does the knowledge of the test result affect the patient's non-health-related quality of life?

Topic: Quality of life

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It can improve or worsen the quality of life. Test result may alleviate symptoms although there is no effectiveness to the primary outcome. It can also trigger or worsen symptoms.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Qualitative research, observational studies, trials

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social and ethical domain

Sequential relations

D0017 Assessment element card

Issue: Was the use of the technology worthwhile?

Topic: Patient satisfaction

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Patients overall assessment of the worthiness of the intervention.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Qualitative research, observational studies, trials

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D0018 Assessment element card

Issue: Is the patient willing to use the technology?

Topic: Patient satisfaction

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Differences in acceptability may predict the overall uptake of the technology and would impact on the overall effectiveness.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Qualitative research, observational studies, trials

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social domain

Sequential relations

D1001 Assessment element card

Issue: What is the accuracy of the test against reference standard?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy in terms of sensitivity and specificity, and other measures such as likelihood ratios, pre-test probabilities, SDORs, AUC or Q*? In screening programmes one should consider separately the accuracy of the screening test and the accuracy of subsequent diagnostic tests.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies

References
Content relations
Sequential relations

D1003 Assessment element card

Issue: What is the reference standard and how likely does it classify the target condition correctly?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies

References
Content relations
Sequential relations

D1004 Assessment element card

Issue: What are the requirements for accuracy in the context the technology will be used?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Acceptable number of false negative and false positive test results is different e.g. in replacement/ triage/ add-on situations, and in life threatening / harmless conditions. In screening programs one should consider separately the screening test and the subsequent diagnostic tests.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Descriptive literature, expert advice, prevalence data, modelling studies, calculations

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical aspects domain

Sequential relations

D1005 Assessment element card

Issue: What is the optimal threshold value in this context?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Sensitivity and specificity vary according to the threshold value. Optimal combination of sensitivity and specificity defines optimal threshold value. The optimum depends on the consequences of the test results. E.g. whether it does more harm to overlook a case or to treat someone unnecessarily. In screening programs one should consider separately the screening test and the subsequent diagnostic tests.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Screening studies with varying thresholds, accuracy studies with varying thresholds, modelling studies

References
Content relations
Sequential relations

D1006 Assessment element card

Issue: Does the test reliably rule in or rule out the target condition?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

When assessing screening programs one should consider here the combination of the screening test and the subsequent diagnostic tests.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies, modelling studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety, social, ethical domains

Sequential relations

D1007 Assessment element card

Issue: How does test accuracy vary in different settings?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

How do patient spectrum, disease prevalence, disease severity, and properties of the technology itself affect the accuracy of the test? This may have implications on how frequently a test needs to be repeated, optimal age range for a screening programme and adjustments in different populations.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies in different settings, descriptive literature, expert advice

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0004, B0016, B0005, Organisational domain

Sequential relations

D1002 Assessment element card

Issue: How does the test compare to other optional tests in terms of accuracy measures?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Or, how does the technology compare to other development stages of the same technology?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies

References
Content relations
Sequential relations

D1008 Assessment element card

Issue: What is known about the intra- and inter-observer variation in test interpretation?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This is especially relevant in tests with subjective assessments, such as most imaging tests.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies, trials, observational studies

References
Content relations
Sequential relations

D1019 Assessment element card

Issue: Is there evidence that the replacing test is more specific or safer than the old one?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

If there is effective treatment for a condition, then a new diagnostic technology with similar sensitivity but greater safety or specificity may be seen as improved effectiveness. In screening programs one should consider separately the screening test and the subsequent diagnostic test.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Accuracy studies, trials, observational studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety domain

Sequential relations

D0027 Assessment element card

Issue: What are the negative consequences of further testing and delayed treatment in patients with false negative test result?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In screening programmes one should consider separately the false negative screening test results and the subsequent false negative diagnostic test results.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Observational studies, trials, qualitative research

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety domain

Sequential relations

D0028 Assessment element card

Issue: What are the negative consequences of further testing and treatments in patients with false positive test result?

Topic: Test accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In screening programs one should consider separately the false positive screening test results and the subsequent false positive diagnostic test results.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Observational studies, trials, qualitative research

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

C0006, Organizational, costs and ethical domains

Sequential relations

D0029 Assessment element card

Issue: What are the overall benefits and harms of the technology in health outcomes?

Topic: Benefit-harm balance

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This question integrates all benefits and harms concerning mortality, morbidity, QoL and further patient relevant outcomes, also considering the amount of false positive and false negative test results. It is the central question about clinical effectiveness. There is no common quantitative summary measure, and even qualitatively a balanced and meaningful presentation is difficult to reach. In diagnostic technologies one should consider also the benefits and harms of subsequent diagnostic testing and treatments in patients with true positive test result in a prior diagnostic or screening test. For true positive cases there is a benefit-harm balance, because diagnostics and treatment can harm. Consequences for true negative cases are identical with the possible harms of the screening test (see D0004, D0008).The integration of some assessment elements of other domains into the benefit-harm-balance is essential and differs between the core model applications. For screening the frequency of disease and coverage of screening are essential AES

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

RCT, CT, observational studies, modelling studies

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0007, A0011, C0001, C0003, C0004, C0005, C0006, C0007, C0061, Ethical Domain

Sequential relations

References

Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, Elbourne D et al. The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med 2001;134:663-94.

Andermann A, Blancquaert I, Beauchamp S, Déry V. Revisiting Wilson and Jungner in the genomic age: a review of screening criteria over the past 40 years. Bull World Health Organ. 2008 Apr;86(4):317-9.

Ashcroft R; What is clinical effectiveness? Stud. Hist. Phil. Biol. & Biomed. Sci. 2002, 33; 219-233.

Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S et al. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490.

Bossuyt PM, Lijmer JG, Mol BW. Randomised comparisons of medical tests: sometimes invalid, not always efficient. The Lancet, 2000 11/25;356(9244):1844-1847.

Bossuyt PM, Irwig L, Craig J, Glasziou P. Comparative accuracy: assessing new tests against existing diagnostic pathways. BMJ 2006 May 6;332(7549):1089-1092.

Centre for Reviews and Dissemination. Systematic reviews. CRD’s guidance for undertaking reviews in health care. University of York 2008

Cochrane Handbook for Systematic Reviews of Interventions REF TO BE COMPLETED

Deeks JJ, Dinnes J, D'Amico R, Sowden AJ, Sakarovitch C, Song F et al. Evaluating non-randomised intervention studies. Health Technol Assess 2003;7:1-186.

Equator network: reporting guidelines. http://www.equator-network.org/resource-centre/library-of-health-research-reporting/reporting-guidelines/

European Medicines Agency. Benefit-risk methodology project.Work package 2 report: Applicability of current tools and processes for regulatory benefit-risk assessment. 2010 www.ema.europa.eu

Friedrich JO, Adhikari N, Herridge MS, Beyene J. Meta-analysis: low-dose dopamine increases urine output but does not prevent renal dysfunction or death. Ann.Intern.Med 2005;142:510-24.

Gartlehner G, Hansen R, Nissman D, Lohr K, Carey T. Agency for Healthcare Research and Quality. Criteria for distinguishing effectiveness from efficacy trials in systematic reviews. 2006 www.ahrq.gov/downloads/pub/evidence/pdf/efftrials/efftrials.pdf

Gates TJ, Cancer screening in perspective. Am Fam Physican 2001; 63(6):1101-12

Guyatt GH, Tugwell PX, Feeny DH, Haynes RB, Drummond M. A framework for clinical evaluation of diagnostic technologies. CMAJ 1986 Mar 15;134(6):587-594.

Guyatt GH, Tugwell PX, Feeny DH, Drummond MF, Haynes RB. The role of before-after studies of therapeutic impact in the evaluation of diagnostic technologies. J.Chronic Dis. 1986;39(4):295-304.

Guyatt G, Gutterman D, Baumann MH, Addrizzo-Harris D, Hylek EM, Phillips B et al. Grading strength of recommendations and quality of evidence in clinical guidelines: report from an American College of Chest Physicians Task Force. Chest 2006;129:174-81.

Guyatt GH, Rennie D. Users' Guides to the Medical Literature. A Manual for Evidence-Based Clinical Practice. AMA Publishing Group; 2007.

Haynes B, Can it work? Does it work? Is it worth it? BMJ,1999; 319; 653-653

Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60.

Irwig L, Bossuyt P, Glasziou P, Gatsonis C, Lijmer J. Designing studies to ensure that estimates of test accuracy are transferable. BMJ 2002 Mar 16;324(7338):669-671.

Karnon J, Goyder E, Tappenden P, McPhie S, Towers I, Brazier J, Madan J. A review and critique of modelling in prioritising and designing screening programmes. Health Technol Assess. 2007 Dec;11(52):iii-iv, ix-xi, 1-145.

Khan KS, Kunz R, Kleijnen J, Antes G. Systematic Reviews to Support Evidence-based Medicine: How to review and apply findings of health care literature. London: The Royal Society of Medicine Press, 2002.

Leeflang MMG, Scholten RJPM, Rutjes AWS, Reitsma JB, Bossuyt PMM. Use of methodological search filters to identify diagnostic accuracy studies can lead to the omission of relevant studies. Journal of Clinical Epidemiology, 2006 3;59(3):234-240.

Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat.Med. 2002 Jun 15;21(11):1525-1537.

Lijmer, 2009, Various randomized designs can be used to evaluate medical tests. J Clin Epidemiol

MacMahon S,.Collins R. Reliable assessment of the effects of treatment on mortality and major morbidity, II: observational studies. Lancet 2001;357:455-62.

Malmivaara A, Koes BW, Bouter LM, van Tulder MW. Applicability and clinical relevance of results in randomized controlled trials: the Cochrane review on exercise therapy for low back pain as an example. Spine 2006;31:1405-9.

Medical Services Advisory Committee. Guidelines for the assessment of diagnostic technologies. August 2005.

Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al. Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses? Lancet 1998 Aug 22;352(9128):609-613.

Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. Ann.Intern.Med. 2001;134:657-62.

Moons KG, van Es GA, Michel BC, Buller HR, Habbema JD, Grobbee DE. Redundancy of single diagnostic test evaluation. Epidemiology 1999 May;10(3):276-281.

Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat.Med. 1993 Jul 30;12(14):1293-1316.

Philips Z, Bojke L, Sculpher M, Claxton K, Golder S. Good practice guidelines for decision-analytic modelling in health technology assessment: a review and consolidation of quality assessment. Pharmacoeconomics. 2006;24(4):355-71.

Radford MJ,.Foody JM. How do observational studies expand the evidence base for therapy? JAMA 2001;286:1228-30.

Raffle, A, Muir G. Screening.Evidence and practice. Oxford University Press 2007

Reitsma JB, Glas AS, Rutjes AWS, Scholten RJPM, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. Journal of Clinical Epidemiology, 2005 10;58(10):982-990.

Rennie D. How to report randomized controlled trials. The CONSORT statement [editorial; comment]. JAMA 1996;276:649.

Ritchie G, Glanville J, Lefebvre C. Do published search filters to identify diagnostic test accuracy studies perform adequately? Health Information and Libraries Journal 2007 09/24;24(3):188-192.

Rutjes AW, Reitsma JB, Coomarasamy A, Khan KS, Bossuyt PM. Evaluation of diagnostic tests when there is no gold standard. A review of methods. Health Technol.Assess. 2007 Dec;11(50):iii, ix-51.

Sawaya GF, Guirguis_Blake J, LeFevre M, Harris R, Petitti D, for the U.S. Preventive Services Task Force. Update on the methods of the U.S. Preventive Services Task Force: Estimating certainty and magnitude of net benefit.

Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med 2010, 152: 726-32 see also http://www.consort-statement.org/consort-statement/

Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995 Feb 1;273(5):408-412

Sox HC; Comparative effectiveness research: a report from the Institute of Medicine. Ann Int Med 2009, 151, 203-205.

Strech et al. Value judgements in the analysis and synthesis of evidence. J Clin Epidemiol 2008, 61 521-4

Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008-12.

Tatsioni A, Zarin DA, Aronson N, Samson DJ, Flamm CR, Schmid C, et al. Challenges in systematic reviews of diagnostic technologies. Ann.Intern.Med. 2005 Jun 21;142(12 Pt 2):1048-1055.

Trikalinos TA, Siebert U, Lau J. Decision-Analytic Modeling to Evaluate Benefits and Harms of Medical Tests: Uses and Limitations. Med Decis Making. 2009 Sep-Oct;29(5):E22-9.

vanTulder M, Furlan A, Bombardier C, Bouter L, Editorial Board of the Cochrane Collaboration Back Review Group. Updated method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group. Spine 2003;28(12):120-9.

Weinstein MC, O'Brien B, Hornberger J, Jackson J, Johannesson M, McCabe C, Luce BR; ISPOR Task Force on Good Research Practices--Modeling Studies. Principles of good practice for decision analytic modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices--Modeling Studies. Value Health. 2003;6(1):9-17.

Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med.Res.Methodol. 2003 Nov 10;3:25.

Whiting P,  Westwood M, Burke M, Sterne J, Glanville J. Systematic reviews of test accuracy should search a range of databases to identify primary studies.  J.Clin.Epidemiol., 2008, 61, 4, 357.

Costs and economic evaluation

Suvi Mäklin, Sunya-Lee Antoine, Irina Cleemput, Belén Corbacho, Maria Rosaria Perrini, Janek Saluse, Isaura Veira

Description

What is this domain about?

The main aim of the costs and economic evaluation domain is to provide information about the relative costs and cost-effectiveness of health care technologies. It is intended to support decision-making regarding resource allocation for health technologies in the health care sector, to include emerging, new and existing technologies (Kristensen 2007). An economic evaluation identifies, measures, values and compares the costs and outcomes of a technology with its relevant comparator. Its aim is to inform value for money judgements about an intervention (Guidelines for the Economic Evaluation of Health Technologies, 2006).

Central to this area of economics are the concepts of opportunity cost and incremental change. In publicly-funded health care systems, finite resources mean that not all interventions can be provided in every situation for all who need or want it. Choices must be made between effective health care interventions; the decision to fund one intervention may mean that others cannot be funded (Guidelines for the Economic Evaluation of Health Technologies, 2006).

Why is this domain important?

Economic evaluation is an important part of health technology assessment. Over the last two decades, the rate of increase in health-care costs has accelerated, placing increasing pressure on the finite resources available to fund them. This growth in costs has been fuelled in part by the rate of technological development. Increasingly, there is a conflict between what is technologically possible and what is economically feasible. Clinical investigators have begun to recognise the importance of performing economic evaluations alongside RCTs. In evaluating a new technology, it is not sufficient to consider evidence of its efficacy and effectiveness; data on its costs and other outcomes are also needed.

Relation to other domains

Costs Domain requires information from Health problem and current use, Effectiveness, Safety and Organisational Domains.

Specific features in finding, interpreting or implementing information for this domain

An economic evaluation should provide decision makers with information that is useful, relevant, and timely. The economic evaluation component of an HTA should be conducted within a common methodological framework that consists of a well-defined research question depicting a specific health policy problem or question, a perspective and scope of analysis, and a set of alternatives to be assessed comparatively (Liberati 1997). It is important to provide a detailed description of the alternatives and to justify their choice, so that study users can assess their relevance to their own setting. What represents ‘current practice’ may vary over time and between countries. There may also be regional variation in the importance of other elements for the economic evaluation. Therefore, transparency in reporting of economic evaluations is critical to allow the applicability and relevance to economic evaluations performed as part of an HTA to be assessed for different settings.

Issues specific for screening technologies

The overall costs and benefits (effects) of a screening programme should be assessed prior to its implementation (organisation in real life). The economic evaluation of a screening programme differs in a number of respects to that of other health care interventions. In general, the costs of screening programmes are high. Screening may not be limited to a single screening test, but may include follow-up screenings and interventions for those with a positive result; the evaluation of a screening programme therefore needs to incorporate other health care interventions in the analysis. The interventions chosen, the rationale for their inclusion and the measurement of the resources consumed should be clearly described. A decision to implement a programme should take into account the sensitivity and specificity of the screening technology, the number of positive and negative results (true and false, ie. positive predictive value PPV and negative predictive value NPV) and the implications of false-positive and false-negative results. Potential benefits of screening include a more timely diagnosis, allowing more timely treatment with associated reductions in morbidity and, or mortality.

The generalisability of clinical trial data may be limited due to the range of choices for the screening test, screening interval, the eligible population and the organisation of the screening programme. There may also be difficulties in extrapolating benefits from clinical trial data due to the long time interval between screening and the development or progression of the condition of interest (Karnon et al. 2007).

The long time horizon has particular implications for discounting. A decision to discount costs or outcomes, or both, and the choice of the discount rate(s) may have a significant impact on the cost-effectiveness of the intervention and needs to be carefully considered. Most of the costs of a screening programme are incurred within a relatively short time period, whereas the benefits (e.g. life years gained) may not be accrued for many years; for many curative interventions both the costs and the effects occur immediately. The decisions regarding discounting should be made explicit and according to available, e.g. country-specific, guidelines.

Another issue to be considered is the incorporation of utilities in the analyses. Screening programmes profoundly differ from the situation where a patient seeks care due to symptoms, as screening targets populations who are mostly healthy. “Healthy” people may become patients due to their screening result and thus the effect of screening on their utility may be significant, though data on this is fairly limited (Karnon et al 2007). Furthermore, false positive and false negative test results may have impact on peoples’ behaviour, and this in turn, may change the results of the analysis. The data on these issues are limited, some implications exist that false negative test result might lead to more risk-taking behaviour, e.g. a person gets a low cholesterol reading chooses a less healthy diet. The researchers should consider such possible effects and try to assess their impact (e.g. how would the ICER change if false negative screens changes peoples’ behaviour in a specific direction).

Methodology

Where to find information?

Databases and search strategies

There are two main purposes for searching for information in economic evaluation. First, when planning and scoping an economic evaluation on any topic, it is useful to search for what is already published on that topic elsewhere. A systematic review of previously published economic evaluations may be done. Furthermore, relevant literature and other data sources may be searched in order to find information on different aspects (e.g. clinical effectiveness, quality of life, resource use, costs) to be combined in a modelling study.

The key sources for published economic evaluations are MEDLINE, EMBASE, EconLit and the CRD-databases, especially NHS Economic Evaluation Database (NHS-EED).

The Hedges Project at McMaster University in Canada develops search strategies to improve retrieval of published studies from large databases. Examples of search strategies for cost and economic studies from MEDLINE are available:

http://hiru.mcmaster.ca/hiru/HIRU_Hedges_MEDLINE_Strategies.aspx#Costs

http://hiru.mcmaster.ca/hiru/HIRU_Hedges_MEDLINE_Strategies.aspx#Economics

Useful other sources
  • Registries (e.g. national screening registries),
  • international, national and regional statistics,
  • national and regional guidelines,
  • hospital databases (costs, resource use data),
  • patient reported outcome and quality of life instruments database (http://www.proqolid.org),
  • expert opinions and
  • manufacturers' handbooks.

What kind of information is required?

 Study types, design, outcomes measures

Four main types of economic evaluation can be part of HTA; cost-minimisation, cost-effectiveness, cost-utility and cost-benefit analyses (Table 1, Modification from Drummond 2005). The difference between them is based on how health outcomes are measured and valued. The choice between the different types of economic evaluations for answering a specific question depends on the purpose of the evaluation, the availability of specific data and potentially the guidelines for economic evaluations that are to be followed in a specific context.

The objective of economic evaluations -the main types of which are described in Table 1- is different from the objective of a budget impact analysis (BIA). While economic evaluations attempt to inform policy makers about the most efficient way to allocate the available health care resources, given the objective to optimize health outcomes of the population, BIA estimates the financial consequences of adopting a new intervention in health care without taking the health consequences into account. In the Core Model, BIA is included in the Organisational Domain.

Table 1. Types of full economic evaluation.

Type of economic evaluation

Appropriate if ...

Valuation of costs

Valuation of outcomes

The question to be answered

Cost-minimisation analysis (CMA)

the compared technologies are equally effective; data on costs suffice.

Monetary units

None

Which intervention is the least costly?

Cost-effectiveness analysis (CEA)

the effectiveness of the compared technologies is different (e.g. the difference in costs have to be weighted against the difference in effectiveness);

activities with the same aim and measure of effectiveness are compared.

Monetary units

Natural units (e.g. life years gained, disability-days saved, points of blood pressure reduction, etc.)

What is the intervention’s incremental cost per additional unit of outcome as compared to its best alternative?

Cost-utility analysis (CUA)

HRQoL is an important health outcome;

and/or

activities across specialities or departments in the health care sector are compared.

Monetary units

QALYs, HYEs

What is the intervention’s incremental cost per additional unit of outcome as compared to its best alternative?

Cost-benefit analysis (CBA)

non-health effects are also of importance (e.g. the treatment process itself, utility of information);

or only one technology is assessed (net benefit);

or

there is a wish that individual life's are valued in monetary units;

or

activities across different sectors in society have to be compared.

Monetary units

Monetary units

What is the economic trade-off between different activities that matter for society?

Perspective

The perspective of the economic evaluation is a key element in defining which costs and outcomes should be included in the analysis. For instance short stay at hospital may be cost-effective from the perspective of the hospital but it may be more costly to society if the cost of home care is taken into account.

In the ideal situation the economic evaluation is conducted from the broadest possible perspective. The most comprehensive perspective is societal and then all relevant costs and outcomes of the technologies have to be identified, measured and valued, no matter whom these costs and outcomes fall on (Drummond 2005). Other possible perspectives are the health care sector’s perspective, third party payer’s perspective, hospital perspective or patients' perspective. The perspective chosen ultimately depends on the purpose of the economic evaluation. If the purpose is to inform societal resource allocation, the societal perspective should be taken. For hospital HTA, the hospital perspective may be more appropriate.

Costs

The costing procedure can be divided into three phases: identification, measurement and valuation of resource use. First of all the relevant resources used have to be identified, then the volume or number of units of the resource use has to be measured and finally these volumes have to be valuated. An important consideration is also the choice of time period, i.e. the choice for how long the resource use should be tracked and measured. The length of time period depends on what is relevant to the topic of evaluation, which in some cases may include lifetime.

Direct costs are all costs directly related to a disease or technology. They include costs borne inside the health care sector (e.g. materials, equipment, personnel, tests – direct health care costs) as well as outside the health care sector (e.g. patients’ travel time – direct non-health care costs). A broad agreement exists that all costs related to the disease or technology in question should be included in the analysis. A more debated issue is whether to include the unrelated future health care costs or not, such as health care costs of other diseases which people experience when they live longer thanks to a certain treatment or screening. Whether related or unrelated, future costs should be discounted according to national guidelines, if such exist.

Indirect costs include the patient’s temporary absence from work due to illness, reduced working capacity due to illness and disablement, or lost production due to an early death. The lost production can be measured either by means of the human capital method or the friction cost method. Lost production is most often reported separately and not integrated in the cost estimate used for the calculation of the incremental cost-effectiveness or cost-utility ratio. Its valuation is made only in situations where it is judged to be relevant. The concept of lost production should not be confused with a transfer payment like sickness benefit. Inclusion of transfer payments depends on the perspective of the analysis; they are a cost to the paying organisation (e.g. government), a gain to the recipient, but from a societal point of view, not a cost nor a gain.

Physical units or volumes of resources used should be reported separately from the unit costs of resources to allow decision makers to assess the applicability of resource use estimates to their own setting. In addition it is recommended to report direct costs separately from indirect costs.

All costs should be adjusted to a common price level (usually the year of analysis).

Issues specific for screening technologies

When identifying the costs of screening, all the costs associated to the screening programme should be included. This means, that in addition to the costs of screening test itself, also costs of the screening organisation, invitations to screening, further examinations as well as possible treatment costs need to be included. Also, travel costs to and from the screening facility, if relevant, should also be taken into account when identifying the costs.

Population on screening programmes can be considered as healthy people not unable to work because of any health related condition. In that sense, the lost time as a consequence of undergoing the screening programme should be considered as lost productivity and be included as a cost in the economic evaluation.

Outcomes

Health outcomes of interventions can be measured by natural units of health (e.g. deaths, life years gained (LYG)), valuations of health states or utilities, or in monetary terms (Table 1).

If the intervention affects both the length and the quality of life, a composite outcome measure, such as Quality-Adjusted Life Years (QALYs) or Healthy Years Equivalent (HYEs) should be used. The QALY-approach and similar approaches are useful in policy analysis and program decision-making because they are generic and consequently allow broad comparisons between interventions and across diseases. They can in principle be estimated for any population, any disease, any intervention, and can be used to compare across diverse programs, assuming that studies use the same methodology.  Health-related quality of life (HRQoL) refers to aspects of quality of life that are related to health. There are different tools to measure HRQoL and there is no single measure which has been accepted as the gold standard. Patient outcome measures that extend beyond traditional measures of mortality and morbidity, to include physical, social, and emotional aspects that are relevant and important to an individual's wellbeing can be assessed using a disease-specific, generic, or a preference-based instrument. However, for economic evaluation an index measure is at least needed. To be able to compare outcomes in different disease areas, a generic measure should moreover be used. Single index HRQoL instruments combine the answers of individual questions into a single index number (usually ranging between 0 and 1, although negative values for states worse than death are possible). Generic instruments providing a single index number suitable for the calculation of QALYs include AQoL (Assessment of Quality of Life), EQ-5D (EuroQol), 15D, HUI (Health Utilities Index Mark II/Mark III), QWB (Quality-of-Well Being Scale), Rosser-Kind and SF-6D (based on RAND-36/SF-36).

Future outcomes should be discounted according to national guidelines, if such exist.

Issues specific for screening technologies

In assessment of outcomes, the definition of the intervention and comparator is critical. With regards to screening, it is critical to define the entire care pathway following the screening test (as well as following the comparator).

Screening programmes profoundly differ from the situation where a patient seeks care due to symptoms, as screening is usually targeted to populations who are mostly healthy. This implies that these “healthy” people may become patients due to the screening results and thus the effect of screening on their utility may be significant though data on this is fairly limited (Karnon et al 2007). Also the screening may cause anxiety and concern, especially in the case of false positive test result. The effects on the patients' utility or HRQoL are still fairly unknown, yet some qualitative evidence exists from cancer screenings that abnormal and false-positive screening results have a negative impact on certain psychosocial domains (Brodersen et al 2007).

Tools for critical appraisal

There are several methodological characteristics to consider, when assessing the quality of an economic evaluation. A report of an analysis should inform the reader about all the important aspects of an analysis. Several checklists have been published, in order to use when reporting an economic evaluation, but also to help in identifying the strengths and weaknesses of different studies (e.g. Siegel et al 1996; Drummond et al 2005; BMJ guidelines for authors and peer reviewers of economic submissions to BMJ). Below is presented an example, a summary of a checklist by Drummond et al (2005):

  1. Was a well-defined question posed in answerable form?
  2. Was a comprehensive description of the competing alternatives given?
  3. Was the effectiveness of the programmes or services established?
  4. Were all the important and relevant costs and consequences for each alternative identified?
  5. Were costs and consequences measured accurately in appropriate physical units?
  6. Were costs and consequences valued credibly?
  7. Were costs and consequences adjusted for differential timing?
  8. Was an incremental analysis of costs and consequences of alternatives performed?
  9. Was allowance made for uncertainty in the estimates of costs and consequences?
  10. Did the presentation and discussion of study results include all issues of concern to users?

Analysing and synthesizing evidence

As all relevant evidence is rarely available from a single source, the mostly used approach in economic evaluation is modelling: collecting the best available evidence from various sources and synthesising it using appropriate modelling techniques.

Study frame and scoping of the economic evaluation of screening technologies

A coherent and manageable economic analysis needs a framing or scoping of the analysis that defines the following aspects of the analysis:

Target population

The population or group of people at risk of a disease that the screening is aimed at

Intervention

The screening technology being studied

Comparators

The alternative technologies that the screening is being compared to (often including, but not limited to, current practice or “no systematic screening”)

With respect to evaluation of screening, two main types exist: comparison of screening vs. no screening, and comparison of different screening tests within one screening (e.g. faecal occult blood test vs. colonoscopy in colorectal cancer screening).

Outcomes

The positive or negative health outcomes that are included in the analysis.

Specific to screening are the outcomes caused by screening to people who would not have been examined or treated in absence of screening.

Costs

The costs of the compared screening technologies and further examinations and treatments

Organisational and management costs

Time horizon

The time frame during which cost and outcomes are assessed

Perspective

The perspective from which costs and outcomes are assessed

Evaluation type

The chosen type of economic evaluation (e.g. cost-effectiveness, cost-utility, cost-benefit analysis)

Analysis methods and modelling

The statistical tests/models for analyzing the data

Discounting  

Rate at which future costs and outcomes are discounted

Sensitivity analysis

The chosen type of sensitivity analysis (e.g. one-way SA, probabilistic SA)

The chosen variables whose values are uncertain are subjected to a sensitivity analysis

Modelling

There are several reasons for carrying out an economic evaluation with modelling, for example in a situation where economic and clinical data are missing or when there is a need for extrapolation of short-term clinical data to the long run. Decision trees and Markov models are the most frequently used types of models, but also other approaches are used (e.g. discrete-event simulation, micro simulation). 

Useful links:

International Society for Pharmacoeconomics and Outcomes Research (ISPOR) has published guidelines for conducting and reporting modelling studies (Weinstein et al 2003) at http://www.ispor.org/workpaper/healthscience/TFModeling.asp 

More detailed guidelines are in development for e.g.

–        state-transition http://www.ispor.org/workpaper/modeling_methods/state-based-modeling.asp

–        discrete event http://www.ispor.org/workpaper/modeling_methods/Modeling-discrete-event-simulation.asp

–        dynamic transmission modelling http://www.ispor.org/workpaper/modeling_methods/Dynamic-transmission-modeling.asp  .

Sensitivity analysis

Economic evaluation is often based upon estimates of variables that are characterised by a specific distribution. Besides parameter uncertainty, economic evaluations –and more specifically economic models- are often based on assumptions about the relationship between parameters which are also uncertain. It is important to take this uncertainty into account in the evaluation, either parameter or model uncertainty. Sensitivity analysis will show the decision maker, how robust (trustworthy) the results and conclusions of the economic analysis are. Deterministic and/or probabilistic sensitivity analyses should always be a part of an economic analysis (Guidelines for the Economic Evaluation of Health Technologies: Canada, 3rd edition, 2006; Guidelines for pharmacoeconomic evaluations in Belgium: Brussels, 2008). Especially in economic models it is very important to conduct a complete sensitivity analysis for all uncertain model inputs to determine the impact on the results. Omission of any model input from the sensitivity analysis should be justified. Different methods to handle uncertainty are presented by Briggs et al 1994 and Briggs et al 2006.

Discounting

Cost and outcomes in the economic analysis that occur in the future should be discounted. Discounting, or calculating the present values of future costs and consequences, makes it possible to compare health technologies in an economic analysis whose costs and outcomes do not occur at the same time. Discounting should not be confused with inflation.

Issues specific to screening

Most of the costs of a screening programme incur within a relatively short time period and typically the benefits (e.g. life years gained or quality-adjusted life years gained) incur after a longer time period, while in many curative interventions both the costs and the effects occur immediately. .The consequences of discounting in cost effectiveness analysis are often substantial This means that the questions related to discounting need to be carefully examined. By attaching a lower weight to future health makes preventive health care seem less cost effective because such interventions typically involve current costs and future effects. The decisions to be made are; whether to discount both costs and effect or not, which discount rate to use, and should both, costs and effects, be discounted using the same discount rate? On this issue, please refer to possible national guidelines.

Meta-analysis

Theoretically, it is possible to conduct meta-analysis of economic evaluations, but is not generally used. The existing heterogeneity between studies would demand a great deal of adjustments, which are often not possible. Not only the methods used in economic evaluations vary across studies, but also more profound elements of the research questions, comparators, perspectives, health care systems, clinical guidelines, resource use and time horizon differ significantly (CRD, 2009).

Synthesis
Incremental Cost-Effectiveness Ratio, ICER

To be able to conclude which health technology is cost-effective, both the total costs and the effectiveness of at least two interventions have to be compared. The comparison may lead to nine different situations, as described in the decision matrix below.

Table 2. The cost-effectiveness decision matrix (Kristensen 2007)

A new technology compared with an old one

Less effective

Same effectiveness

More effective

Less costly

1.No clear decision

non-dominance =>

Incremental analysis needed

4.Adopt the new technology

the new dominates the old (weak dominance)

7.Adopt the new technology

the new dominates the old (strong dominance)

Same costs

2.Keep the old technology

the old dominates the new (weak dominance)

5.The technologies are equal

8.Adopt the new technology

the new dominates the old (weak dominance)

More costly

3.Keep the old technology

the old dominates the new (strong dominance)

6.Keep the old technology

the old dominates the new (weak dominance)

9.No clear decision

non-dominance => incremental analysis needed

In situations described in cells 1 and 9 incremental analysis is needed to decide, which technology is preferable. For that purpose an incremental cost-effectiveness ratio (ICER) has to be calculated. It is a ratio of the difference in costs of interventions to the difference in outcomes. The ICER indicates the costs of achieving one extra unit of health benefit when switching from one alternative to another. The new intervention is cost-effective if the society is willing to pay for the additional benefits (cell 9) or if the society considers that the cost savings compensate for the lower effectiveness (cell 1).

Threshold cost-effectiveness and net benefit approach

Whether an intervention is cost-effective depends on its relation to the maximum willingness-to-pay for a unit of outcome, or the so-called ICER threshold. If the ICER of the intervention is lower than the threshold, the intervention is considered cost-effective (i.e. improving efficiency in health care). If it is higher than the ICER threshold, the intervention is not considered cost-effective and resource allocation to this intervention would not increase efficiency in health care. 

The ICER seems to be most popular method but the ratio gives no idea of the size or scale of the interventions being considered. Presenting ICER alone, however, is not sufficient and it should be presented along with other separate relevant outputs of the economic analysis (absolute health benefits, number of patients, etc). ICER is one of the decision elements, alongside others. The net benefit approach is an alternative summary measure of the value for money. Net monetary benefit (NMB) and net health benefit (NHB)) will be used to overcome problems with cost-effectiveness ratios. Both NMB and NHB are functions of the threshold cost-effectiveness ratio (Drummond 2005).

Cost-Effectiveness Plane and CEAC

Incremental Cost-Effectiveness Plane

The incremental cost and incremental effect can be represented visually using the incremental cost-effectiveness plane (Black 1990), which is divided into four quadrants through the origin. The horizontal axis divides the plane according to incremental cost (positive above, negative below) and the vertical axis divides the plane according to incremental effect (positive to the right, negative to the left).

Cost-Effectiveness Acceptability Curve

Cost-effectiveness acceptability curve is a method of summarizing the information about uncertainty in cost-effectiveness. The CEAC shows the probability that an intervention is cost-effective at each ceiling ratio (or willingness-to-pay threshold), according to the available data. More detailed information about CEAC can be found in, for example, Briggs et al 2006; Fenwick et al 2006.

Reporting and interpreting

A common reporting format increases transparency of studies and facilitates comparison between studies. Several guidelines for economic evaluation have also suggested reporting formats and most of them include at least following items (Drummond & Jefferson 1996; Drummond 2005; CADTH guidelines 2006):

  • Costs (direct and indirect costs) and effectiveness (life years gained, quality-adjusted life years gained, etc.) should be reported both in disaggregated and aggregated form. Also undiscounted values should be reported.
  • An incremental analysis (ICER, ICUR), comparing the relevant alternatives.
  • Conclusion drawn from the analysis, answering the original question of the study. Strengths and limitations of the study should also be reported.
Transferability of resource utilization and unit cost elements

Costs of technologies are generally not transferable from one country to another. However, transferability of individual elements of data differs. Table 3 contains our assessment of transferability for each element. Although the resource utilization and unit cost elements are only partially transferable or not transferable at all, they are all essential parts of an economic assessment. The relevance of economic evaluations cannot be judged without information on these elements. Moreover, data on types and amounts of resources used in one country are often valuable information for researchers performing anHTA in another country.

Table 3 Transferability of resource utilization and unit cost elements

Data Element

Transferability

What types of resources are used when delivering the assessed technology and its comparators?

Partially transferable. In most cases types of resources are completely transferable, but this should be tested, if appropriate.

What amounts of resources are used when delivering the assessed technology and its comparators?

Partially transferable. It is a well-known fact that resource utilization when delivering a specific technology can differ between countries, e.g. the average number of hospital days for a specific procedure may vary considerably. Other types of resource utilization may vary little between countries. Transferability for this issue is an empirical question that needs to be addressed carefully.

What are the unit costs of the resources used when delivering the assessed technology and its comparators?

Not transferable. Although some unit prices are comparable between countries, it cannot generally be assumed that unit costs are transferable.

Health-economic data can be collected alongside a randomized clinical trial, so called piggyback evaluation. Advantages of this are the internal validity by trial design and practicality in collection of data on resource use and effectiveness simultaneously. The aims of clinical trials and economic evaluations, however, differ in significant ways, which can lead to disagreements in many aspects (time horizon, sample size, etc). (Drummond et al 2005)

As all relevant evidence is rarely available from a single source, the mostly used approach in economic evaluation is modelling: collecting the best available evidence from various sources and synthesising it using appropriate modelling techniques.

Assessment elements

E0001 Assessment element card

Issue: What types of resources are used when delivering the assessed technology and its comparators (resource use identification)?

Topic: Resource utilization

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

In order to do an economic evaluation all types of resource utilization must be identified. The study perspective determines what kinds of resource utilization must be identified. A societal perspective implies identifying all kinds of resource utilization irrespective of who pays for the resources or whether the costs are born inside or outside the health care sector. If a health care provider perspective is applied, then resource utilization paid for by the patient is not relevant and if a health care payer perspective is applied, non-health care costs should not be taken into account. In identifying the resource use of a screening programme, the screening test, further examinations and treatments, as well as administration and organisation of the screening programme need to be taken into account.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Health care registers and databases, RCT’s with resource utilization data, reimbursement databases, micro-level costing studies/ABC-costing studies

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0011, A0013, A0014, A0015, A0016, A0017, A0018) G0001, G0003, G0004, G0005, G0006, G0007, G0010, B0007, B0008, B0009

Sequential relations

E0002 Assessment element card

Issue: What amounts of resources are used when delivering the assessed technology and its comparators (resource use measurement)?

Topic: Resource utilization

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

After identifying the types of resources used, also the quantities of resources must be measured, for all types of resource utilization of implementing the technology and its comparators. Resource use data may be collected prospectively (e.g. alongside a clinical trial) or estimated retrospectively by reviewing patient registries, hospital or reimbursement databases, or other routine data collection.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Health care registers and databases, RCT’s with resource utilization data, reimbursement databases, micro-level costing studies/ABC-costing studies

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational, Health Problem and Current Use, B0007, B0008, B0009

Sequential relations

E0003 Assessment element card

Issue: What are the unit costs of the resources used when delivering the assessed technology and its comparators?

Topic: Unit costs

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Ideally unit cost estimates should be (proxies for) opportunity costs. By the opportunity cost is understood the value of the (lost) health gains that could have been achieved from an alternative technology, which, however, cannot be introduced or retained, because the resources e.g. manpower, are used on the new technology. Market prices or shadow prices (e.g. for voluntary work) are often used as proxies for opportunity costs. Also costs caused by a false negative or false positive screening test result should be included.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Market prices, companies, hospital accounting systems, reimbursement databases, micro level costing studies/ABC-costing studies

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Sequential relations

E0004 Assessment element card

Issue: What is the impact of the technology on indirect costs?

Topic: Indirect Costs

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Indirect costs include costs to society of lost production. This can be due to patient’s temporary absence from work due to examinations, treatments, or illness; reduced working capacity due to illness and disablement; or lost production due to an early death. Depending on the perspective of analysis, also indirect costs related to patients and relatives (e.g. income loss, transportation costs) should be examined.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

The data are available from different registers e.g. register on sick leave, sickness allowance, patient administration systems/ clinical databases, earlier studies, cost diaries.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

D0014, Social

Sequential relations

E0005 Assessment element card

Issue: What are the incremental effects of the technology relative to its comparator(s)?

Topic: Outcomes

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The calculation of an incremental cost-effectiveness ratio (ICER) requires the estimation of the incremental effectiveness/utility/benefit of an intervention relative to its comparator(s). Estimation of utility related to screening differs from many curative interventions, since the target population of screening is healthy or at least asymptomatic, who might become patients due to the screening. Benefits of screening include improved diagnosis, timely and appropriate treatment and reduction in mortality and morbidity. Also the number of detected positives and false positives (specificity and sensitivity) are important aspects in evaluation of effects of the assessed screening programme.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Estimation of the incremental effects can be based on information provided in the effectiveness domain (e.g. mortality data). Additional information collection may be needed (e.g. on health-related quality of life indices). The incremental effectiveness may result from an economic model, where inputs from the effectiveness domain are used.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0004, A0005, A0006, A0009, A0023) D0001, D0002, D0003, D0004, D0005, D0006, D0008, D0011, D0012, D0013, D0023, D0030, C0001, C0002, C0003, C0004, C0006, C0008

Sequential relations

E0008 Assessment element card

Issue: What is the method of analysis?

Topic: Cost-effectiveness

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Clinical trials usually compare a limited number of screening options over a relatively short time horizon and it is unlikely that trial data will inform all relevant aspects of a screening programme. Decision analytic models provide a structure for synthezising information from various sources as well as analysing how the uncertainty affects the results.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Sequential relations

E0007 Assessment element card

Issue: What is the appropriate time horizon?

Topic: Cost-effectiveness

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Both costs and effects should be modelled over an appropriate time horizon. In most curative interventions both costs and effects occur in a relatively short time period, while in screening the effects occur later in the future. Effectiveness data is rarely available for the whole appropriate time horizon and economic evaluation needs to link intermediate endpoints to final endpoints and/or extrapolate the effectiveness. Thus it is often argued that the effects are penalized by discounting and there is controversy on this issue. One needs to take into account any relevant official guidance when choosing specific discount rate for analysis. After that it is important to decide whether to discount both costs and effects, and whether to use uniform discount rate.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Effectiveness domain (effectiveness data may need extrapolation)

Sequential relations

E0006 Assessment element card

Issue: What is the incremental cost-effectiveness ratio?

Topic: Cost-effectiveness

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The result of the economic analysis will most often be an incremental cost-effectiveness ratio eg. costs/QALY. if quality-adjusted life years is used as the main outcome indicator. The incremental cost-effectiveness ratio does not in itself determine that a technology is desirable. Decision makers need – implicitly or explicitly – to weigh the benefits of an intervention against the costs. The concept of a cost-effectiveness threshold is one way of expressing decision-makers willingness-to-pay for health benefits. If other type of economic evaluation is chosen, eg. cost benefit analysis, other types of measures are used to express results of the analysis, but most current economic analysis within HTA’s are done within the cost-effectiveness/cost-utility framework.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Sources of data used are specified under relevant issues under domains safety, effectiveness and costs. The ICER estimate might result from the economic model, using inputs from the safety and effectiveness domain.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Guidelines for economic evaluation of Health Technologies: Canada, 3rd edition, 2006, Guidelines for Pharmacoeconomic Evaluations in Belgium, 2008

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety, Effectiveness

Sequential relations

References

  1. Black WC. The CE plane: a graphic representation of cost-effectiveness. Med Decis Making 1990;10(3):212-4.
  2. Briggs A, Schulpher M, Buxton M. Uncertainty in the economic evaluation of health care technologies: The role of sensitivity analysis. Health Economics 1994;3:95-104.
  3. Briggs A, Claxton K, Sculpher M. Decision Making for Health Economic Evaluation.New York : Oxford University Press, 2006.
  4. Brodersen J, McKenna SP, Doward LC, Thorsen H. Measuring the psychosocial consequences of screening. Health and Quality of Life Outcomes 2007:5;3.
  5. Centre for Reviews and Dissemination. Systematic Reviews: CRD's Guidance for Undertaking Reviews in Healthcare. CRD; University of York, 2009.
  6. Cleemput I, Van Wilder P, Vrijens F, Huybrechts M, Ramaekers D. Guidelines for Pharmacoeconomic
  7. Danish Centre for Health Technology Assessment, National Board of Health, 2007. http://www.sst.dk/publ/Publ2008/MTV/Metode/HTA_Handbook_net_final.pdf
  8. Drummond M.F, Sculpher M.J, Torrance G.W., O`Brien B.J., Stoddart G.L, Methods for the Economic Evaluation of Health Care Programmes. Third Edition. Oxford University Press, 2005.
  9. Drummond M, Manca A, Sculpher M. Increasing the generalizability of economic evaluations: Recommendations for the design, analysis, and reporting of studies. International Journal of Technology Assessment in Health Care 2005;21: 165-71.
  10. Drummond MF, Jefferson TO, on behalf of the British Medical Journal Economic Evaluation Working Party. Guidelines for authors and peer reviewers of economic submissions to the BMJ. British Medical Journal 1996;313:275-283
  11. Evaluations in Belgium. Health Technology Assessment (HTA). Brussels: Health Care Knowledge Centre (KCE); 2008. KCE Reports 78C (D/2008/10.273/27)
  12. Fenwick E, Marshall DA, Levy AR, Nichol G. Using and interpreting cost-effectiveness acceptability curves: an example using data from a trial of management strategies for atrial fibrillation. BMC Health Serv Res 2006;6:52.
  13. Guidelines for the Economic Evaluation of Health Technologies: Canada, 3rd Edition, 2006. http://www.cadth.ca/media/pdf/186_EconomicGuidelines_e.pdf
  14. Karnon J, Goyder E, Tappenden P, McPhie S, Towers I, Brazier J, et al. A review and critique of modelling in prioritising and designing screening programmes. Health Technol Assess 2007;11(52).
  15. Kristensen FB, Sigmund H (ed.). Health Technology Assessment Handbook. Copenhagen: Danish Centre for Health Technology Assessment, National Board of Health, 2007. http://www.sst.dk/publ/Publ2008/MTV/Metode/HTA_Handbook_net_final.pdf
  16. Liberati A, Sheldon TA, Banta HD. EU-ASSESS project subgroup report on methodology. Chapter 13: Economic evaluation.  International Journal of Technology Assessment in Health Care 1997;13:206-10.
  17. NHS NICE. Guide to the methods of technology appraisal.Issued: June 2008. http://www.nice.org.uk/media/B52/A7/TAMethodsGuideUpdatedJune2008.pdf
  18. O'Sullivan AK, Thompson D, Drummond MF. Collection of Health-Economic Data Alongside Clinical Trials: Is There a Future for Piggyback Evaluations? Value in Health 2005;8:67-79.
  19. Siegel et al 1996
  20. Weinstein MC, O'Brien B, Hornberger J, et al. Principles of good practice of decision analytic modeling in health care evaluation: Report of the ISPOR Task Force on Good Research Practices-Modeling Studies. Value Health 2003; 6:9-17

Ethical analysis

Mirella Marlow, Ilona Autti-Rämö, Bjørn Hofmann, Samuli Saarni, Sinikka Sihvo, Aleksandra Zagórska

Description

What is this domain about?

The term “ethics” is broadly used to describe activities relating to the understanding and study of “the moral life”.  The term “morality” encompasses beliefs, standards of conduct, principles and rules which may guide personal and professional behaviour and the behaviour of institutions.  Morals are standards that are widely shared, and that form some degree of social consensus (Beauchamp and Childress, 2001).

The ethical aspects domain encompasses the ethical issues raised by the health technology itself and by its implementation. The issues stem from the general values of the population, the aims of the healthcare system and from values arising from use of a technology. Ethical analysis also addresses specific issues inherent in the process of health technology assessment (HTA).  In carrying out ethical analysis, prevalent norms and values in society relevant to HTA are considered. The weight given to these norms can differ between societies and countries. Socio-political, cultural, legal, religious and economic differences also have a major impact on the moral value societies will attribute to the consequences of implementation of a technology.

Why is this domain important?

Ethical analysis aims to provide a thorough understanding of norms and values that need to be taken into account during the HTA and in the decision making process. Moral values and norms form the basis of social life and they play a key role in shaping the context in which health technologies are used. Ethical analysis also reflects the fact that HTA is a value-laden process which should not be considered as a purely technical tool for maximising the health benefits of technology, since benefit maximising is of itself a normative aim that carries a priori assumptions about the goals of healthcare and healthcare expenditure.

Although addressing ethical issues is generally accepted as an important component of the HTA process, their integration to date has often been limited. It can be argued that “integration” is not the right word since ethics is already a part of HTA (Hofmann 2008). The challenge is to make it more explicit and visible. The need for, and weight placed, on ethical analysis can differ greatly between technologies depending on the purpose and context of their use (Grunwald 2004). For example, a new test that targets the same biomarker as the one it is intended to replace but does so with better specificity, sensitivity, safety and at lower cost is likely to be less problematic than a new, risky technology for a previously undiagnosable disorder.

It should be noted that in taking ethical considerations into account in HTA, two separate but interconnected activities must be conducted.  One is the identification of moral issues relevant to the HTA, and the other is ethical analysis that will be used to draw conclusions about use of the technology, and, in some settings, for decision-making bodies.  The analysis will generally consist of using structured methods for exposing the relevant (often competing) moral values in the HTA, and weighing their relative merits (see potential ethical analysis methods below).  Those who are drawing conclusions about the use of the technology will need to apply the framework(s) in the course of the HTA to decide which of these possibly competing values should dominate.

Relation to other domains

Although ethical analysis is a separate domain in the HTA Core Model, moral issues are relevant to all HTA domains and the methods of ethical analysis should take this into account.. Although some argue that ethical and legal issues should be kept separate from the rest of the HTA process (Duthie & Bond 2011) , it can be important to integrate the ethical analysis in the entire HTA process, including assessment and decision making. Ethical issues, rather than being a "one session" task or an add-on, the various topics and issues described in the assessment element have to be identified and addressed at different phases of the assessment process. This is important in order to ensure that decision-makers are presented with a complete picture, but also because not all ethical considerations are apparent early in the HTA: sometimes they emerge as the clinical or cost-effectiveness evidence emerges.  For example, the assessment might indicate that the proposed technology is not cost-effective for subgroups who are protected by equalities legislation.  

Specific features in finding, interpreting or implementing information for this domain

Values are inseparable from HTA (Hofmann 2005a), so the question is whether to address them explicitly or implicitly. The relative weight placed on the ethical analysis and the selection of methods depends heavily on the technology being evaluated (Hofmann 2005b; 2008). The more the technology presents new, severe or fundamental value conflicts, or challenges to everyday norms or beliefs, the more emphasis should be placed on the ethical analysis. Methods and significance of integrating ethical analysis in HTA have been explored and actively advocated in the INAHTA ethics working group (Andersen et al 2005; Burns et al 2011 (in press)).

HTA organisations differ in their resources and mandate for decision-making: while some only provide synthesis of evidence, others conduct appraisal of evidence and formulate recommendations or produce clinical practice guidelines. Hence the available methods, weight and ways of reporting an ethical analysis might vary accordingly. For example, the more guiding authority the HTA organisation has, the more weight should be devoted to a balanced explication of the normative bearings of the recommendations. If the HTA organisation is clearly separated from decision-makers, it may be enough to describe the different values, attitudes and arguments that should be considered by the decision-makers. The “first” ethical question to emerge – whether to select a topic for HTA – might also be outside the scope of some HTA organisations. Furthermore, successful integration of ethical analysis into the HTA process depends on recognising its importance and aligning its processes with those of the entire HTA organisation, not carrying it out as an add-on to selected HTA projects (ten Have 2004). HTA organisations will need the appropriate skills, understanding and resources to do this. According to recent study, only 17% of Canadian HTA reports addressed ethical issues (DeJean 2009). Separate sections on ethical aspects were rare in the reports: instead, superficial remarks about possible ethical issues were more common, or ethical issues were raised but not solved. Further, use of ethical experts was rare.

Integration of ethical analysis may take various forms in HTA organizations. Some methods align well with the more traditional approach of conducting HTA, e.g. hiring a bioethicist to conduct a separate chapter on ethics, or conducting meetings for HTA researchers to reflect on the issues raised by their HTA project. Other initiatives are more challenging to the traditional HTA culture, e.g. developing “interactive” or “constructive” HTA processes that involve stakeholders’ participation.

Issues specific for (diagnostic technologies in) screening programmes

Screening technologies bring many ethical questions to participants, their relatives, health care system and the society as stated in the criteria for a screening programme (Wilson 1968 and the Danish council of ethics 2002). The condition sought should be a sufficiently important health problem both to the individual and to society to warrant considering allocating resources to a screening programme, but the decision to define a disease as an important health problem is of itself a value-laden one (Hofmann 2001). Ethical considerations will vary depending on whether the subject of the HTA is a diagnostic test used in primary or secondary screening. Primary screening deals with asymptomatic populations in which disease is possible if not actually yet suspected.  In secondary screening, the population has already come into contact with the healthcare system because symptoms have arisen.  In secondary screening for conditions with known adverse effects there may therefore be a greater imperative to identify and treat the condition, because the natural history of the disease, once it has been found, might dictate early treatment.  For primary screening, the test is being given to an asymptomatic individual and this raises significant ethical issues that are discussed further below. 

There are a number of considerations that govern the introduction of organised screening programmes.  Some bodies have criteria to determine the appropriateness of programmes being considered for introduction across the population (eg UK National Screening Programme criteria, criteria for screening programmes in Finland).  Such criteria can form a useful basis for the classification of issues to consider when initiating HTA on screening technologies.  Some of these considerations are now discussed in more detail.

Organised screening programmes are usually targeted at healthy individuals, and involve the health care system contacting an individual and proposing an intervention to prevent disease and promote health. This implies a special responsibility for the health care system; the effectiveness of the screening must be guaranteed as well as the treatment that follows if the patient is found to have the disease.  It increases the importance of clear and balanced patient information and decision aids in order to ensure informed consent to participation. The participants need to be well informed about the options they may face if the test is positive. Ethical analysis needs to be applied to the consequences of "false positive" and "false negative" test results as well as consequences of possible over-diagnosis and over-treatment have to be carefully evaluated and weighed against the expected benefits. Any of these may affect the medical, economic or legal status of individuals who participate in the programme.

There should be a suitable test or examination for screening, for which the following characteristics are known (eg UK national screening programme criteria):

  • validity of the testing system
  • sensitivity and specificity
  • predictive value of the test(s)
  • any concerns about safety or adverse events.

The screening test should be acceptable to the population. Where to set the limits for test accuracy (sensitivity and specificity) and who to include in the assessment of this (i.e. the acceptability to the population) are normative issues.  If the proposed screening is for a disease that the programme planners wish to have identified at a latent or early symptomatic stage, it will result in people who feel healthy learning that they have a “disease”. The natural history of the condition, including development from latent to declared disease, should therefore be adequately understood.

Equity of access is a further consideration. Some technologies may be expressly addressed to reduce inequalities (for example, self sampling HPV tests or mobile mammography), while other technologies may carry a risk of decreased equity of access, such as regionalised assessment or colonoscopy vs. faecal occult blood testing. Information materials may, in attempting to be scientifically correct, be too difficult to understand, and thus act as a barrier for less educated people. The evaluation should also consider whether participating in the screening programme might stigmatize the participants or the test positive individuals.

Ethical evaluation of a screening programme has multiple perspectives as it may encompass the health care system from primary to tertiary level. General and technology specific ethical issues and consequences for various stakeholders (e.g. participants, their relatives in case of hereditary disorders, various levels of health care organization, screening test providers, screening health care professionals) need to be identified both before and during the HTA process. For each stakeholder, possible consequences of proceeding with or refraining from the implementation of the screening technology have to be identified.

There may be different ethical considerations for “case-finding” and screening carried out with the intention of treating. If screening is being carried out with the purpose of finding patients who need treatment, there needs to be an effective treatment available for the condition being screened for, and a clear referral protocol for subsequent treatment (as measured on, for example, physiological or other characteristics which may be found by the test). The costs of both screening and subsequent treatment will form part of the HTA.

Methodology

Where to find information?

Issues requiring ethical analysis should be identified systematically at the start of the HTA but assessors and decision-makers should be prepared to consider relevant issues that arise at any point in the HTA process. Information and evidence required to carry out ethical analysis in HTAs of screening technologies may need to be gathered from a number of sources, using various procedures. These may include:

  • standard literature searching, which for ethical analysis will need to be carried out in a broader range of sources than for standard HTA;
  • expert opinion, elicitation and professional guidelines;
  • patient/service user opinion;
  • views of organisational stakeholders, for example, the health system within which the technology is to be used.

The information gathering phase may require several iterations, where previous phases identify new needs and questions that might then be answered from other sources (Figure 1). Thus, it may be useful to repeat some phases following new insights.

{{51-Eth1}}

Figure 1. Process of ethical analysis

Databases and search strategies

Evaluation of the principal questions about the technology, and the consequences of implementing or not implementing it are based on the information received from ongoing research on efficacy, safety, effectiveness and cost-implications of the technology. Organisations carrying out ethical analysis in HTA will need to consult a wider range of sources of literature than would normally be considered for scientific evidence on clinical effectiveness. Academic sources encompassing philosophy, particularly ethics, law and social sciences should be searched. Grey literature, including legal case law, books and other monographs may also be of interest. Information retrieval for ethical assessment is likely to require more hand searching than information retrieval for effectiveness assessments. If these sources do not contain suitable literature in relation to the screening technology under consideration, searching should be extended to include other related technologies with similar ethical challenges (see casuistry below). A suggestion for databases and MeSH terms that can be useful has been identified by Droste et al (Droste 2003). Droste et al (Droste 2011) propose a methodological approach to literature searching for ethical analysis in HTA.

Expert and stakeholder opinion

Discussions among the working group and with experts are effective in identifying important ethical issues related to the technology. The questions in the assessment elements table of this domain are a good starting point for discussions with experts and other stakeholders, but additional content-specific ethical issues or challenges may also be identified during the discussions.  Qualitative analysis of the expectations and fears of various stakeholders may reveal questions that cannot be identified by the content or methodological expert group or from the literature review. This information can be derived from stakeholder meetings or by conducting primary studies.

What kind of information is required?

The focus of the assessment, the specific questions to be answered, the study inclusion criteria, and the primary outcome points for the analysis of the consequences of implementing a technology are defined by the entire working group, and may be incorporated into a formal scope or decision problem document. These choices are value laden and they need to be carefully scrutinized before proceeding to literature review as they can have a major impact on the content and conclusions of the HTA report.

It is important to consider whether there are issues of potential ethical significance related to the disease or health problem, even before any factual considerations about consequences of implementing or not implementing the related technology. For example:

  • some conditions being screened for might be considered “self-inflicted”, and could lead to debates about access to treatment, such as liver damage arising from alcohol consumption where the patient needs a scarce donor organ;
  • prenatal screening tests may raise fundamental questions about the value of life and autonomy, and highlight competing interests of the embryo, mother, father, siblings or future possible siblings. 

Some issues in the Assessment elements table deal with the direct consequences of the implementation of a technology (e.g. can the technology harm the patient?). Others relate to questions of value that need to be addressed when deciding on implementation, such as the impact of the technology on availability of healthcare resources for different patient groups, or the balance of benefit and harm for the population as a whole. Competing ethical considerations generally do not lead to clear conclusions and therefore judgement must be applied by assessors and decision-makers.  Philosophical techniques such as deductive reasoning may be helpful in testing the logic and coherence of the arguments for stakeholders’ different viewpoints.

The perspective of all relevant stakeholders should be reflected in the process. It is usually fairly easy to identify the primary stakeholders for each technology - patients, clinicians, patient organizations, industry, providers etc. Making HTA project plans public as early as possible and allowing for public consultation may help identify relevant stakeholders and their fears early in the process. It is equally important to identify those stakeholders who will be indirectly affected if the technology is implemented, such as patient groups with competing interests in accessing healthcare resources. The views of stakeholders are best acknowledged early on in the process rather than during the external peer review process.

Ethical assessment and analysis

As we have seen, ethical analysis is an ongoing process that lasts throughout the HTA project. Ideally, many of the ethical and moral issues should be considered early on while still analysing other aspects of the technology.  The results and insights gathered for the other domains guide ethical analysis. However, the ethical analysis phase should add to the process in a way that the other domains cannot. For each Core HTA project there should be a person responsible for facilitating and reporting the ethical analysis. For a successful ethical analysis, it is necessary that it is done together with scientific and clinical experts. If expert ethical advice is available within the HTA organisation this resource should be used. If it is not available, it should be acquired if possible at least for the application of the methods.

Although there is wide consensus that ethical analysis should be a mandatory element of HTA, there has been no generally accepted, structured method for performing ethical analysis. Identifying and defining the various methodological approaches has been conducted by the INAHTA ethics working group (Andersen 2005). Most of them emphasise the need to consider issues extending past utilitarian maximisation of cost-benefits of technology.

The methods must be tailored to suit the HTA organisation, the topic under study as well as the local culture and health care system. Standard HTA practices such as evidence grading are redundant in this context. The choice of methods to conduct a formal analysis of ethical aspects depends on a number of interacting factors:

a)      The type of technology being assessed.  The following aspects determine the relative importance of ethical analysis in HTA:

-         The intervention is innovative, or appears to challenge commonly held values or societal beliefs.  For example, HPV-screening is seen by some groups as “promiscuity testing”; prenatal screening (PND) and preimplantation genetic screening (PGS) are considered to be offensive by some people with the conditions that are screened for (the so-called expressivist argument).

-         In cases where screening (encompassing diagnosis and treatment) of an individual may have an impact on the health or treatment of relatives.

-         In cases where there is a pre-treatment test to identify a responding subgroup (stratified or personalised medicine), which may lead to restricted access to treatment.

-         When there are uncertainties about safe use of the technology or the long-term outcomes of both the diagnostic and the subsequent therapeutic technology

-         In cases where the intervention predominately affects a group protected by equalities legislation.

-         The accuracy level of the diagnostic test

b)      The role and authority of the HTA organisation in the national decision making procedure. Decision making bodies and agencies providing guidance may have more explicit requirements for transparency for their stakeholders than academic or other bodies carrying out HTA.  They may also have legal duties requiring them to avoid discrimination and promote equality.  This may affect their approach to ethical analysis.

c)      The methodological expertise and experience with ethical analysis that are available within or to the organisation.

d)     Time and resource constraints for the assessment.

Methods for ethical analysis

The following approaches have been presented (and used) for ethical analysis in HTA.

Casuistry

Casuistry means solving morally challenging situations ("cases") by referring to relevantly similar "paradigmatic" cases for which an undisputed solution has been found (Jonsen 2001,2005, van Willigenburg 2005, Giacomini 2005).

The methodology of casuistry comprises three steps. First, the case at hand is sorted to a broad category of problems, "topics" (e. g. medical indications, patient preferences, quality of life, contextual features). Details should be described in a standardised way (who, what, where, when, why, how, by what means). Second, common sense moral rules, “maxims”, related to the case are explored (e.g. “the wish of the patient has to be respected"). If the maxims are contentious, the moral principles that underlie them in the case at hand are explored. Third, the case at hand is compared with a set of paradigmatic cases on the same topic that have been solved in agreement previously. Comparing the details of the case at hand, including the underlying maxims and principles, with the details of the paradigmatic case then may suggest a solution for the current problem (Neitzke 2005).

In HTA, especially for coverage decisions, a casuistic approach (precedence method) is suggested as at least a part of the ethical analysis. It means first establishing an inventory of past coverage decisions. The aim is to generate a typology of paradigmatic, covered technologies, which would represent the basic moral principles that underlie decision-making in the respective health care system. Next, the relevant qualitative and quantitative characteristics of the new technology are identified, and the technology is compared to similar, preceding paradigmatic cases. Ideally their solution may then be applied to the new technology. However, in addition to applying the solutions of past precedents to current cases, it is also necessary to reflect on the possibility that the value base has changed since the paradigmatic decisions were made. It may be that this reflection leads to a need to reconsider previous decisions.

In pure casuistry, cases are approached without referring to ethical principles, norms or theories. The process might resemble coherence analysis in that coherence between solutions to similar cases is searched for, or interactive approaches that aim for consensus of relevant stakeholders. A pragmatic, “moderate” form of casuistry as described above can include an element of principlism in that referring to ethical maxims and principles is done if comparison to previous cases does not provide clear enough solution. It also includes an element of wide reflective equilibrium, in that applying past precedents to new cases might reveal a need to reconsider previous decisions. 

Coherence analysis (CA)

The main idea of CA is to reflect upon the consistency of ethical argumentations or broader theories on different levels, without prescribing which facts, arguments or principles are prima facie relevant. It is a procedural, pragmatic approach, i.e. describes a procedure of approaching moral issues without claims of providing direct answers on “right or wrong”. CA can be compared to test-reliability and internal consistency of tests in empirical research. It cannot ensure validity: an immoral system can be as coherent as a morally justified one. (Grunwald 2004, Musschenga 2005).

CA considers the logical (possibly also emotional or intuitive) consistency of facts, norms and arguments relevant for the HTA. Thus CA is critically dependent on the material input, i.e. the comprehensive identification of facts, values and principles the coherence of which is to be considered.

Some kind of consideration of logical coherence is necessary for any ethical analysis of HTA. The more “extraordinary” the technology under evaluation is, the more useful a formal CA can be. 

For CA the evidence can be summarized in regard to

  1. society’s normative framework relevant to the technology (legislation, practice norms and guidelines, decision making procedures)
  2. society’s, patients' and scientists' expectations regarding the impact of the technology (fears, expectations)
  3. society’s general objectives and visions (concepts of justice, autonomy, reasonable development and other ideals)
  4. Interpretation of the past and present `biography´ of society or parts of it (deeply held, fundamental values and views central to individuals and societies self-image)

CA can be conducted by one expert or by a group. It is a reflective procedure (internal monologue / group discussion) trying to help achieve a logically consistent HTA. The identification of inconsistencies should lead to attempts to solve these (using, for example, discussions, wide reflective equilibrium, interactive technology assessment, normative approaches based on common principles etc.). Higher consistency of the whole is the norm, on which conflicting ideas are evaluated, edited and possibly abandoned. Thus and in contrast to interactive approaches (see below), opinions of important stakeholders can but need not be taken into account.

Reaching consistency might not succeed, so the end result might as well be identification of incommensurable beliefs or values, or contradictions between empirical claims, normative frameworks, or scientific and societal understandings and needs.

In conclusion, CA does not provide an unequivocal normative “ethical recommendation”, but CA is an essential part of all ethics analysis. It may be especially useful early on in the HTA process, to help identify central issues in need of further scrutiny.

Interactive, participatory HTA approach (iHTA)

iHTA aims for intersubjective consensus on ethically problematic issues, reached through real discourse It integrates patients, professionals and other stakeholders’ perspectives into HTA. It is a procedural approach (like coherence analysis) meaning that it describes a procedure to approach ethical problems, not any ideal solution to these problems. In contrast to coherence analysis, however, iHTA also aims to improve the validity of the whole HTA process through empowering and involving the stakeholders to participate. Although iHTA aims for consensus, this may not always be reached together with the stakeholders. It may also be decided that the conclusions are drawn from the stakeholder hearing by the method experts. (van der Wilt 2000, Reuzel 2004, McGee 1999, Habermas 1981, Skorupinski 2000).

The iHTA process begins by asking what kind of values are at stake, whose values they are, who are the important stakeholders and what values of theirs are at stake. Second, an interactive procedure to clarify these values is chosen, depending on presumed severity of value conflicts and the resources available. For example, the Delphi procedure, citizen juries, focus groups or deliberative polls could be used. The results of the interactive process inform the HTA process, i.e. help to identify relevant questions and relevant parameters to assess the (health) effects of the technology, but can also be reported as such.

iHTA informs, but does not dictate, the normative ethical conclusions needed in reporting the results of the HTA. The iHTA can bring into the expert group important opinions and values that may otherwise have been ignored. Ethical conclusions can not, however, be directly derived from any naturalistic population consultation: it is not possible to deduce how things ought to be from how things are. But the description of possibly differing valuations of different stakeholders, discovered with the iHTA process can be important for the application of the results.

Principlism

Principlism is based on the idea that there are principles, rooted in society, that are based on a common morality. These principles form a core dimension of all morals occurring in the world, and are presumed to be shared by every serious moral person. Principlism does not imply a specific method of reasoning, but describes a specific content of ethics: the principles form the essence of considered judgments. Principlism considers the validity of ethical analysis. (Beauchamp 2001, Vieth 2002).

Principlism recognises that there are several ethical principles, in contrast to foundational theories like utilitarianism or Kantian deontology that recognise only one supreme principle. The most influential principlist approach to bioethics (Beauchamp 2001) comprises four principles, representing clusters of practice norms:

  • Respect for autonomy: a norm of respecting the decision making capacities of autonomous persons,
  • Non-maleficence: a norm of avoiding the causation of harm,
  • Beneficience: a group of norms for providing benefits and balancing benefits against risks and costs - also referred to as the `proportionality principle´, highly relevant for HTA and research ethics and
  • Justice: a group of norms for distributing benefits, risks and costs fairly.

These norms are assumed to form a comprehensive analytical framework for bioethics. The principles are ‘prima facie’ binding, meaning that they are always important in every situation, but they are not absolute, because they can conflict. Highly relevant for HTA is, for example, the conflict between autonomy and beneficence for single persons on the one hand, and the just distribution of resources and beneficence for society on the other.

In practice, as the principles are abstract, they must always first be specified according to the current context. Then, if all principles can not be realised fully (as is most often the case), the specified principles must be balanced with each other. A principle should only be overridden if:

  • Better reasons can be offered to act on the overriding one,
  • The moral objective which justifies the infringement must have a realistic chance of being achieved,
  • The infringement must be the only way to realize one principle at the cost of the other,
  • The form of the infringement must be commensurate with achieving the primary goal,
  • Any negative effects of the infringement must be minimized and
  • The decision must be impartial in regard to all affected parties.

The major advantage of principlism is that it delivers a comprehensive, normative framework for ethical analysis, in contrast to procedural, non-normative approaches like CA, iHTA, wide reflexive equilibrium and casuistry. Conversely, normativity is also the main problem of principlism, as not all ethicists agree in that these and only these principles are universal. If so, the normative framework of four principles might not be valid for every technology and every population.

Explicit principlistic considerations are useful for increasing the transparency and transferability of the ethical analysis. To balance the principles in a context-sensitive manner in practice, WRE (see below) or participatory methods can be useful.

Social shaping of technology

The social shaping of technology (SST) approach (Rip 1995, Clausen 2004, Reuzel 2004) views technology as the product of societal processes (within industry, research institutes, governmental bodies, and society at large) rather than an independent artefact that has a certain, measurable impact on its target. The aim is to understand what technology is and how its development is interwoven with its social context (e.g. the engagement and strategies of various actors, and the way various problems are defined and resolved).

Assessing the role, merit, and value of technology becomes important. The social shaping perspective also implies an opportunity to manage technology through its social context. If technology in fact is technology-in-context, then both technology and its context can be influenced or adjusted to improve the outcomes of using technology. The societal processes underlying technology development can be explained to some extent by the values relevant in different contexts.

From the ethics point of view, the SST approach emphasizes

a)      reflexive focus on the range and values of relevant actors and their conditions of involvement

b)      considering how technology can influence society and how technology can be best managed by society

c)      the inadequacy of evaluating a technology without considering the local social environment.

Within this framework, many of the other methodological approaches to ethical questions in HTA can also be applied (e. g. participatory approaches such as iHTA).

Wide reflective equilibrium (WRE)

The WRE (Rawls 1971, 1993, Daniels 1979, 1996) is an ideal, perpetual goal of justification in modern philosophical inquiry. It is based on pragmatism and social constructivism, which claim that ethical truths can not be revealed or directly experienced, and that there are no static, fundamental a priori valid universal principles. On one hand, the normative framework of society may change over time. On the other hand, humans need stability, cognitive coherence and some degree of reconciliation of individual and social norms and values. WRE is a central methodological part of the ‘four principles’ approach, discussed above (Beauchamp 2001).

When using WRE, the reflection starts from the most considered judgments and moral feelings that have a prima facie credibility. This has to be done behind a ‘veil of ignorance’ (i.e. imagining we do not know which position we would have in the society our decisions concern) to try to be as impartial as possible.  To approximate WRE, all possible situations, arguments, and judgments need to be taken into account and brought into a coherent whole through rational reflection (see coherence analysis above). This might entail that some of our primary considered judgments have to be adjusted.

WRE is an important political and philosophical goal of coherence analysis and discourse ethics in regard to decision making. However, it is an ideal goal of a theoretical procedure, which may be difficult to apply in real-world HTA processes. As a goal emphasizing individual and inter-subjective consensus, WRE may also neglect true conflicts between incommensurable arguments. Essentially, WRE emphasizes open, honest and impartial discourse, conducted by rational, sensible actors in democratic, pluralistic societies who want to reach consensus through finding the most validity of claims.

The “triangular model” for ethical analysis based on human person - centred approach

The triangular model is centred on a substantial conception of human person. It considers the man as reference-value in the reality, around which all the ethical judgements are coordinated. Based on a cognitivist approach to the ethics, this model considers that it is possible to get some truths, concerning man and his/her praxis, recognizable by everyone through a rational activity. (Sgreccia 2007).

The methodology of the triangular model comprises three steps of analysis: 1. data collection; 2. anthropological aspects, 3. ethical-normative evaluation. The first step, “scientific moment” consists of an in-depth study of all facts/data, including qualitative and relational ones. The second step, “anthropological moment”, consists of the anthropological understanding of facts; in other words, the analysis of eventual values at stake, related to human life, integrity and dignity. According to this analysis it is possible to find values which should be promoted and defended, and norms which should guide human action on individual and societal levels. The third, “ethical-normative” step consists of evaluation of practical choices that should be made.

This model highlights a triangular connection between bio-medicine, anthropology and ethics, settled on two levels: the explanation of a certain topic (descriptive step), followed by a normative phase, in which we can get conclusions within a debate of the meta-empirical perspectives i.e. relating to the steps 2 and 3 described above. It is evident that such an ideal process needs all three theoretical steps in order to be possible.

This model presumes a normative framework for ethical analysis (Sacchini 2005, 2007). It consists of four principles of reference: 1) the defence of human physical life as a whole, and its integrity; 2) the principles of freedom (capability of the human will) and responsibility (anintra-and inter-subjective evaluation of subject’s own acts and will); 3) the therapeutic principle, according to which the human person has to be treated as a whole of body-mind reality; 4) the principles of sociality and subsidiarity, according to which public or private authority is called to intervene and to help the person only if he is not able to manage, to promote or safeguard him/herself (Sgreccia 2007).

Axiological (Socratic) approach

The axiological approach is based on the insight that science and technology is a social activity governed by a wide variety of norms and values. Health technology is applied in a social setting where there is interplay of different kinds of norms and values, HTA should highlight and address the norms and values involved in the implementation and use of a health technology. The reason why it is also called a Socratic method, is because it is based on a set of questions which are aimed at highlighting normative issues in the HTA as well as in the decision making process.

The (32) questions relate to:

  • General moral issues, such as integrity, human rights, patient autonomy, benefit, harm, respecting social and religious convictions
  • Moral issues related to stakeholders (patients, relatives, health care providers, industry, policy makers)
  • Moral issues due to methodological challenges (end-point selection, quality assessment of study design)
  • Issues typical to the technology (function, purpose, intention)
  • Moral issues related to the HTA process itself.

The axiological/Socratic approach consists in six steps (Hofmann 2008).

  1. Identify and analyze the moral challenges that are typical for the health technology.
  2. Identify stakeholders.
  3. Select a set of morally relevant questions by selecting from a list of questions (Hofmann 2005a; 2008) which highlight value issues in regard to the implementation of health technology. Justify the selection.
  4. Perform literature search on the basis of the steps 1-3.
  5. Analyze the selected questions (in step 3) on the basis of the literature search (step 4), hearings with stakeholders, and results from qualitative research.
  6. Summarize the analysis and highlight the most important value issues.

The aim with addressing norms and values through the set of morally relevant questions is to provide an open, transparent and informed decision making process.

The axiological/Socratic approach has been applied to bariatric surgery (Hofmann 2010), newborn screening (Vist et al 2007; Heiberg 2009; Hofmann 2010), HPV-vaccine (Hofmann 2008; 2009), welfare technology (Hofmann 2008;2009), palliative surgery (Hofmann et al 2005), obstibation treatment in cancer care (Movik et al 2009), ICSI (Holte et al 2007), amalgam replacement (Håheim et al 2006), autologous stem cell transplantation in advanced breast cancer (Droste 2011), and other technologies. Moreover several HTAs include subsets of the questions in the axiological approach (DeJean et al 2009).

Examples of local application of these and other methods, see Appendix 2
Qualitative synthesis

The methods described above can be used to guide the elicitation of information, but their main use will be in presenting, analysing and balancing that information so that conclusions may be drawn, and the presentation can be used by decision-makers. The core set of questions in the Assessment element table is intended especially for identifying ethically relevant issues. The morally relevant issues and moral conflicts have to be synthesized and reported transparently so that they can be considered when deciding whether to implement a technology. No single solution to every moral problem exists; neither is it possible to list moral issues according to a commonly agreed weighted value. Answers to the core set of issues may also reflect the variation in morals and values found within most societies. The synthesis of ethical analysis has to be performed in an open way so that the interests of various stakeholders are kept as "unweighted" as possible, or the weighing is done transparently i.e. describing the procedure and participants of the analysis. Ideally, the decision on "whose values are to be weighted" need to be in the hands of the decision makers. The decision makers can be different both within the same country between technologies and / or institutions and also between countries. Thus the ideal way to present the synthesis of the analysis may vary accordingly.

Ethical analysis on the consequences of implementing or not implementing the technology may be handled using an open framework (Autti-Rämö 2007). The possible consequences of proceeding with or refraining from the implementation of the technology can be listed separately for each stakeholder in an open table as the answers for various parties may differ largely (table 1). The identified issues are not valued-weighted against each other but the table offers a transferable list of aspects that need to be appreciated in the final decision making process.

Table 1. Example of a framework for ethical analysis

Stakeholder

Benefits when proceeding with implementation

Adverse consequences when proceeding

Benefits when refraining from implementation

Adverse consequences when refraining

Patient

    

Family

    

Healthcare organisations (ie the organisations that own hospitals and provide healthcare)

    

Other patient groups within the specialty

    

Primary Health care providers

    

Secondary Health care providers

    

Tertiary health care providers

    

Non-governmental organisations (NGO) representing patients who need the technology

    

NGOs representing patients needing another technology which is withdrawn due to implementing the technology in question

    

Payers

    

Society

    

Producers/Industry

    

Decision makers

    

HTA organisations

    

It is important to identify also those areas where values may differ significantly between the various stakeholders (eg. attitude towards the care of patients with non-treatable diseases, treatments of extreme cost or conditions perceived as ‘self-inflicted’). The main areas of ethical controversy and competing interests should be clearly stated in the final document.

Reporting and interpreting

The results of the ethical assessment or analysis will usually be reported as a separate chapter, in order to assure transparent reporting of value issues. The ethical implications of implementing or refraining from the implementation of technology need, however, to be discussed in a balanced way so that the health policy makers have a wider view on all possible consequences of their decision. The open framework as presented in table 2 can be a helpful tool in this process. The decision to implement a new technology requires careful decision on the balance between benefit and harm, cost-effectiveness, reallocation of resources etc. Discussing the context-specific moral issues within the respective chapter (e.g. effectiveness, safety, and costs) may thus also help the decision makers to identify various scenarios and set them out transparently.

Overlap with legal and social aspects

The results of the ethical assessment or analysis closely relate to the evaluation of legal and social aspects, although Duthie and Bond (2011) argue they should be clearly distinguished from one another. These domains overlap the ethical analysis, though the angle of evaluation may differ. The legal framework forms a basis for professional ethics (e.g. abortion, prenatal screening, and euthanasia). The social consequences of implementing a technology may differ largely from those of primary outcomes at patient level (f.i. avoidance of death at patient level, avoidance of impaired working ability at societal level). The implementation of new technology will not only have an effect on health, functional abilities and psychosocial well-being but also on social networks and need of support.

Transferability of ethical analysis

The ethical assessment or analysis and its outcome have to be described in an open way in order to judge their transferability. Many of the ethical implications are common to various nations but some value laden issues are likely to be country specific, and will crucially relate to factors such as the ‘social contract’, the funding system used for the country’s healthcare and the country’s GDP growth prospects. Analyses relating to ethical principles, coherence or paradigmatic cases are likely to be more easily transferable than argumentation based on interactive approaches relying on local values, stakeholder attitudes and available health care resources.

Assessment elements

F0001 Assessment element card

Issue: Is the technology a new, innovative mode of care, an add-on to or modification of a standard mode of care or a replacement of a standard?

Topic: Principal questions about the ethical aspects of technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The consequences of totally new screening programmes are likely to be more difficult to predict than the consequences of changing methods within an existing screening programme (breast screening and digital imaging), for individual values, attitudes and expectations as well as for health care systems. Novel screening programmes (screening for rare metabolic disorders in newborn), improved specificity of screening methods (ultrasound for fetal abormalities), or totally new screening tests (screening for maternal drug and alcohol abuse from hair or meconium) - may have far-reaching consequences on health care. They may require more emphasis on ethical analysis than replacing a test already in use with another testing the same diagnostic marker, although the literature and research base on the topic may be narrow.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

DTC, Organisational.

Sequential relations

F0002 Assessment element card

Issue: Can the technology challenge religious, cultural or moral convictions or beliefs of some groups or change current social arrangements?

Topic: Principal questions about the ethical aspects of technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It is important to identify those groups within the society for whom the use of the technology may pose serious challenges due to their beliefs, convictions or current social arrangements. Finding other acceptable possibilities for these groups is important. Identifying the conceptions behind the beliefs and values may help put them in perspective, when considering the overall acceptability of the technology. Technology may also change generally accepted social arrangements by challenging traditional conceptions (e.g. screening for fetal abnormalities and on the other hand the concept of “design babys” through development of preimplantation diagnostics).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Ogletree 2004

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social

Sequential relations

F0003 Assessment element card

Issue: What can be the hidden or unintended consequences of the technology and its applications for different stakeholders.

Topic: Principal questions about the ethical aspects of technology

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The technology may be used for other purposes and have side-effects in addition to those following from the intended use. E.g. screening for fetal abnormalities may give information on gender. Unintended consequences may be difficult to predict (eg abortion due to unwished gender), but the intended purpose and uses of the technology should be evaluated against the likely uses and consequences of the technology. New technologies give rise to new ethical questions (e.g. screening for metabolic disorders in newborn with non-existing early treatment options). As pre-symptomatic screening tests have become available, the health care system has to be prepared to handle moral issues raised by true positive and false negative findings. Screening positive and being diagnosed with the disease may have effects on relatives as a all diagnoses of hereditary disorders, also provide knowledge of relatives. Screening positive may also affect social relations. In screening programmes by definition diagnostic information necessitates further action, so all screening programmes may have large impact on the health care processes and systems and on individuals. They may even change the concepts of disease if the definition of whom to treat as a patient is unclear (eg screening for aorta aneurysm).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Ogletree 2004, Hofmann 2005b, Hofmann 2002b

Content relations
Sequential relations

F0004 Assessment element card

Issue: Does the implementation or use of the technology challenge patient autonomy?

Topic: Autonomy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Patients have in most cases a right to autonomy. This means both the right to decide, but also right to relevant information. The information should enable understanding the issues, enable considering it in relation to personal values, and deciding accordingly. Screening programmes represent complex technologies that may be difficult to be understandably explained to the patient (e.g.meaning of screening positive or negative and the possible risks associated with diagnostic tests and/or treatment) , as are screening programmes that require patients to behave in a certain way (e.g.dietary restrictions for fecal blood test). The practical challenge with screening programmes is that in order to be fully autonomous, the participating person should understand all alternatives following different test results and be able to make informed consent at every step.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Miller 2004

Content relations
Sequential relations

F0005 Assessment element card

Issue: Is the technology used for patients/people that are especially vulnerable?

Topic: Autonomy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The right and justification to use the technology for persons who are vulnerable (critically ill or have otherwise reduced decision making capacity, like children, mentally retarded, patients that have due to their illness/state limited decision making capacity, pregnant women etc) has to be clarified. Who has the right to balance the benefit against possible harm in these situations? On what grounds can these decisions be made? Is the technology so valuable, as to justify its use on people who cannot give informed consent to it?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Miller 2004

Content relations
Sequential relations

F0006 Assessment element card

Issue: Can the technology entail special challenges/risk that the patient/person needs to be informed of?

Topic: Autonomy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Is the common professional practice of discussing the technology with patients enough, or is special care needed with this technology? Should the patient be explicitly informed, for example, that false positive results may lead unnecessary further investigations and treatments with serious harms? Screening programmes to be used for early identification of life-threatening situations may have life-threatening side effects (e.g. treatment is invasive surgery with risk of death). Technology used to get exact diagnostic information for those screening positive may have unexpected severe side-effects (e.g. miscarriage due to amniocentesis).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Registers

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Miller 2004

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety

Sequential relations

F0007 Assessment element card

Issue: Does the implementation challenge or change professional values, ethics or traditional roles?

Topic: Autonomy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technologies may change the relationship between physician and patient, challenge professional autonomy or otherwise interfere with professional ethics and values. The patient-physician relationship is traditionally based on mutual trust, confidentiality and professional autonomy so that individual treatment decisions can be made in the best interest of the patient. Technologies that interfere with core virtues and principles of medical and professional ethics challenge the professional integrity of the physicians or other health care professionals (eg. screening for drug abuse when use is denied). Technologies that align with professional ethics are more likely to be implemented successfully. For example, people may require a test or intervention for many reasons, even if the professionals think them unnecessary and potentially harmful (eg whole body MRI scans).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Expert opinion

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Hofmann 2005b. Medical Professionalism Project 2002

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

DTC, Organisational.

Sequential relations

F0008 Assessment element card

Issue: Does the implementation or use of the technology affect human dignity?

Topic: Human Dignity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Especially technologies that are applied for persons with reduced autonomy may violate a person's dignity (children, mentally impaired, severely ill), i.e. challenge the idea that all human beings have intrinsic moral value, and should thus not be seen as means to others ends. Labelling people may also threaten their dignity (eg. screening children for fetal alcohol spectrum disorders). Some screening tests may label healthy people as sick (eg PSA for prostate cancer) or otherwise less worthy (screening for a non-dominant gene defect in fertile aged, screening for STD in school aged girls). Handicapped people may be labelled by prenatal screening programmes which imply that their handicap is an indication for abortion.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kilner 2004

Content relations
Sequential relations

F0009 Assessment element card

Issue: Does the implementation or use of the technology affect human integrity?

Topic: Human integrity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technology can challenge human integrity by preventing (or even tempting) people (patients or professionals) to live according their moral convictions, preferences or commitments. This is especially important for vulnerable patient groups. Integrity can also be seen as a coherent image or identity of oneself. Institutions that discourage honesty or ethical conduct are detrimental to integrity; for example, systems where lying about one’s health state might lead to better treatment than being honest. Prenatal screening programmes might challenge the integrity of people who value new life as gift; screening for cervical cancer and/or HPV may be problematic for some religious groups.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kilner 2004

Content relations
Sequential relations

F0010 Assessment element card

Issue: What are the benefits and harms for patients, and what is the balance between the benefits and harms when implementing and when not implementing the technology? Who will balance the risks and benefits in practice and how?

Topic: Beneficence/nonmaleficence

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The decision to implement a technology requires careful decision on the balance between benefit and harm, cost-effectiveness, reallocation of resources etc. When this decision has been made on the system level, the decision on individual patient level rests on both the professional who offers the technology and the patient who autonomously accepts to participate at every possible step. The individual decision has to be based on objective information on possible benefit and risks. Risks are only justified to the extent they are needed to create benefits. If not proven otherwise, the individual patient is generally to be seen as the best judge of risks and benefits for her/himself.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Autti-Rämö 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety.and Effectiveness

Sequential relations

F0011 Assessment element card

Issue: Can the technology harm any other stakeholders? What are the potential benefits and harms for other stakeholders, what is the balance between them? Who will balance the risks and benefits in practice and how?

Topic: Beneficence/nonmaleficence

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Some technologies have the potential to unfold unwanted or harmful effects not only on the patients that the technology is directly applied to but also indirectly on other stakeholders (relatives, other patients, organisations, society etc.) Benefits and harms to individuals must be balanced with benefits and harms that can befall society as a whole (social utility, maximizing public health). These harmful effects may manifest in the physical, social, financial or even other domains of life. For example results of prenatal screening and screening for metabolic disorders in newborn may negatively interfere with the family planning and social life of not only the individual being tested but also of his or her relatives. Changes in the availability of treatment facilities may significantly alter the requirements placed on the health care system.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Autti-Rämö 2007 Beauchamp and Childress 2001

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational, Social

Sequential relations

F0012 Assessment element card

Issue: What are the consequences of implementing / not implementing the technology on justice in the health care system? Are principles of fairness, justness and solidarity respected?

Topic: Justice and Equity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

A new intervention may require reallocation of human resources, funding and training. A large reallocation of resources may seriously jeopardize other patient groups. How this reallocation affects the existing health care system has to be studied for all stakeholders? Can the technology be applied in a way that there is equal access to those in equal need and who would equally benefit for the programme? How can this be guaranteed? Could potential discrimination or other inequalities (geographic, gender, ethnic, religious, employment, insurance) prevent access? Are specific safeguards needed? How will possible caregivers’ burden and well-being be influenced? Potential inequalities and discrimination should be justified. Screening technologies sometimes acquire significant symbolic value (e.g. fetal ultrasound, PSA) that may create demands for tests that are not justified on health grounds.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Sterba 2004 Daniels 2001

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Cost-effectiveness. Organisational. Social

Sequential relations

F0013 Assessment element card

Issue: How are technologies presenting with relevantly similar (ethical) problems treated in health care system?

Topic: Justice and Equity

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Clearly presenting how relevantly similar technologies are treated in a health care system may help to adopt coherent and just health policies, either by applying past precedents to current cases, or showing that past cases need reconsideration. Similarity is to be defined individually for each technology. The idea is to concentrate only on the similarities relevant for solving the ethical problems found important for the current HTA project. The similarity may be, for example, of medical, technological, economical, ethical, social, organisational or legal nature.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Littrature search. Expert opinion

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Hofmann 2005b

Content relations
Sequential relations

F0014 Assessment element card

Issue: Does the implementation or use of the technology affect the realisation of basic human rights?

Topic: Rights

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Human rights exist both in ethics and legislation, most notably in the United Nations declarations and related statements, like the European Council Biomedicine convention. Basic human rights are universal and consider the most important goods, protections and freedoms. Classes of rights are civil and political rights, social rights, minority and group rights and environmental rights. For HTA, perhaps the most relevant are the rights to equality, non-discrimination, safety, adequate standard of living and health care. For example: -Right to life, liberty and security of person. -Right to a standard of living adequate for the health and well-being of himself and of his family, including medical care and necessary social services, and the right to security in the event of sickness, disability or old age. -Right of the child to the enjoyment of the highest attainable standard of health and to facilities for the treatment of illness and rehabilitation of health. For screening programmes, issues of access to screening and diagnostic tests and treatments as well as labelling and potential discrimination of diagnosed persons may be relevant issues.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search. Law, rules and regulations. Expert opinion. Stakeholder hearing

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Marks 2004

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social. Legal

Sequential relations

F0016 Assessment element card

Issue: Is legislation and regulation to use the technology fair and adequate?

Topic: Legislation

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Technology may lead to ethical problems that make current regulation inadequate. Screening and diagnostic technologies are commonly differently regulated than treatments, especially medications. Ethical reflection is needed when considering what kind of regulation is needed. This consideration is done on the basis and in combination with the legal domain. Emphasis should be put on considering the ethically relevant aspects and consequences of current law, needs for legal regulation that have arisen from the ethical analysis, and a global assessment of the adequacy of the legislation based on all available information. For example, who has a right to get the results and for what purposes? Is legislation needed to ensure equal access? What kind of rules and regulations are needed to ensure good quality of high risk diagnostic tests and treatments.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Law, rules and regulations. Stakeholder hearing. Expert opinion

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Capron 2004

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Legal

Sequential relations

F0017 Assessment element card

Issue: What are the proper end-points for assessment and how should they be investigated?

Topic: Questions about effectiveness and accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The acceptable and feasible endpoints must be carefully considered early in the analysis. The context must be especially considered; some technologies require extensive interpretative skills, and sometimes the consequences will depend on the target population. This is especially true in disorders related to life style. The importance of context relates to what kinds of studies are deemed acceptable. For diagnostic technologies and screening programmes, clinical effectiveness – improved health outcome -should ideally be directly investigated. This is not always possible so other endpoints may have to be used. In addition, screening programmes may have several aims (e.g. screening for hearing disorder in newborn - early institution of therapy and possibility for cochelar implant).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Other domains of analysis: accuracy, safety, effectiveness. Expert opinion

References
Content relations
Sequential relations

F0018 Assessment element card

Issue: Are the accuracy measures decided and balanced on a transparent and acceptable way?

Topic: Questions about effectiveness and accuracy

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Are the accuracy measures (sensitivity and specificity) chosen so that they accord with the purpose of the HTA? How and by whom are cut-off values decided? How and by whom has balancing sensitivity and specificity been done? This should be done considering the moral value of different results – for example, high specificity is required if false positives have serious consequences.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Other domains of analysis: accuracy, safety, effectiveness. Expert opinion

References
Content relations
Sequential relations

References

Andersen S, Caron L, Langavant GC, Dundorp W, Hagenfelt K, Hofmann B, Lopez-Argumedo M, Noorani H. INAHTA’s Working Group on Handling Moral issues. INAHTA Report, June 2005.

Autti-Rämö I, Mäkelä M: Ethical evaluation in health technology assessment reports: An eclectic approach. IJTAHC 2007:23:1-8.

Beauchamp T, Childress J. The Principles of Biomedical Ethics. Oxford. 2001.

Clausen C, Yoshinaka Y (2004) Social shaping of technology in TA and HTA. Poiesis Prax 2(2,3):221–246.

Daniels, N (1996): Wide reflective Equilibrium in Practice, in Sumner L, Boyle J (ed.), Philosophical Perspectives on Bioethics, Toronto, Buffalo, London, pp. 96-114.

Daniels N (1979): Wide reflective Equilibrium and Theory Acceptance in Ethics. Journal of Philosophy, 76, 5:256-282.

DeJean D, Giacomini M, Schwartz L, Miller FA. Ethics in Canadian health technology assessment: A descriptive review. International Journal of Technology Assessment in Health Care, 25:4 (2009), 463–469.

Droste S, Herrmann-Frank A, Scheibler F, Krones T. Ethical issues in autologous stem cell transplantation (ASCT) in advanced breast cancer: A systematic literature review. IJTAHC (In press).

Droste S, Dintsios CM, Gerber A. Information on ethical issues in health technology assessment: how and where to find them. Int J Technol Assess Health Care 2010; 26(4): 441-449. Additional supplemental material publishers web page.

Droste S, Gerhardus A, Kollek R. Methoden zur Erfassung ethischer Aspekte und gesellschaftlicher Wertvorstellungen in Kurz-HTA-Berichten – eine internationale Bestandsaufnahme. DAHTA@DIMDI Köln: DIMDI, 2003.

Duthie K, Bond, K. Improving ethics analysis in health technology assessment.  International Journal of Technology Assessment in Health Care, 27:1 (2011), 64-70.

Giacomini M (2005) One of these things is not like the others: The idea of Precedence in Health Technology Assessment and Coverage Decisions. The Milbank Quarterly 83(2). Available at http://www.milbank.org/quarterly/8302feat.html

Grunwald A. The normative basis of (health) technology assessment and the role of ethical expertise. Poiesis Prax 2004; 2: 175-193.

Habermas, J (1981): Theorie des kommunikativen Handelns. Band 1, Handlungsrationalität und gesellschaftliche Rationalisierung / Band 2 Zur Kritik der funktionalistischen Vernunft. Suhrkamp, Frankfurt a.M.

Heiberg, Arvid; Frydenberg, Karin; Grønvik, Oddrun; Hofmann, Bjørn; Klingenberg, Claus Andreas; Markestad, Trond; Martiniussen, Marit; Pettersen, Folf Dagfinn; Rogne, Sissel. Reccomendations on extension of newborn screening. Oslo: Directorate of Health, 2009. In Norwegian.Hofmann B. Toward a procedure for integrating moral issues in health technology assessment. Int J Technol Asess Health Care 2005; 21:312-318.

Hofmann B. Nyfødtscreening – mer skjult tvang? [Newborn Screening: More covert enforcement?] Tidsskrift for Den norske legeforening 2010; 130(3): 291-293.

Hofmann B. Addressing moral aspects in implementing prohylactic vaccines against human papilloma virus (HPV). Annals of the Academy of Medicine, Singapore 2009; 38(6): S70-1.

Hofmann B. Ethical aspects of HPV vaccination. Oslo: Norwegian Knowledge Centre for the Health Services, 2008. In Norwegian with English summary.

Hofmann B. Ethics in the assessment of health technology [In Norwegian with an English summary:]. Oslo: Norwegian Knowledge Center for the Health Services, 2008.

Hofmann B. Why ethics should be part of health technology assessment. Int J Techn Assess Health Care 2008;24:423-429.

Hofmann B. Håheim LL, Søreide JA. Ethics of palliative surgery in patients with cancer. British Journal of Surgery 2005; 92(7): 802-9.

Hofmann B. On value-judgements and ethics in health technology assessment. Poiesis & Praxis: International Journal of Technology Assessment and Ethics of Science 2005; 3(4): 277-295.

Hofmann B. Complexity of the concept of disease as shown through rival theoretical frameworks. Theoretical Medicine and Bioethics 2001; 22(3): 211-37.

Holte, Therese Opsahl; Hofmann, Bjørn; Lie, Rolv Terje; Norderhaug, Inger Natvig; Romundstad, Pål; Sæterdal, Ingvil; Ørstavik, Karen Helene; Tanbo, Tom. Male infertility: intracytoplasmic sperm injection. Oslo: Norwegian Knowledge Centre for the Health Services, 2007. In Norwegian with English summary.

Håheim LL, Dalen K, Eide R, Karlsson S, Lygre GB, Lyngstadaas SP, Nordmo AM. Health effects from removing amalgam fillings with suspicion of or health related injuries from amalgam. Oslo: Norwegian Knowledge Centre for the Health Services, 2006. In Norwegian with English summary.  

Jonsen AR (2005) Casuistical Reasoning in Medical Ethics. In: Düwell M, Neumann JN: Wieviel Ethik verträgt die Medizin? mentis, PaderbornMeth, p: 147-164.

McGee, G (1999): Pragmatic Bioethics. Nashville, London.

Movik, Espen; Ringerike, Tove; Linnestad, Kristin Kamilla; Hofmann, Bjørn; Harboe, Ingrid; Gjertsen, Marianne Klemp. Assessment of metylantrexon on obstibation in cancer treatment. Oslo: Norwegian Knowledge Centre for the Health Services, 2009. In Norwegian with an English summary.

Musschenga AM (2005): Empirical Ethics, Context –Sensitivity, and Contextualism. Journal of Medicine and Philosophy, 30:467-490.

Neitzke G (2005) Was ist der Fall? Argumente für eine Zuspitzung der kasuistischen Methodik. In: Düwell M, Neumann JN: Wieviel Ethik verträgt die Medizin? mentis, PaderbornMeth, p: 211-224.

Rawls, J (1971): A Theory of Justice. Harvard University Press, Cambridge.

Rip A, Misa TJ, Schot J (eds) (1995) Managing technology in society: the approach of constructive technology assessment. Pinter, London.

Sacchini D., Refolo P., L’Health Technology Assessment (Hta) e i suoi aspetti etici, Medicina e Morale 2007, 1: 101-139.

Sacchini D., Spagnolo A.G., Minacori R., Carrasco de Paula I., HTA and ethics: the framework of ethical positions and the proposal of a person-centred model, Italian Journal of Public Health 2005, 2 (2): 304.

Sgreccia E., Manuale di Bioetica. I. Fondamenti ed etica biomedica, Milano: Vita e Pensiero, 2007 (4.a ed.).

Skorupinski, B, Ott, K (2000): Technikfolgenabschätzung und Ethik. ETH-Hochschulverlag, Zürich.

Ten Have H. Ethical perspectives on health technology assessment. Intl J Technology Assessment in Health Care, 2004; 20(1): 71-76.

The Danish council of ethics. Screening - a report.: 1-188. Copenhagen 2001.

van Willigenburg T (2005) Casuistry in Medical Ethics. In: Düwell M, Neumann JN: Wieviel Ethik verträgt die Medizin? mentis, PaderbornMeth, p: 165-178.

Vieth, A (2002): Defending Principlism well understood. Journal of Medicine and Philosophy, 27:623-51.Vist GE, Frønsdal KB, Johansen M, Hofmann B, Fretheim A. Health effects of screening for metabolic diseases in newborn. Rapport nr 22-2007. Oslo: Norwegian Knowledge Centre for the Health Services, 2007. In Norwegian with English summary.

Wilson JMG, Jungner G. Principles and practice of screening for disease. WHO. Geneve 1968

Appendices

Appendix 2: Examples of local approaches to ethical analysis

AETMIS: Promoting context-specific, integrated approaches to analysing ethical issues in HTA

At AETMIS the ultimate objective is to integrate a context-sensitive ethical inquiry right from the beginning of the HTA (Caron 2005, 2006). Several approaches were developed for different HTA needs that apply at different times in the process of HTA:

  • “Start-up” meetings, which is an institutional process to promote context-based, ethically-informed HTA projects. These are conducted at the very beginning of selected HTA projects;
  • The “comprehensive” ethical approach, where ethical inquiry is an integral part of the evaluative framework. This means that ethical inquiry is “active” throughout the entire HTA process. Such approach is only used for specific HTA reports (e.g. genetic testing); and
  • The more traditional ethical analysis, which refers to the write up of a seperate section on ethical issues in an HTA report. Such “add-on” ethical inquiry is usually performed by an ethical expert in collaboration with the assessors.

Integration of ethical analysis throughout the entire HTA process is achieved by teaming a bioethicist with the assessment team responsible for the project. The assessment team can also be advised by a technology–specific advisory committee (e.g. for genetic testing). An “integrated” ethical inquiry involves a reflection on value-laden choices at all levels of the HTA process, namely in: a) defining the scope of assessment, b) performing literature review and primary research to document the experience of patients and their families as well as the context of service delivery, c) establishing a framework for appraisal of technologies and modes of intervention, d) conducting the appraisal of those strategies, e) highlighting specific ethical and social issues, and f) formulating recommendations. In addition to literature review, primary research can be conducted to better document the situation in the local jurisdiction, and to explore the perspectives of different stakeholders on the various issues linked with technology use. Ethical and social considerations pertaining to technology use are also documented in a specific section of the HTA report.

The eclectic approach of FINOHTA

In Finohta, each HTA report is produced in co-operation with the methodological experts from Finohta and clinical experts from health care organizations (Autti-Rämö and Mäkelä 2007). Professional ethicists are included either during the HTA or peer review process depending on the technology to be evaluated.

General and technology specific ethical issues and consequences for various stakeholders are identified during the HTA process by the content experts, through literature search and (when possible) by stakeholder hearing. For each stakeholder, a) possible consequences of proceeding with or b) refraining from the implementation of the technology (as compared with other options) are listed. Including patient representatives is an option in this process.

A repetitive exchange of opinions and weighing different values has been the core of a successful ethical discussion and when making a summary of the evaluation process. New moral issues often emerge during the HTA process and novel aspects have come up even at the final comment round. Ethical evaluation is written as a separate chapter in Finohta reports, but its main aspects are interwoven in the discussion chapter so that evidence is balanced against ethical consequences.

Value analysis of NKCHC

This method is used at the Norwegian Knowledge Centre for Health Services (NKCHC) and it is based on value analysis (axiology) developed with regard to technology, according to which technology is a part of human activity that is related to values in different ways (Hofmann 2002, Hofmann 2006):

  • Function (value-ladenness, e.g. visualizing extracorporeal structures by ultrasound for a diagnostic ultrasound machine)
  • Purpose (primary value of technology use, e.g. knowledge gained by diagnostic ultrasound)
  • Intention (secondary value of technology use, e.g. possible actions as a result of diagnostic ultrasound)
  • Intention (social values attributed to technology, e.g. social and professional status of diagnostic ultrasound)

Values come to play in many ways with regard to the implementation and application of health technology, such as:

  • general moral issues (consequences, autonomy, integrity, human rights, dignity),
  • issues related to stakeholders (professionals, users, industry, patient organisations, assessors),
  • issues related to methodological choices (end points, level of evidence)
  • issues related to technology assessment (selection of technology to be assessed)(Hofmann 2005a)

A Socratic approach has been applied in this framework through a set of questions which are applied to highlight the value issues at stake in the different areas. (Hofmann 2005b) In the Norwegian context the method has been normatively open, i.e. the value analysis has not resulted in explicit normative advice, but only outlined the important normative issues. This restrictive use is due to the context and not due to the method.

The method has been applied to a series of HTA reports by the NKCHC, such as proton therapy, treatment of CFS/ME, intracytoplasmic sperm injection, palliation of cancer patients, transfusion versus other methods at blood loss, effects of snuff use, methods for age estimation in asylum seekers, methods for removing amalgam fillings, benzodiazepines treatment for drug-dependent subjects, palliative surgery for cancer patients, and use of hemopoietic stem cells from cord blood. As the technologies are different, so are the values involved. Accordingly, only a subset of the questions is applied in each HTA.

Organisational aspects

Ulla Saalasti-Koskinen, Ilona Autti-Rämö, Mirella Corio, Paolo Giorgi Rossi, Lotte Groth Jensen, Suvi Mäklin, Camilla Palmhøj Nielsen, Eva Turk, Aleksandra Zagorska

Description

What is this domain about?

The organisational aspects domain considers what kind of resources (e.g. material artefacts, human skills and knowledge, money, attitudes, work culture) have to be mobilised and organised when implementing a technology, and the consequences they may further produce in the organisation and the health care system as a whole. The issues include e.g. quality and sustainability assurance, centralization, communication, managerial structure, and acceptance.

There are three levels to consider organisational aspects: Intra-organisational (e.g. how information about the new technology is provided to the patients in the organisation), inter-organisational (e.g. how the communications between different organisations occur), and health care system level (e.g. how to set down national objectives). There are various stakeholders, besides staff and patients, at various levels, e.g. payers, providers and suppliers. These groups have usually different aims and expectations of the technology. Some issues are relevant at all levels (e.g. approval of a new technology), and some mostly at one level. Viewpoints may be different in the various levels.

The elements that constitute an organisation have been defined in many ways in different approaches, for example the physical structure, social relations, technology and organisational culture. A structure of the organisation defines its assignment of tasks, reporting systems and the mechanisms of interaction and coordination. In addition, other elements of society and its culture have influences on organisation and its function. Different types of organisations exist, e.g. the profit centre organisation, the matrix organisation and the network organisation. (Kristensen 2001)

Why is this domain important?

Organizational aspects have not been a visible part of HTA: focus has been more on the clinical aspects (Banta 2003, Draborg 2005). The growing focus of organisational issues in HTA indicates a recognition that many decisions on resource allocation in provision of technologies are of crucial importance. Organisational aspects in an HTA influence the behaviour of managers and health professionals (Battista, 2006). Also policymakers on the national level need knowledge on organisational aspects when making decisions on the use of technologies. Organisational aspects in HTA may clarify challenges and barriers in implementing health technologies.

Relation to other domains

The organisational domain might overlap with other domains especially with the current use, cost and economic evaluation, ethical aspects (e.g. acceptance and accessibility), and with the social aspects, and legal domains (e.g. privacy).

Specific features in finding, interpreting or implementing information for this domain

The complexity of health care systems and processes challenges the assessment of organisational issues. Due to the multiplicity of objectives and criteria in organisational analysis, it will be less pre-determined and more variable than for example economic and clinical effectiveness analyses. In addition, the findings are expected to be more context-dependent and less transferable than e.g. in the effectiveness and safety domains of an HTA. The choice of the areas of assessment should be guided by the information needs of the end users of HTA (e.g. regional health authorities'  focus may differ from that of hospital managers).

Issues specific for screening technologies

A screening program is a system incorporating all necessary steps, from identifying and providing information to the eligible population, through actual screening, to diagnostic testing and treatment. The assessment of a screening technology implies thus an assessment of a complex organization where organisational changes and relations within and between organisations are considered.

The screening technology under assessment can have various objectives and thus various implications for organisational aspects assessment. For example, when assessing mammography screening program, the focus can be either in a new screening test (digital mammography), or population eligible for screening (screening for women less than 50 years old), or varying screening interval (1 to 3 years), or the way to deliver the test (e.g. calling people to attend the fecal test versus mailing the test kit to them in colorectal cancer screening).

Regarding the population eligible for screening, the extent of the use of screening and waiting times defined in the Health problem and current use Domain, are of importance to the Organisational Domain. In the Description and technical characteristics Domain issues concerning definitions of the screening test and further investigations (diagnostic tests) are important also in the organisational domain. 

Methodology

Where to find information?

To reduce publication bias, it is recommended that a wide range of sources of information should be searched (Bidwell 2003). These should include published literature, as well as grey literature, hand searching of journals, contacting experts and scanning reference lists of relevant papers. Sometimes it is needed to carry out primary study about specific issues for example work processes.

Databases and search strategies

Organisational studies could be found in different databases. Selection of databases depends on the context. The most important databases are:

  • Medical databases: Medline, Medline in Process, Cochrane Library, HTA, DARE, NHS EED, Cinahl
  • Social Science databases: Sociological Abstracts, Social Services Abstracts, Social Care on line / Caredata and SocINDEX, PsycInfo, ASSIA (Applied Social Sciences Index and Abstracts)
  • Administrative studies: General science publishers' databases such as Emerald Library, Science Direct and Ebsco Academic Search Elite, Pub Med Central (PMC) and Bio Med Central (BMC), ProQuest Health Management, HealthSTAR (Health Services Technology, Administration, and Research
  • Educational database: ERIC   
  • Gray literature: Dissertational Abstracts, Scirus (reports of hospital studies and doctoral thesis), OAIster 
  • GIN guidelines
  • Registers,  e.g. national screening registry;
  • international, national and regional routine collected statistics (Health Information Database DRG);
  • conference proceedings (Web of Science database);
  • national and regional guidelines;
  • expert opinions;
  • experience of organisations e.g. NHS Technology Adoption Centre http://www.technologyadoptionhub.nhs.uk/; and
  • manufacturers' handbooks.
Other useful sources and links
Own research

When necessary, primary research could be carried out according to the co-production approach, but it will usually be very time-consuming. There are several possibile study methods to choose from, e.g. interviews, questionnaires, observation, an analysis of written material. If the resources available for the assessment project does not allow carrying out own primary research, it can be useful to consult health care professionals or other content experts.

What kind of information is required?

Framework

In a complex system, such as health care, the boundaries are typically fuzzy and activities of different agents are not predictable.  Multiple approaches are needed in this kind of systems. Through different theoretical frameworks we can understand how various organisational functions operate.

One approach to address health care systems is to divide them into micro level (patient interaction), meso level (health care organization and community) and the macro level (health policy). All these levels have been taken into account while defining the issues of the organisational domain. Some issues are relevant at all levels (e.g. approval of a new technology) and some mostly in one level, for example issues related to the staff which affect mostly in the hospital level. In addition, different viewpoints have been noticed. There are issues related to the patients in nearly all topics.

The relation between technology and organisation can be tackled in different ways. At least two different and incompatible views on causality and transferability can be differentiated with respect to the organisational issues: the diffusion model and the translation model, see Appendix 3 (Kristensen 2001, Latour 1987).  Parallel viewpoint is seen in the social domain.

The definition of organizational analysis in this document is based on the loose approach called co-production of technology and its context and especially on the translation model. Its main thesis is that a technology needs a context or a network to function. In addition to the translation model, other approaches that form the co-production approach are for example constructive technology assessment (Schot 1992; Douma 2007), the systems approach (Hughes 1983) and social construction of technology (Bijker 1987).

Both organisational and administrative perspective can be used in the organisational analysis (Kristensen, 2007). Administrative analysis uses a managerial perspective (e.g. decision making, co-ordination and managerial tools) and organisational analysis deals with changes in relation to the executing /producing function (e.g. organisational conditions, change processes).

Usually, it will be difficult to isolate and measure the output effects of given organisational initiatives. More realistic is to describe the various process dimensions in relationship between a technology and organisational behaviour.

Study types, design, outcome measures

The natural starting point of an analysis of change in processes will be to map the current work-flow / patient-flow. Therefore, the methods for data collections involve qualitative methods such as interviews or observations, or quantitative methods such as surveys (Kristensen 2001).

Qualitative study is the mostly used study type in organisational domain. In this kind of search approach the scope of relevant evidence is not known in advance and therefore the search method is usually iterative. The collected information of iterative search could be systematic only if the search steps have been documented carefully. 

The review question should be based on PICO (Patient, Interventions, Control, Outcomes), see Appendix 3. Within qualitative evidence synthesis SPICE (Setting, Perspective, Intervention/Interest, Comparison, Evaluation) (Booth, 2003) or ProPheT (Problem, Phenomenon of Interest, Time) (Joanna Briggs Institute, 2008) could be more eligible for formulating a question.

It depends on the research question what kind of study design gives the most reliable answer to it. Both quantitative and qualitative studies and their synthesis are essential in the organisational domain. Although the most important sources of information are observational and qualitative studies, it is good to check if there are controlled or quasi-experimental studies available (See appendix). Other types of relevant information for organisational issues can be found in national and international guidelines, statistics and registers and handbooks.

Issues specific for screening technologies

Policy measures, such as the choice between organised and opportunistic screening, or the reimbursement/funding strategies are implemented at the macro level and are likely assessed more appropriately by observational/qualitative studies; the organisation of screening services delivery at the institutional (meso- level) might be studied using qualitative research designs, but experimental studies may offer valuable and crucial information; similarly at the micro level of the interaction between provider and patients both experimental and qualitative evidence are important to assess screening technology. Of course there are interactions across the levels and different actors may be involved at more than one level (i.e. the provider is involved both at the meso and at the micro level).

Tools for critical appraisal

There are different study types used in the organisational domain and therefore the range of quality assessment and appraisal instruments available to assess studies is wide.  These are presented in table 1. Examples of quality assessment and checklists of different study types are shown in the appendix 3. Some of the appraisal instruments are generic and others targeted to specified contexts. For quantitative studies assessment of quality is clearer than for qualitative studies. It has been claimed that quality of qualitative study cannot be determined by prescribed instruments. Therefore using checklist or scales on quality assessment of observational or especially of qualitative studies is not always relevant.

Table 1

Issue

Study type

Quality assessment

Systematic vs other

Synthesis

What kind of work flow, participant flow and other processes are needed?

Guidelines, observational, mostly qualitative

AGREE, or other methods to evaluate guideline quality, tools for qualitative study appraisal

Not necessarily systematic

narrative

What kind of quality assurance is needed and how it should be organised?

Obervational, qualitative and quantitative. Intervention studies are possible, usually not controlled (pre-post), randomisation is not possible for most of the interventios

Relevant. Tools for RCT and observational study evaluation, tools for qualitative study appraisal

Not necessarily systematic, systematic for national and regional reports

narrative

What kind of involvement has to be mobilized for participants and important others?

RCT or systematic reviews of RCTs; observational quantitative and qualitative. Guidelines.

Relevant. Tools for RCT evaluation. AGREE.

Systematic

meta-analysis for most commonly evaluated intervention, narrative for less common and complex interventions

What kind of staff, training and other human resources are required?

Guidelines, scientific soc. consensus, Obervational, qualitative and quantitative

Not relevant, tools for qualitative study appraisal

Not necessarily systematic, systematic for national and regional reports

narrative

What kind of co-operation and communication of activities have to be mobilised ?

Observational, mostly qualitative. Guidelines.

Not relevant, tools for qualitative study appraisal.

not necessarily systematic, systematic for national and regional reports

narrative

What influence decentralisation or centralization of the technology will have?

Guidelines, observational, mostly qualitative. Health Information Databases (DRG etc.)

Not relevant, tools for qualitative study appraisal

not necessarily systematic

narrative

What kinds of investments are needed (material or premises)?

Guidelines, producer technical handbooks.

Not relevant

systematic at least for technical requirements

narrative

What is the likely budget impact of the implementation of the technology for the payers (e.g. government)?

Costing and buget impact analyses

Tools for the evaluation of economic studies

systematic

narrative

What management problems and opportunities are attached to the technology?

guidelines, observational studies mostly qualitative

Not relevant, tools for qualitative study appraisal

not necessarily systematic

narrative

What kind of monitoring systems are there for the technology?

guidelines, consensus, registries

AGREE, or other methods to evaluat guideline quality

systematic

narrative

Who decides which people are eligible for the technology and on what basis?

guidelines, consensus, protocols

AGREE, or other methods to evaluat guideline quality

not necessarily systematic, systematic for national and regional reports

narrative

How is the technology accepted?

observational, mostly qualitative. Scientific societies websites

Not relevant, tools for qualitative study appraisal

not necessarily systematic, systematic for national and regional reports

narrative

How are the other interest groups taken into account in the planning / implementation of the technology?

observational, mostly qualitative. Scientific societies websites

Not relevant, tools for qualitative study appraisal

not necessarily systematic, systematic for national and regional reports

narrative

Analysing and synthesizing evidence

Data extraction

Data extraction approach must be appropriate to the review question, the type of review and the available evidence. It needs to be systematic and transparent. Data extraction can be a subjective process and therefore the design of these forms should be undertaken carefully (CRD guidance, 2009). The amount of information to be extracted should be directly related to the questions posed and must be balanced detail with usefulness (overly inclusive / minimalist data extraction form).

Key components of data extraction (especially of quantitative studies) are identifying features of the study (title, authors, journal, publication details), population characteristics and care setting, methodological quality, interventions, outcomes, length of follow-up, drops-outs, missing data, data of the results, effect measures and notes. Different form may be necessary if there are findings from qualitative studies. Example of data extraction form for qualitative studies is SUMARI done by Joanna Briggs institute.

Biases

Triangulation is a way to reduce bias in research, and thus should be done when assessing organisational issues. Triangulation compares the results from either two or more different methods of data collection (for example, interview and observation) or two or more data sources (for example, interviews with members of different interest groups). The researcher looks for patterns of convergence to develop or corroborate an overall interpretation. Triangulation can be seen as a way to ensuring comprehensiveness and encouraging a more reflexive analysis of data than as a pure test of validity. (Mays and Pope, 2000)

Synthesis

Meta-analysis is rarely used in the organisational domain because most of studies are qualitative. Qualitative evidence synthesis is a process of combining evidence from individual qualitative studies to create new understanding by comparing and analyzing concepts and findings from different sources of evidence with a focus on the same topic of interest. It can be an aggregative or interpretive process which requires authors to identify and extract evidence: categorizing the evidence, and combine categories to develop synthesized findings. Important is to understand why people feel or behave certain way and not just make a description of it. (Noyes: Cochrane handbook).

There is range of methods available for synthesizing diverse forms of evidence, for example meta-ethnography, grounded theory, thematic synthesis, narrative synthesis, realist synthesis, content analysis. Some of the methods maintain the qualitative form of the evidence such as meta-ethnography and some involve converting qualitative findings into a quantitative form such as content analysis.

  • meta-ethnography
  • thematic analysis / synthesis
    • There are toolkits for coding the text e.g. EPPI-Reviewer, which includes a component designed to support thematic synthesis (Thomas and Harden 2008).
  • grounded theory
  • qualitative research synthesis (qualitative meta-synthesis)
  • content analysis
  • case survey
  • qualitative comparative analysis
  • qualitative research synthesis (qualitative meta-summary)
  • narrative synthesis

Synthesis methods could be classified different ways and it has been argued weather it is acceptable to conduct syntheses of qualitative evidence at all, and whether it is acceptable to synthesize qualitative studies derived from different traditions. (Thomas and Harden 2008, Dixon-Woods et al 2007, CRD Guidance 2009)

Qualitative and quantitative findings could be synthesized in two ways: multilevel synthesis (separate and combined synthesis) and parallel (separate and juxtaposed synthesis) (Noyes, Cochrane handbook). Quantitative and qualitative studies can be synthesize together, one example is systematic review on teenage pregnancy and social disadvantage (Harden and Oakley 2009)

Reporting and interpreting

The transferability of the research identified in literature searches, will have to be assessed very carefully, since this domain is in general to be considered highly context-specific. It is possible, that in many cases, the results from the literature review, can be considered to be hypothesis generating, and be useful for planning primary research in the own context.

Assessment elements

G0001 Assessment element card

Issue: What kind of work flow, participant flow and other processes are needed?

Topic: Process

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Current tasks and work processes and participant path should be described. Preparations participants need to do before and after the intervention (e.g. diet before bariatric surgery) must be taken into account, as well as need for self/home monitoring. There are many actors at different levels (intra-organisational, inter-organisational and health care system level) in the process. Continuity should be ensured so that there will be no gaps between the steps of the process. It has to be described how the screening process has been organised, e.g.: 1) how the target population is chosen, 2) how and by whom the invitation is carried out (open/fixed invitation, announcement/personal invitation letter), 3) how and by whom the information for consent is given, 4) how, where and by whom the test is executed, 5) how, where and by whom the further investigations and treatment are carried out, 6) how, when, and by whom the follow up services are carried out (e.g notifying results, recalls, reminders ).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, annual reports and statistics, reports and own study (e.g. questionnaires and interviews of different actors)

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2001, Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Mandatory: A0007, A0023, A0011, A0013, A0014, A0015, A0016, A0017. Other: B0004, B0005, B0016. Order of doing: to be answered prior to E0001

Sequential relations

G0012 Assessment element card

Issue: What kind of quality assurance is needed and how should it be organised?

Topic: Process

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

A new technology usually have an effect on current quality assurance not only inside the organization but also outside in different health care levels. To assure the quality, a monitoring system with standards and indicators are needed. Screening involves asymptomatic participants and therefore quality control is crucial. There are national, regional and/or (cross)organisational (screening unit) demands for quality assurance. Quality control needs to be systematic at every step of the screening process steps and throughout the screening programme. Acceptable delay from screening test to test positive result and finally to treatment must be specify. Special attention has to be paid to the control when the programme is provided by several providers (e.g. a combination of private and public health care organisations) when test and further investigations are separated.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, annual reports and statistics reports of hospitals and own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratories). Information from manufacturers. .

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0012, C0007

Sequential relations

G0002 Assessment element card

Issue: What kind of involvement has to be mobilized for participants and important others?

Topic: Process

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

A new technology may require distribution of tasks among the people involved in the treatment and care. Participants and their important others may be more actively involved in own care and treatment – or tasks they used to carry out may be taken over by health professionals. The screening has to be organised in the way that the test and the further investigations are easily attainable e.g. mobile mammography.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, annual reports and statistics reports, hospital documents and own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory, participants).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0006, A0007, A0023, B0014, B0015, H0002, H0003

Sequential relations

G0003 Assessment element card

Issue: What kind of staff, training and other human resources are required?

Topic: Process

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

How much staff is needed and what kind? How much trained resources are needed to ensure proper functioning? Different health care levels (e.g. primary and secondary care) should be taken into account. What are the needs for training and expert advice? Are there national, regional or in-house registries and criteria for personnel and training? Implementing a technology can change the job and have thus influence on job satisfaction.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, reports and documents of the hospital or hospital districts and own study: interview or questionnaires of different actors of the process.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Busse 2002, Kristensen 2001, Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0013, C0063, E0001

Sequential relations

G0004 Assessment element card

Issue: What kind of co-operation and communication of activities have to be mobilised?

Topic: Process

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Implementing a technology can demand new co-operation and communication in- and outside the organization, e.g. other hospitals, pharmacies. Also interaction and communication with patients/participants and their important others will change. Adaptation of self/home monitoring needs close co-operation and fluent communication. Screening needs close co-operation and fluent communication between all actors of the screening process in all steps (e.g. screening unit, laboratory, hospital, registry, participants). There are actors at different levels which make the communication and co-operation challenging, especially when making up a new screening. The information must be fluent and electronic communication (software) is crucial. Adequate communication with participants and their important others must be taken into account. Different kinds of "patient information" could be defined for screening. For example: 1. "promotional/educational information" with the aim to involve target population and to promote participation 2. "screening related information" to communicate with participant the "phase related information" in the different phases of the process (e.g. sending invitation; communicating the test results etc.). Information strategies should be tailored to the specific subgroup of the target population (depending on socio-economic status, cultural background, epidemiological features, etc.). Risk families need special information.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, reports and documents of hospital and hospital districts, guidelines, own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory, participants).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2001, Kristensen 2007, Senter för Medisinsk metodevurdering (SMM) 2003

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0015, C0063, H0007, H0008

Sequential relations

G0005 Assessment element card

Issue: How does de-centralisation or centralization requirements influence the implementation of the technology?

Topic: Structure

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The setting (primary - secondary - tertiary care) can vary between different countries depending on the health care system. (De)centralisation could have some economical and qualitative benefits. Centralisation could make the technology more difficult to access. Sometimes screening test (for example maternal ultrasound) needs special experience from personnel which is possible after education and sufficient amount of patients. Centralisation could make screening or further investigation more difficult to access. For example timing is important in foetal screening. Decentralisation makes screening more attainable but the quality can weaken.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, reports and documents of hospital and hospital districts, health information databases (DRG etc.), own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory, participants).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Busse 2002, Kristensen 2001, Kristensen 2007, Senter för Medisinsk metodevurdering (SMM) 2003

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0005

Sequential relations

G0006 Assessment element card

Issue: What kinds of investments are needed (material or premises) and who are responsible for those?

Topic: Structure

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Implementing the required changes in e.g. premises may be costly for the organisation. High costs can influence the decision of purchasing the new technology. There may be division of costs so that some organisation(s) take the acquisition costs and others the running costs. Investments of all steps and actors of the process must be perceived. When building up a new screening programme, there's need for many investments (e.g. equipments, education and implementation support, training).

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, reports and documents of hospitals and hospital districts and manufacturers (e.g. producer handbook), own study: questionnaires and interviews of different actors of the screening process (monitoring authorities, hospitals, hospital districts, laboratory)

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0011, A0012, A0019, B0008,, D0023, E0001. Order of doing: to be answered prior to E0001.

Sequential relations

G0007 Assessment element card

Issue: What is the likely budget impact of the implementation of the technology for the payers (e.g. government)?

Topic: Structure

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Whenever a technology is reimbursed, there will be an impact on the health care budget. Budget impact analysis examines the likely impact of the reimbursement of the technology on financial outlays from the perspective of the payers (e.g. government). Different payers include: government, region, municipalities, employer, insurance company or participant. Payer can change during the management process. Incentives are connected to this issue: What kind of incentives the budget impact imposes on different actors? How this potentially impact on the organization? Payer could change during the screening process (e.g. municipality pays screening test but hospital district pays further investigations). Screening is usually free of charge for people, but sometimes participants have to pay e.g. hospital fee for further investigations. It must be noticed that when building up a new screening, initial costs are needed.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, reports questionnaires and interviews of different actors of the screening process (monitoring authorities, hospitals, hospital districts, laboratory), information from manufacturers.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Mauskopf et al. 2007 Principles of Good Practice for Budget Impact Analysis, Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0011, E0001

Sequential relations

G0008 Assessment element card

Issue: What management problems and opportunities are attached to the technology?

Topic: Management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The issue concerns the administrative / managerial questions of technology: management of resources (e.g. investments), co-ordination (in relation to different levels and different steps of the process), establishment of objectives, monitoring and control, evaluation and sanctioning. Data/information management systems connected to each of these points have to take account.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, reports and documents of hospitals, own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Report of the ISPOR Task Force on Good Research Practices

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0011, A0012, A0015, A0016, A0017, H0009

Sequential relations

G0013 Assessment element card

Issue: What kind of monitoring requirements and opportunities are there for the technology?

Topic: Management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

There may be different monitoring systems for different phases of the process where the technology is used (e.g. personnel registry or quality control system) These registries are part of quality assurance. It is necessary to define validated/recommended indicators (guidelines for QA, or other documents). A core data set is needed to monitor the phases and to produce the recommended indicators. The information flow should be analysed.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, reports and documents of hospitals and hospital districts, guidelines, own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory).

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

A0013, A0014,

Sequential relations

G0009 Assessment element card

Issue: Who decides which people are eligible for the technology and on what basis?

Topic: Management

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Information about the possible variations in the decision level and criteria has ethical implications. Decisions about the people eligible for screening is done in the beginning of the screening. Usually, it has been made nationally or regionally (in municipalities) but also locally (by employers). In systematic screening, the screening unit does not make decisions about who is eligible for screening. The management of positive test result needs systems to guarantee proper follow up and sometimes case specific evaluation. In this topic responsibilities should be identified.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, guidelines, documents of hospitals, own study: questionnaires and interviews of different actors of the process (monitoring authorities, hospitals, hospital districts, laboratory).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

F0012, I0012

Sequential relations

G0010 Assessment element card

Issue: How is the technology accepted?

Topic: Culture

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Acceptance should be looked at by different perspectives: by organisation, by personnel and by participants. Organisational view can be separated out intra-organisational (primary care), inter-organisational (secondary care) and health care system level. In all these actors/views acceptance could vary. Alternative ways to introduce a new technology into the organisation could influence problems e.g. resistance among staff and dysfunction of processes. Acceptance could vary in the same screening process for example in foetal screening someone accepts ultrasound but not chromosomal (serum) test. Example of organisational acceptance: Sometimes screening could consist of elements which are not suitable for the image of the organisation. Screening is voluntary and for persons eligible for screening both decisions are right decisions: to participate or not. Giving understandable information on pros and cons of screening is important. Communicational skills of personnel may have an influence on acceptance of screening.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, own study: questionnaires and interviews of different actors of the screening process (monitoring authorities, hospitals, hospital districts, screening units, laboratory, staff, participants).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Finohta's EUnetHTA workshop 2006, Kristensen 2007

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

F0007, H0006

Sequential relations

G0011 Assessment element card

Issue: How are the other interest groups taken into account in the planning / implementation of the technology?

Topic: Culture

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It may be useful to know who are the possible stakeholders, as well as what kind of co-operation exists and what kind of interaction is needed. The stakeholders could be e.g. the pharmaceutical industry and companies offering technologies for screening, authorities (national / regional), registry, administrative parties, municipalities, policy makers / decision makers, staff groups, GPs/primary care physicians and patient organisation. One can also ask: Has the patient organisation taken part into the evaluation process? Has it been involved from the beginning (in the planning) or in the later stages for example as commentator?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Literature search, reports and documents of hospitals, own study: questionnaires and interviews of different actors of the screening process (monitoring authorities, hospitals, hospital districts, screening units, laboratory, manufacturers, registry, participants).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Kristensen 2001, Kristensen 2007, Senter för Medisinsk metodevurdering (SMM) 2003

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

F0003, F0011

Sequential relations

References

Banta D. The development of health technology assessment. Health Policy 63 (2003) 121-132.

Battista RN. Expanding the scientific basis of health technology assessment: A research agenda for the next decade. International Journal of Technology Assessment in Health Care 22:3 (2006), 275-282.

Bidwell S, Jensen MF. E-text on Health Technology Assessment (HTA) Information Resources. National Information Center on Health Services Research and Health Care Technology (NICHSR). 2003.

Bijker W, E, Hughes, T.P., & Pinch, T. The social construction of technological ssssystems: New directions in the sociology and history of technology. Cambridge, MIT press, 1987.

Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy : the STARD Initiative. Ann Intern Med 2003; 138 (1): 40-4

Busse R, Orvain J, Velasco M etc. Best practice in undertaking and reporting health technology assessment. Working Group 4 Report. International Journal of Technology Assessment in Health Care, 18:2 (2002), 361-422.

Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0. Noyes et al. Qualitative research and Cochrane reviews.The Cochrane Collaboration. http://www.cochrane.org/  . 13.11.2008.

Dixon-Woods M, Booth A, Sutton AJ. Synthesizing qualitative research: a review of published reports. Qualitative Research 2007; Vol. 7, No.3, 375-422.Draborg E, Gyrd-Hansen D, Poulsen PB, Horder M. International comparison of the definition and the practical application of health technology assessment. International Journal of Technology Assessment in Health Care, 21:1 (2005), 89-95.  

Douma K.F.L, Karsenberg K, Hummel M.J.M, Bueno-deMesquita J.M, van Harten W.H. Methodology of constructive technology assessment in health care. International Journal of Technology Assessment in Health Care, 23:2(2007), 162-168.

Draborg E, Gyrd-Hansen D, Poulsen PB, Horder M. International comparison of the definition and the practical application of health technology assessment. International Journal of Technology Assessment in Health Care, 21:1 (2005), 89-95.  

EUnetHTA Workshop. Finohta 20.6.2006.

Harden A, Oakey A. Teenage pregnancy and social disadvantage: systematic review integrating controlled trials and qualitative studies BMJ 2009; 339:b4254

Hughes, T.P. Networks of power: Electrification in Western society, 1880-1930. Baltimore, MD, John Hopkings University press, 1983.

Khan K, Riet G, Glanville J et al. Undertaking Systematic Review of Research on Effectiveness. CRD's Guidance for those Carrying Out or Commissioning Reviews. CRD Report Number 4. NHS Centre for Reviews and Dissemination, University of York. 2001

Kristensen FB, Horder M, Poulsen PB (eds). Health technology assessment handbook.  Danish Institute for Health Technology Assessment (DIHTA). 2001

Kristensen FB, Sigmund H (ed.). Health Technology Assessment Handbook. Copenhagen: Danish Centre for Health Technology Assessment. National Board of Helath, 2007.

Latour B. Science in Action. How to follow scientists and engineers through the society. Harvard University Press, Cambridge, MA, 1987

Leavitt HJ. Applying Organisational Change in Industry: Structural, Technological and Humanistic Approaches. In Handbook of Organisations, edited by James G March. Chicago: Rand McNally. 1965.

Leyes M. Health technology assessment: the contribution of qualitative research. International Journal of Technology Assessment in Mays N, Pope C. Qualtative research in health care: Assessing quality in qualitative research. BMJ 2000;320:50-52.

Plsek PE, Gleenhalgh T. The challenge of complexity in health care. BMJ 2001;323:625-8.

Popay J, Rogers A, William G. Rationale and standards for the systematic review of qualitative literature in health services research. Qualitative Health Research 1998; 8(3), 341-351.

Robbins SP. Organisational theory: Structure, design, and applications. Prentice-Hall. 1987.

Ryan M, Scott DA, Reeves C et al. Eliciting public preferences for healthcare: a systematic review of techniques. Health Technology Assessment 2001; Vol.5: No.5

Schein, E. Organisational Culture and Leadership. San Francisco: Jossey-Bass; 1985.

Schot, J.W. Constructive Technology Assessment. Science, Technology & Human Values, Vol. 17, No. 1, 36-56.

Senter för medisinsk metodevurdering (SMM). Medisinsk metodevurdering. En innforing. 2003. Oslo, Norge.

Systematic Reviewes. CDR's guidance for undertaking reviews in health care. Centre for Reviews and Dissemination. university of York, 2009.

Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews. BMC Med Res methodol. 2008; Jul 10;8:45.

Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care; 2007;19 (6): 349-57.

Social aspects

Alessandra Lo Scalzo, Lotte Groth Jensen, Suvi Mäklin, Ulla Saalasti-Koskinen, Ingrid Wilbacher, Aleksandra Zagorska

Description

What is this domain about?

Social domain takes the patient or individual as a point of departure in an HTA. A technology may be practiced in hospital, primary care or at home. Implications for patients may though extend far beyond the original setting of the technology. The patient is not just a passive target for interventions in health care. He is also a human being with different roles – a family member, a citizen, an employee, a consumer etc. (Hansen 2007). His life takes place in various arenas: everyday life, homes, schools, workplace, health services, etc. The use of the technology may change the roles, skills and positions in both negative and positive ways.  A new role can strongly affect all the arenas of one’s everyday life and all the important others. Considerations of power, empowerment and stigmatisation are therefore essential (Hastrup 1997, Goffman 1990, Devereux 1963, Rose 1993).

Patients and carers give specific meanings and significance for health technologies. Perceptions are attached to feelings of hope, fear, or perhaps uncertainty as well as values of society (Hansen 2007, Lehoux 2006, Whyte 1997, Bech 1992, Douglas 1996). The social analysis is interested in all these aspects.

The analysis of social aspects of health technology can include at least two kinds of questions. The first set of issues focuses on the kinds of resources (people, support, money and so on) that have to be enacted and mobilised from the point of view of a patient before, during and after the implementation of the technology. The other set of issues focuses on the experiences, actions and reactions of patients with respect to the technologies as well as on the changes and consequences that the enactment of the technologies may further generate. These are for example changes that occur with respect to a person’s working capacity, social relationships, coping with illness and treatment, or attitudes towards a person who uses the technology.

The social analysis of a health technology can be considered at two levels: micro and macro sociological. The first is related to the individual (inter-individual relationship, direct environment of a person, direct effects on an individual), while the second focuses on the society as a whole (views, attitudes, culture, norms and values). From a macro sociological point of view core questions are aimed at understanding the benefits and consequences of the technology for the target population, for specific groups (religious, ethnic etc.) and for the general population. 

Figure 1 provides a view of different social aspects that are relevant from a patient's perspective (Hansen 2007). The model intends to show and map different aspects, which could be considered of relevance for a specific HTA analysis. Social domain chooses mainly to focus on the individual topics, communicative topics, and topics of major life areas such as family life, work life, and leisure time. These topics are underlined in figure 1.

 {{48-Soc2}}

Figure 1. Social aspects of relevance from a patient perspective in HTA. Modified from (Hansen 2007).

Issues specific for screening technologies

Issues important in screening:

  • attendance/participation to screening
  • compliance to further assessment tests and treatment protocols
  • patient and operator preference for the screening organisation and setting (in particular the between organised and spontaneous screening)
  • acceptability of intervals (longer or shorter)
  • attitude of the patient organisations to propose very aggressive and invasive screening protocols
  • attitude of clinicians to apply clinical protocol for differential diagnosis to screening protocols.

All these issues should be seen - from a social point of view as in the aspects of the Council Conclusions on Common values and principles in European Union Health Systems (2006/C 146/01 )that include quality, safety, care based on evidence and ethics, patient involvement, redress, privacy and confidentially.

Why is this domain important?

The technology does not produce the good results alone. Social analysis reveals the resources needed in individual's daily activities in order to achieve satisfactory results. The use of technology always produces some kind of changes or consequences in different spheres of social life, which should be anticipated. These can be positive or negative, or even unexpected (Rapp 1999, Kaufert 2000, Cambrosio 2000). The different meanings individuals give to a technology and its implication are important to recognize (Dreier 2000, Bourdieu 2000).

Relations to other domains

Patient perspectives is present n several other Domains:

–        Ethical analysis domain

–        Effectiveness and safety domains

–        Costs and economic evaluation domain

–        Legal domain.

Coordination is needed across these domains when preparing a Core HTA.

Specific features in finding, interpreting or implementing information for this domain

Technologies are not applicable everywhere. They work within networks of different human and non-human elements. Implementing a technology means that the technology and its entailing network has to be re-built in a new place (Koivisto 2007, Koivisto 2008). This is equally true for simple technologies, such as a single drug or a single device, as for complex interventions like screening or disease management programmes. Transferability of the social analysis results requires careful consideration of the comparability of the social and cultural circumstances presented in the published literature to the circumstances at hand.

Furthermore, social implications change over time as people put the technology to use, get accustomed to it, and find new ways of using it in combination with other technologies or practices. An analysis of social aspects can never foresee the exact social implications and consequences of the use of a given technology. It may however, provide us with important knowledge of aspects that need to be taken into consideration.

Issues specific for screening technologies

Equity in access is essential for the participation in the screening and thus the success of the screening program. The delivery modes of screening may have an impact on this. Self-sampler devices and the possibility to mail the sample instead of clinic visit and telephone reminder messages can affect participation, as well as mass media campaigns.

Correct and balanced information on benefits and harms of screening is essential for an individual to be able to make informed decision to participate screening.

Figure 2 illustrates the scope of social analysis by an example of the individual’s itinerary in and outside the health services during screening procedure.

{{169-Social2}}

Figure 2.

Methodology

Where to find information?

Issues on the social aspects of technologies can be subject of the following fields:

-          Medical Anthropology,

-          Medical Decision-Making,

-          Medical Sociology,

-          Science and Technology Studies,

-          Governance of Innovation Studies,

-          Medical Ethics,

-          Social Psychology,

-          Communication science, and

-          Health Services Research

-           Health Sociology

Examples of relevant scientific journals: Health Expectations, Medical Anthropology Quarterly, Social Science and Medicine, Anthropology and Medicine, Sociology of Health and Illness, Qualitative Health Research, Values in Health, Medical Decision Making.

Databases and search strategies

Psychological/sociological databases such as

-          Psychinfo,

-          ASSIA (Applied Social Sciences Index and Abstracts),

-          Sociological Abstracts and

-          ISI Web of Science

-          Social Services Abstracts,

-          Social Care on line / Caredata

-          SocINDEX

Euroethics (European Database Network on Ethics in Medicine, including:

-          Biogea (Italy),

-          Cendibem (Spanish),

-          CRIB (Belgium),

-          ETHINSERM (France),

-          ETHMED (Austria, Germany, Switzerland),

-          EUROETHIK (Germany),

-          MIKS (Sweden).

Medical databases such as

-          Medline,

-          Embase,

-          Cinahl.

Suggested search terms include: "social aspects of", "medical decision making process", "patient education", depending on the PICO question.

Useful other sources

Other sources of qualitative studies can be

-          Citizens and patients associations

-          WHO, OECD, ILO, UNESCO homepages and databases

-          Citizens and patients associations

-          Patients’ (virtual) forums

-          Structured systematic content analysis of Patients’ (virtual) forums

-          Structured systematic content analysis of Mass media

The use of qualitative sources should always be done in caution do to the high possibility for the subjective bias.

Own research
Primary study

If no relevant studies could be identified, it could be worthwhile to carry out primary studies, e.g. interviews and questionnaires. Timing of the primary study must be considered thoroughly. Appropriate time point for assessing the patient experience will be different with different technologies. Both ethical and practical considerations must be taken into account when deciding on whether to study people before, during or after using the technology. This choice may also have considerable significance for the results. Primary study, as any intervention, affects behaviour and practice. There must be clarity whether the effects noticed in e.g. an interview are related to the implementation of the technology or to the interview itself.

Consultation

If there is not enough time to perform a primary study, the opinion of health care professionals and content experts or other stakeholders can be consulted. However, one needs to be aware of that the amount of knowledge on the views of respondents may be limited as it reflects participants' willingness to listen and talk. Even when talking the information is influenced by the positions and power relations of the professionals and patients, knowledge asymmetry, patient's dependency on doctor's goodwill and time constraints. Stakeholders may represent patient’s perspective, but the evaluator should be critical to any political agenda.

Social analysis is both theoretically and empirically complex and demanding. Advanced skills in social analysis are required from the person conducting this part of the HTA. An assessment of patient and social aspects should not be a separate process within an HTA. Co-operation and interaction between the HTA team members is essential.

What kind of information is required?

Analytical framework

At least two different approaches can be differentiated with respect to the assessment of the social sphere. The approaches are 1) the diffusion model and 2) the translation model (see also Appendix 3). They imply different study questions and methods for the analysis of social issues. They were originally characterized by Latour (1987) and elaborated by Tryggestad & Borum (2001).

Diffusion model

The diffusion model bases on a linear, one-directional conception of causality. This model supposes that a technology has an inner causal power that can affect and change the individuals' life (micro-level), the organizations such as hospitals or health care centers (meso-level) or the national and international systems (macro-level). From the point of view of citizens the model implies that a health technology can cause the people to work longer, it can change the way people live and it can improve the quality of life of people.

The adoption process of a technology typically includes (Rogers 2003):

  • Technical knowledge about the technology 
  • Persuasion for the participation
  • Decision for the participation or use 
  • Implementation of changes to decrease risk
  • Confirmation for further use according to the time schedule of the program

According to the diffusion model, it can be asked

  • which social impact will the implementation of the technology have?
  • how does the technology change the social or working life of people?
  • which strategies should be adopted to facilitate diffusion of the technology?

Translation model

The translation model sees technology as something endogenous. It cannot be separated from the health system, its users, and the context of use; it is not an independent and stable entity. Technology is a network of human and non-human elements that produces change. From the point of view of a citizen it is up to the perceptions and discretion of the people what they make with the technology or with the possibilities it offers. Constant interaction between the technology and people determines whether, in what ways, and how often the technology is used. Therefore the actual implementation of the technology may be different from policy makers’ analytical expectations. The task of the evaluator is to reconstruct the chains of empirical events which are related to the implementation and utilization of the technology.

According to the translation model, it can be asked

  • how much and what kind of resources (material entities, time, money, people, etc.) must be mobilized and organized in order to produce satisfactory result?
  • what kind of behavioral patterns (such as resistance or compliance) or attitudes can influence/interfere use of the technology? 
  • how do potential users perceive its benefits and risks?
Study types, design, outcome measures

When estimating the applicability of published literature, it is important to consider contextual factors. There is no hierarchy in study designs of social research. Studies have to be evaluated according to their relevance for the issue at stake and quality.

A number of study designs, both quantitative and qualitative, are relevant. These include randomized or non randomized controlled trials, observational studies and open or semi structured individual or group interviews. For qualitative studies the relevance refers to the ‘transferability’ of the concepts to our setting, in knowing how far the findings help us to understand ‘what is going on’ in our setting (Green & Thorogood 2005).

Patient related outcomes are relevant also for many questions in effectiveness and safety domains. When these issues are brought into the analysis of social aspects, focus is on the interrelation between biological, individual and social aspects. Patient related outcomes can result in major consideration and impact on the content and conclusions of a HTA report. The technology may for instance have other patient related consequences than intended.

Tools for critical appraisal
Quality assessments should evaluate (Facey 2010).
–        the purpose of the study and relevance to study question,
–        context (population/setting/values),
–        appropriateness of methods and theoretical framework,
–        transparency of data generation, analysis and interpretation (avoidance of bias),
–        connection between research question and conclusions (internal consistency in relation to the theoretical framework of the study) and
–        the account of the knowledge generated given the methods (relevance for practice)
Quality assessment of qualitative research
In assessing qualitative studies it should be noted that generalizability of findings in statistical terms is often not the aim. In qualitative works study samples are rarely randomly selected because the logic of generalizability is here different. The aim is to provide in-depth (‘thick’) descriptions or to address particularities rather than to provide generalizable findings (Greeen & Thorogood 2005).
Another point is that researchers’ judgment sometimes applies to the interpretations provided by qualitative studies. Although the researcher describes a certain issue from the point of view of participants, s/he simultaneously unpacks the issue in such a way that broader meanings and connections can be elicited. Therefore, the presence of researcher’s perspective does not per se discredit the study. So long as the judgment is made consciously and articulated explicitly in the study, it may not be considered as a source of bias.

Guidelines for standards on qualitative research vary and are currently debated and developed. For further guidance, see e.g. Malterud et al 2001 or Hansen et al 2007. Another tool can be found in Green & Thorogood 2005, page 241 and Tong 2007. 

Analysing and synthesizing evidence

Data extraction

Publication details: First author, year

 

Social topic(s)/issue(s): to be categorized by the reviewer

 

Nature of the study: aims/objectives, user/carer involvement in the design/conduct of study, country, site (setting, key characteristics of the context), details of theory/conceptual model.

 

Methods: study type and design, study date and duration, sampling/recruitment, methods of data collection, data collector, used research tools (if any), analysis methods

 

Participant characteristics: gender, age, ethnicity, types of practitioners, policy makers or patients

 

Features the studied intervention (when applicable): aim of the intervention, intervention process (description of how was the intervention/service delivered)

 

Outcomes and results: outcome measures, details of findings, strengths/limitations of the study, author's conclusions.

 

Reviewers' comments: e.g. remarks of quality issues

 
Qualitative synthesis
-           Thematic mapping

Qualitative studies often involve generating evidence in the form of certain themes, concepts and trends. Thematic mapping means mapping out relevant sub-themes, and the assessment of the quantity, quality of existing literature related to them. Applicability of published information depends on its ability to give insight into social processes. Examples of sub-themes may be: how do illness or risk perceptions change family relations, roles, people's interaction with technology, unforeseen and unintended social consequences, or risk management. A thorough description of relevant themes and dimensions is more important than finding all relevant studies. It is also important to define the questions that cannot be answered on basis of the existing literature.

-          Other methods

The synthesis of qualitative studies can be done according to different methods such as meta-ethnography (Noblitt 1988) or narrative analyses (Popay 2004). Guidance for making synthesis of qualitative literature can be found in method books (Petticrew 2006, Coren 2006, Social Care Institute for Excellence 2006). A critical interpretive synthesis on literature considering access to healthcare by vulnerable groups provides one example (Dixon-Woods 2006).

Assessment elements

H0001 Assessment element card

Issue: Which social areas does the use of the technology influence?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Map the major life areas of the patients or citizens using the technology, and their important others. Major life areas include family life, day care, school, work, leisure time, lifestyle, or other daily activities. The use of the technology can affect the final decision of the individual about participating.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Hansen 2007

Content relations
Sequential relations

H0002 Assessment element card

Issue: Who are the important others that may be affected, in addition to the individual using the technology?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

E.g. the results of screening or genetic and prenatal testing, may affect relatives.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical and Legal domains

Sequential relations

H0004 Assessment element card

Issue: What kind of changes may the use of the technology generate in the individual's role in the major life areas?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue is about the patient’s social roles and ability to manage and maintain relations with other people in a socially appropriate (associated by the social norms and values defining the role) manner in major life areas.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted

References
Common to all used applications

None

Specific to Screening Technologies 1.0

ICF 2001: activities and participation, interpersonal interactions and relationships (chapter 7, d710-779), community, social and civic life (chapter 9:d910-d999). Douglas 1996, Goffman 1990, Hoffman 2005, Becker 1997

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical, Effectiveness, Safety and Legal domains

Sequential relations

H0005 Assessment element card

Issue: What kind of physical and psychological changes does the implementation and use of the technology bring about and what kind of changes do patients or citizens expect?

Topic: Individual

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue covers whether, from a patient perspective, the technology leads to improvements or harms, or generates any other unexpected effects on functioning.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

ICF 2001, Good 1994

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Effectiveness and Safety Domains

Sequential relations

H0006 Assessment element card

Issue: How do patients, citizens and the important others using the technology react and act upon the technology?

Topic: Individual

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Micro sociological aspect: This issue is about the attitudes, perceptions, preferences, and satisfaction of the patients, citizens using the technology and their important other in relation to the technology. This covers whether, from a patient perspective, any positive or negative issues arise as a consequence of using the technology e.g. feelings of unity or empowerment and existential experiences, e.g. insecurity, worries, hope, anxiety, stigmatisation, person's value as a human being or social status, courage to face life, satisfaction, changes in self-conception.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

ICF 2001: body functions: mental functions (chapter 1:b110-b199), environmental factors: attitudes (chapter 4:, e410-499), Whyte 1997

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Effectiveness and Ethical Domains

Sequential relations

H0007 Assessment element card

Issue: What is the knowledge and understanding of the technology in patients and citizens?

Topic: Communication

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue explores the understanding of the technology in order to describe and decide what guidance and help (e.g. patient information leaflets, counselling processes, need of follow up consultation or help from other professionals) is needed before, during and after the use of the technology.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Health problem and current use, Safety, and Organisational Domains

Sequential relations

H0008 Assessment element card

Issue: How do patients and citizens perceive the information they receive or require about the technology?

Topic: Communication

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue is about the exchange of information from the patients' and important others' perspectives. What are their questions? How do they receive answers?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review,or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational Domain, B0014, B0015

Sequential relations

H0013 Assessment element card

Issue: What are the social obstacles or prospects in the communication about the technology?

Topic: Communication

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

E.g. limitations to decision making in participating or using the technology (dependent, passive user), and possibilities (empowered, active user)..

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review,or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Ethical Domains

Sequential relations

H0009 Assessment element card

Issue: What influences patients’ or citizens’ decisions to use the technology?

Topic: Communication

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

What kind of societal influences lead patients to decide to participate? How do the provisional perceptions about the outcome influence the use of the technology?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, about what works and what does not.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical and Effectiveness Domains

Sequential relations

H0003 Assessment element card

Issue: What kind of support and resources are needed for the patient or citizen as the technology is introduced?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue is about any kind of support and resources (practical, physical, emotional, information, personal, social, nurturing, financial etc.) to ensure the access and satisfactory results. It covers all arrangements or adjustments that may be needed in the major life areas (e.g. alteration of special tasks, working time, adjustments in the physical environment, emotional support, attitudes, reasons for (non)-participation.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Common to all used applications

None

Specific to Screening Technologies 1.0

ICF 2001: environmental factors: support and relationships (chapter 3: e310-399); " activities and participation, chapter 6: d698, structural arrangements of patient’s environment. Dreier 2000, Rapp 1999, Kaufert 2000

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Costs domains

Sequential relations

H0010 Assessment element card

Issue: What kind of social support and resources are needed for the providers as the technology is introduced?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

This issue is about any kind of support and resources (attitude of providers, social gap between providers and patients, number of providers, time, documentation, flow for additional diagnostic or treatment, financial etc.) that need to be mobilized, and organized - or might be released to use the technology with satisfactory results.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a primary study; if there's no time for primary study, the opinion of health care professionals, patients, citizens, or important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational domain

Sequential relations

H0011 Assessment element card

Issue: What kinds of reactions and consequences can the introduction of the technology cause at the overall societal level?

Topic: Major life areas

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Macro sociological aspect: This issue is about the broader society. What social reactions can be expected for example from religious groups, specific patients and citizens organisations and associations and from any other stakeholder groups (social burden with accepted versus stigmatising diseases)? Are special (social) risk groups defined (ethnic, age etc.) and their possible reactions assessed?

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Search for existing literature review, or collect primary studies and if possible conduct a litterateur review, or, if relevant data is not available, conduct a stakeholder analysis and a qualitative/quantitative primary study; if there's no time the systematic collection of opinion of some of the involved stakeholders and interest groups can be done. Patients, citizens and important others can be consulted.

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical, organizational and Legal domains

Sequential relations

H0012 Assessment element card

Issue: Are there factors that could prevent a group or persons to participate?

Topic: Individual

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Do providers select? Are special groups discriminated? It should reflect how the legal regulation takes place in practice. Ethical and social issues have often been considered in academic articles and discussions in the HTA field, but they have rarely been translated into practice.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Implement the best available evidence about social restrictions, social pressure, social attitudes

References
Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Legal domain

Sequential relations

References

Bech U. Risk society. Towards a new modernity. New Delhi: Sage; 1992.

Becker H. Social impact assessment. Social Research Today Series. 1997.

Bijker WE, Law J. (eds). Shaping technology/building society. Studies in sociotechnical change. Cambridge, MA: MIT Press; 1992.

Bourdieu P. Social being, time and the sense of existence. In: Bourdieu P, editor. Pascalian Meditations New York: Polity Press; 2000. p. 206-245.

Cambrosio A, Loch  M, Young A. Introduction. In: Lock M, Young A, Cambrosio A, editors. Living and Working with new medical technologies: intersections and inquiries Cambridge: Cambridge University Press; 2000. p. 1-18.

Coren E, Fisher M. Systematic research review. The conduct of systematic research reviews for SCIE knowledge reviews. Using knowledge in social care research, Research Resource 1. London: Social Care Institute for Excellence; 2006.

Council Conclusions on Common values and principles in European Union Health Systems (2006/C 146/01); http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:C:2006:146:0001:0003:EN:PDF

Devereux G. Primitive psychiatric diagnosis: A general theory of the diagnostic process I. In: Galdston I, editor. Man’s image in medicine and anthropology. Monograph IV ed.: New York academy of medicine; 1963. p. 337.

Dixon-Woods M et al.: Conducting a critical interpretive synthesis of the literature on access to healthcare by vulnerable groups. BMC Medical Research Methodology 2006;6:35.

Douglas M. Risk and blame essays in cultural theory. New York: Routledge; 1996.

Dreier O. Psychotherapy in clients’ trajectories across contexts. In: Mattingly C, Garro LC, editors. Narrative and the cultural construction of illness and healing Berkeley: University of California Press; 2000. p. 237-258.

Facey K.et al. Patients' perspectives in health technology assessment: A route to robust evidence and fair deliberation International Journal of Technology Assessment in Health Care (2010), 26: 334-340 Cambridge University Press)

Foucault M, Martin LH, Gutman H, Hutton PH. Technologies of the self : A seminar with Michel Foucault. Amherst: University of Massachusetts Press; 1988

Goffman E. Stigma. Notes on the management of spoiled identity. London: Penguin Books; 1990.

Good BJ. Medicine, rationality, and experience an anthropological perspective. Cambridge: Cambridge University Press; 1994.

Green J. & N. Thorogood (2005) Qualitative methods for health research. London: Sage publication

Hansen HP. Assessment of literature. In: Kristensen F, Sigmund H, editors. Health Technology Assessment Handbook Danish Centre for Health Technology Assessment, The National Board of Health: Copenhagen; 2007. p. 57-67.

Hansen HP. The patient. In: Kristensen F, Sigmund H, editors. Health Technology Assessment Handbook Danish Centre for Health Technology Assessment, The National Board of Health: Copenhagen; 2007. p. 104-1152. Lehoux P. The problem of health technology. New York: Routledge; 2006.

Harbers H. Inside the politics of technology: Agency and normativity in the co-production of technology and society. Amsterdam: Amsterdam University Press; 2005.

Hastrup K, Olwig KF. Introduction. Siting culture: The shifting anthropological object London: Routledge; 1997. p. 1-16.

Hoffmann B. Toward a procedure for integrating moral issues in health technology assessment. International Journal of Technology Assessment in Health Care 2005;21(3):312-18.

ICF, International Classification of Functioning, Disability and Health. Geneva: World Health Organization; 2001.

Jackson M. Minimal ethnographical intersubjectivity and the anthropological project. Chicago: University of Chicago Press; 1998.

Kaufert PA. Screening the body: the Pap smear and the mammogram. In: Lock M, Young A, Cambrosio A, editors. Living and working with the new medical technologies: Intersections and inquiry Cambridge: Cambridge University Press; 2000. p. 165-183.

Kleinman A. Introduction: Medical Anthropology as Intellectual Career. In: Kleinman A, editor. Writing at the Margin. Discourse between Anthropology and Medicine London: University of California Press; 1995. p. 1-20.

Koivisto J, Vataja K, Seppänen-Järvelä R. Relational evaluation of organizational development activities. International Journal of Public Administration, Evaluation of Welfare Policies, Reforms and Services 2008;31:10-11.

Koivisto J. What evidence-base? Steps towards relational evaluation of social interventions. Evidence and Policy 2007;3(4):527-537.

Latour B. ’Constitution’ I: We have never been modern. London: Prentice Hall; 1993.

Latour B. Science in Action. How to follow scientists and engineers through the society. Cambridge, MA: Harvard University Press; 1987.

Lehoux P. The problem of health technology. New York: Routledge; 2006.

Malterud K. Qualitative research: Standards, challenges, and guidelines. Lancet 2001;358:483-488.

Noblitt GW, Hare RD. Meta-Etnography: Synthesizing Qualitative Studies.California: Sage Publications; 1988

Petticrew M, Roberts H. Systematic reviews in the social sciences. A practical guide. Oxford: Blackwell; 2006.

Popay J, Roberts H, Sowden A et al. Guidance on the conduct of narrative synthesis in systematic reviews. Draft report from ESRC Methods Programme. University of Lancaster: Institute for Health Research; 2004

Rapp R. Testing women, testing the fetus the social impact of amniocentesis in America. New York: Routledge; 1999.

Rogers, Everett M.: Diffusion of innovations / Everett M. Rogers . - 5. ed. . - New York, NY [u.a]  : Free Press , 2003

Rose N. Government, authority and expertise in advanced liberalism. Economy and Society 1993; 22(3):283-299.

Social Care Institute for Excellence. Using qualitative research in systematic review: Older people's views of hospital discharge. London: Social Care Institute for Excellence; 2006. Report No: 9.

Strauss A, Fagerhaugh S, Suczek B, Weiner C. Social organization of medical work. New Brunswick, NJ: Transaction Publishers; 1997.

Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care; 2007;19 (6): 349–57.

Whyte SR. Questioning misfortune the pragmatics of uncertainty in Eastern Uganda. Cambridge: Cambridge University Press; 1997.

Legal aspects

Ingrid Wilbacher, Mirella Marlow, Sirpa Soini

Description

What is this domain about?

The focus of the legal domain in a Core HTA is to detect rules and regulations that have been established to protect the patients’ rights and societal interests. They may be part of patient rights legislation, data protection legislation, or provisions concerning health care personnel and their rights and duties in general. They may also incorporate prior approval processes by competent bodies.

The questions that arise in the legal domain can be roughly divided into six categories of issues which operate at different levels in health care:

  1. Issues related to the central question of who the end-user of the diagnostic technology is;
  2. Issues directly related to the patient and his/her basic rights and freedoms, such as issues of autonomy, informed consent, privacy and confidentiality as well as his/her safety;
  3. Issues related to health care professionals rights and duties;
  4. Issues related directly to the technology in question such as proper authorisation, patent/license issues, price and reimbursement regulation and product safety, guarantee and liability issues
  5. Issues related to the process of acquisition of the technology; and
  6. Issues related to the health care policy at the local, national, European and/or international level, such as distribution of health services.

Why is this domain important?

Legal issues form a substantial part of HTA in the future, since norms of professional ethics are continuously codified into statutes and European Union is producing ever more health technology related legislation.  At the same time one must bear in mind national characteristics of legal systems and health care systems and policies, and thus be sensitive to the limits of exportation of HTA from one country to another.

Already today proper knowledge of relevant legal questions has significant consequences for the decision making in an HTA process, often perceived as part of sociological issues or so called socio-legal issues (Decker 2004, Møldrup 2002).

Legal domain helps identifying the legal barriers which hinder the export and import of HTA results (Drummond & Weatherly 2000, Henshall et al. 2002, Hofman 2005, Terry 2004). It gives insight into the areas of health care legislation where harmonisation is needed, and provides tools for legislative and policy reforms.

Relations to other domains

Issues in Legal domain may overlap with

-          Ethical/ social aspects: How to deal with the socio-economic impact of an adverse event? How are relatives and their legal rights affected?

-          Costs: What is the impact of the legislation? Are there further costs to fulfill legal acts?

Specific features in finding, interpreting or implementing information for this domain

The systematic consideration of legal aspects is expected to contribute to the implementation of HTA results across the Europe. Information sources are contents of relevant international law, EU law and national law. The interpretation of “evidence” in the legal aspects depends on whether a legal regulation exists or not (i.e. for quality), or is planned. Sometimes existence of governmental guidelines and other soft law material makes detection of de facto applicable legal sources challenging.

Methodology

Where to find the information?

International level/ European Union level

-          Human Rights and Biomedicine Convention with its Additional Protocols

-          European Human Rights Convention

-          European Court of Human Rights

-          internet database EurLex

-          decisions of the European Court of Justice

National level:

-          national legislation

-          precedents of national High Courts

Contract level:

-          provider/ payer

In addition to these, a survey on legal literature may be conducted. At European level such journals as e.g. European Journal of Health Law, Medical Law International, Medical Law Review and Medicine and Law may be scrutinised. It is also advised that national libraries’ electronic databases are used to search for relevant international and national monographs and articles on the issue in question. Especially for medical issues and legal aspects articles can be searched in medical databases like Pubmed, where the term "legal" or "legal issues" can be combined with AND for the medical issue.

Interpreting and reporting

The report should follow the different levels of legal sources according to their power of influence on the implementation of the technology under assessment.

i) International law, particularly generated by the Council of Europe. Themost important document in the field of medicine is the Human Rights and Biomedicine Convention with its Additional Protocols. However, these has not been ratified by all European countries, so their applicability needs to be checked in each case. Also various recommendations given by the parliamentary assembly of the Council of Europe may need to be considered. In addition, it may be necessary to investigate whether the European Court of Human Rights has given a relevant decision on the matter based on the European Convention on Human Rights. As new judgements arise in constant manner, knowledge of these needs to be updated regularly.

ii) The level of European Union. While the doctor-patient relationship does not directly fall

under the authority of the Union, the Union may, however, issue health care related legislation

regarding e.g. patient safety, free movement of (health care) goods and personnel etc. Hence, a search of relevant EC legislation is needed. Also, regulation related to free markets and competion law may become relavant in i.e., public procurement.

iii) The level of national legislation. As most of the EC legislation is given in a form of directives, it is necessary in each country to know the relevant national legislation in order to evaluate the exact manner of implementation. Also much of the health care related EC legislation is given as minimum directives and hence a stricter national control may apply.

iv) Agreements with and documentation provided by the technology supplier (Contract level). These will influence the division of risk and liability between the buyer (health care unit) and the supplier and are hence of economic importance to the health care unit in question. It seems unlikely that any uniform standard agreements emerge and it is advised that the scrutiny of these documents is made by a legally educated person.

Assessment elements

I0002 Assessment element card

Issue: Is the voluntary participation of patients guaranteed properly?

Topic: Autonomy of the patient

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

What kind of informed consent procedure is required by the law/binding rules? The use of screening programs is for symptom free (and probably healthy) people, therefore it must not compromise patient safety. Patients should not be pressured into such programs.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

National, international, databases, legal binding guidelines, published laws, related or affected laws

References
Common to all used applications

None

Specific to Screening Technologies 1.0

EU Charter of fundamental rights (2000/C 364/01) Art 3;

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational domain

Sequential relations

I0034 Assessment element card

Issue: Who is allowed to give consent for minors and incompetent persons?

Topic: Autonomy of the patient

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

National law

References
Common to all used applications

None

Specific to Screening Technologies 1.0

Convention on Human Rights and Biomedicine, Art 6 and 7

Content relations
Sequential relations

I0036 Assessment element card

Issue: Do laws/ binding rules require appropriate counseling and information to be given to the user or patient?

Topic: Autonomy of the patient

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

It is important to provide information on the consequences of using the technology in such a manner that the patient can truly understand it.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Convention on Human Rights and Biomedicine, Art 5; Art 12

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

B0004

Sequential relations

I0008 Assessment element card

Issue: Do laws/ binding rules require informing relatives about the results?

Topic: Privacy of the patient

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

The results of a test, or the incidental findings related to use any technology, may indicate that the relatives of a patient may have a medical condition that would need to be addressed. Do the laws/binding rules require breaking the privacy of the original patient in order to inform the relatives of their situation.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Directive 95/46/EC; Convention on Human Rights and Biomedicine Art 10. ECHR Case Law: Z. v. Finland Appl. 22009/93.

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Ethical aspects, B0004

Sequential relations

I0009 Assessment element card

Issue: Do laws/ binding rules require appropriate measures for securing patient data?

Topic: Privacy of the patient

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

At the era of computer-based patient records it is crucial that the health care unit has taken appropriate measures to secure the patient databases. Negligence may lead to liability. Data security has to be provided within a national legal framework when processing claims data or therapeutic information.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Directive 95/46/EC; Convention on Human Rights and Biomedicine Art 10,

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational aspects

Sequential relations

I0011 Assessment element card

Issue: Do laws/ binding rules require appropriate processes or resources to guarantee equal access to the technology?

Topic: Equality in health care

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalNoneYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Is equitable access prescribed in the law or in practice, both at national and international level? The technology can be part of a public program or opportunistic. In many Constitutions equality of citizens covers also access to health care.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

European Social Charter, 1996, ETS No. 163, Art 11 (1., 3.); Convention on Human Rights and Biomedicine Art 3; UN Covenant on Economic, Social and Cultural Rights (1966), Art 12. (Universal declaration Bioethics UNESCO (2005).)

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Social, Ethical and Organisational Domains

Sequential relations

I0012 Assessment element card

Issue: Is the technology subsidized by the society?

Topic: Equality in health care

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Governmental interventions or the lack of them may affect to the expected number of patients. Does subsidization enhance equal access?

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Charter of Fundamental Rights of the European Union (2000/C 364/01). Art 35

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Organisational and Costs Domains

Sequential relations

I0035 Assessment element card

Issue: Do laws/ binding rules require appropriate preventive or treatment measures available for all?

Topic: Equality in health care

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantNoneNo
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

A screening program without the infrastructure to treat the detected diseases appropriately (and with equal access) would be unethical and senseless.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

Additional protocol to the Convention on human rights and biomedicine on Genetic testing, Art 19 Genetic screening for public health purposes. CETS No 203 (2008).

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

In screening model only

Sequential relations

I0015 Assessment element card

Issue: Has the technology national/EU level authorisation (marketing authorisation, registration, certification of safety, monitoring, qualification control, quality control)?

Topic: Authorisation and safety

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesCriticalPartialYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Does the technology require approval and evaluation of a certain committee? Which? How are professional competences and quality of laboratories being governed? A European database of medical devices (EUDAMED) is under construction.

Information sources
References
Common to all used applications

None

Specific to Screening Technologies 1.0

In vitro diagnostic directive (98/79/EC); EUDAMED;

Content relations
Common to all used applications

None

Specific to Screening Technologies 1.0

Safety domain, B0004, B0011

Sequential relations

I0019 Assessment element card

Issue: Does the technology infringe some intellectual property right?

Topic: Ownership and liability

Application-specific propertiesApplicationUsedImportanceTransferabilityCoreOrder
Screening Technologies 1.0YesImportantCompleteYes
Clarification
Common to all used applications

None

Specific to Screening Technologies 1.0

Infringement of intellectual property rights can reduce the use of the technology. The wording of acquisition contract may affect liability sharing between the manufacturer and health care unit.

Information sources
Common to all used applications

None

Specific to Screening Technologies 1.0

Manufacturer, patent data bases, EPO Web site; C-317/05 (ECJ), C-283/03 (ECJ).

References
Common to all used applications

None

Specific to Screening Technologies 1.0

2004/18/EC on public contracts.

Content relations
Sequential relations

References

Web sites:

-          CURIA (The search page for case law of the European Court of Justice)

-          http://curia.europa.eu/jurisp/cgi-bin/form.pl?lang=en

-          COUNCIL OF EUROPE Treaty Office

-          http://conventions.coe.int/

-          EMEA Product Safety Announcements

-          http://www.emea.europa.eu/htms/human/drugalert/drugalert.htm

-          EMEA Marketing Authorisation Withdrawals and Suspensions

-          http://www.emea.europa.eu/htms/human/withdraw/withdraw.htm

-          European Medical Devices Database (EUDAMED) homepage

-          http://eudamed.cec.eu.int/

-          European Patent Convention

-          http://www.european-patent-office.org/legal/epc/pdf/epc_2006_v5_bm_en.pdf

-          European Patent Office – Search page for European patents

-          http://www.epo.org/patents/patent-information/european-patent-documents.html

-          EUR-Lex (The legislation of the European Union)

-          http://eur-lex.europa.eu/en/index.htm

-          HUDOC (The searh page for the case law of the European Court of Human Rights)

-          http://cmiskp.echr.coe.int/tkp197/search.asp?sessionid=9831593&skin=hudoc-en

-          Medical devises homepage of the European Commission

-          http://ec.europa.eu/enterprise/medical_devices/index_en.htm

Altenstetter Christa: EU and Member State medical devices regulation. International Journal of Technology Assessment in Health Care 19(1):228-248 (2003).

Appelbaum Paul S., Lidz Charles W., Meisel Alan & Berg Jessica W.: Informed Consent: Legal Theory and Clinical Practice. 352 pp. Oxford University Press. 2001.

Case Paula: Confidence matters: The rise and fall of informational autonomy in medical law. Medical Law Review 11:208-238 (2003).

Decker Michael: The role of ethics in interdisciplinary technology assessment. Poiesis Praxis 2:139-156 (2004).

Drummond Michael & Wetherly Helen: Implementing the findings of health technology assessments. International Journal of Technology Assessment in Health Care 16(1):1-12 (2000).

Henshall Chris, Koch Pedro, von Below Georg Carl, Boer Albert, Conde-Olasagasti José, Dillon Andrew, Gibis Bernhard, Grilli Roberto, Hardy Charlie, Liaropoulos Lycurgus, Martín-Moreno José M., Roine Risto, Scherstén Tore, Søreide Odd & Züllig Maya: Health technology assessment in policy and practise. Working Group 6 Report. International Journal of Assessment in Health Care 18(2):447-455 (2002).

Hervey Tamara: EU law and national health policies: problem or opportunity? Health Economics, Policy and Law 2:1-6 (2007).

Hervey Tamara & McHale Jean V.: Health Law and the European Union. 469 pp. Cambridge University Press, 2004.

Hoedemaekers Rogeer & Dekkers Wim: Justice and solidarity in priority setting in health care. Health Care Analysis 11(4):325-343 (2003).

Hofman, B.:On value-judgements and ethics in health technology assessment. Poiesis Praxis 3:277-295 (2005).

Hodges Christopher: Do we need European Medical Devices Agency? Medical Law Review 12:268-284 (2004).

Human Genetics Commission (HGC): More Genes Direct. A report on developments in the availability, marketing and regulation of genetic tests supplied directly to the public. December 2007. HGC, London. At the web: http://www.hgc.gov.uk/UploadDocs/DocPub/Document/More%20Genes%20Direct%20-%20final.pdf (July 4, 2008).

Ibarreta D, Bock A-K, Klein C, Rodriquez-Cerezo E. Towards quality assurance and harmonisation of genetic testing services in the EU. IPTS – Institute of Prospective Technological Studies, Publ. EUR 20977, 2003.

Jones Michael A.: Informed consent and other fairy stories. Medical Law Review 7:103-134 (1999).

Kienle Thomas: New forms of medical data collection – should be complemented by a new European privacy standard? European Journal of Health Law 8:27-39 (2001).

Laurie Graeme: Genetic privacy. A challenge to medico-legal norms. 362 pp. Cambridge University Press 2002.

Leith Phillip: The socio-legal context of privacy. International Journal of Law in Context 2:105-136 (2006).

Lierman Steven: European product safety, liability and single-use devices in medical context. European Journal of Health Law 8:204-227 (2001).

Lötjönen Salla: Medical research on patients with dementia – the role of advancedirectives in European legal instruments.European Journal of Health Law 13:235-261 (2006).

MacLean Alasdair R.: Advance directives, future selves and decision-making. Medical Law Review 14:291-320 (2006).

Mason Kenyon & Laurie Graeme: Mason and McCall Smith's Law and Medical Ethics. 820 pp. Oxford University Press, 2005.

Møldrup Claus: Medical technology assessment of the ethical, social and legal implications of pharmacogenomics. International Journal of Technology Assessment in Health Care 18(3): 728-732 (2002).

Oliver Adam: Inconsistent objectives – reflections on some selective health care policy developments in Europe. Health Economics, Policy and Law 2:93-106 (2007).

Pearson, Helen. Genetic testing for everyone. Special report. Nature 2008;453:570-571.

Permanand, Govin: EU Pharmaceutical Regulation: The Politics of Policy-making. 272 pp. Manchester University Press, 2006.

Senden, Linda. Soft law, self-regulation and co-regulation in European law. Where do they meet? Electronic Journal of Comparative Law, vol. 9.1 (January 2005) at www.ejcl.org.Terry Louise M.: An integrated approach to resource allocation. Health Care Analysis 12(2):171-180 (2004).

Terry Nicholas P.: Privacy and the health information domain: Properties, models and unintended results. European Journal of Health Law 10:223-237 (2003).

Wolfberg, Adam J. Genes on the Web – Direct-to-Consumer Marketing of Genetic Testing. The New England Journal of Medicine 2006;355(6);543-545.

European Union

Treaty of Amsterdam amending the Treaty on European union, the treaties establishing the European Communities and related acts. OJ 1997/C  340, 10 November 1997.

Charter of fundamental rights of the European Union (OJ 2000/C 364/01).

Directive 2007/47/EC of the European Parliament and of the Council of 5 September 2007 amending Council Directive 90/385/EEC on the approximation of the laws of the Member States relating to active implantable medical devices, Council Directive 93/42/EEC concerning medical devices and Directive 98/8/EC concerning the placing of biocidal products on the market.  

COM 567 (2005) final. Commission proposal for a regulation of the European Parliament and of the Council on advanced therapy on medicinal products and amending Directive 2001/83/EC and Regulation (EC) No 726/2004. 

Directive 2004/27/EC of the European Parliament and of the Council of 31 March 2004 amending Directive 2001/83/EC on the Community code relating to medicinal products for human use.

Directive 2004/18/EC of the European Parliament and of the Council of 31 March 2004 on the coordination of procedures for the award of public works contracts, public supply contracts and public service contracts.

Directive 2001/104/EC of the European Parliament and of the Council of 7 December 2001 amending Council Directive 93/42/EEC concerning medical devices.

Directive 2001/95/EC of the European Parliament and of the Council of 3 December 2001 on general product safety.

Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use.

Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001 on the approximation of the laws, regulations and administrative provisions of the Member States relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use.

Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on in vitro diagnostic medical devices.

Directive 98/44/EC of the European Parliament and of the Council of 6 July 1998 on the legal protection of biotechnological inventions.Directive 95/46/EC of the European Parliament and of the Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data.

Council Directive 93/42/EEC of 14 June 1993 concerning medical devices.

Council Directive 90/385/EEC of 20 June 1990 on the approximation of the laws of the Member States relating to active implantable medical devices.

Council Directive 89/105/EEC of 21 December 1988 relating to the transparency of measures regulating the prices of medicinal products for human use and their inclusion in the scope of national health insurance systems.

SEC (2006) 1195/4 Consultation regarding Community action on health services.

Council of Europe

Convention for the protection of Human Rights and dignity of the human being with regard to the application of biology and medicine: Convention on Human Rights and Biomedicine CETS No.: 164.

Additional Protocol  to the Convention on Human Rights and Biomedicine concerning Genetic Testing for Health Purposes, adopted on 7 May 2008. (CETS No still open).

Recommendation R (97) 5 of the Committee of Ministers to Member States on the protection of medical data.

Recommendation R (2006) 18 of the Committee of Ministers to Member States on health services in a multicultural society.

Appendices

Appendix 1 Information sources

[100] Technology and procedure registers

Technology registers gather information on the use of a single technology, for example a register on knee total endoprosthesis. A new case is registered in the database every time the technology is used (i.e. a procedure is done, an intervention takes place). In some countries, there is an obligation of reporting indications and consequences of using a technology before approval, and when there is no high quality evidence to establish effectiveness and/or safety of the technology.

The Register of Induced Abortions and Sterilisations in Finland (http://www.stakes.fi/EN/tilastot/qualityreports/inducedabortions.htm)  

Contains data on the annual numbers of induced abortions and sterilisations. It also gives numbers concerning grounds for induced abortions and sterilisations, the gestational stage at which induced abortions are performed, previous abortions and deliveries among abortion patients, methods of abortion used, and the contraceptive methods that abortion patients have used or that have been planned for them.

SCAAR. The Swedish Coronary Angiography and Angioplasty Register http://www.ucr.uu.se/scaar/ covers 100% of the centres and 99% of the procedures in Sweden. It gathers individual data and is linked to other registries (mortality, medicament use, hospitalization).

The Swedish National Board of Health and Welfare maintain a number of registers including the drug register, the cause of mortality register and the registers containing the diagnoses of all hospitalized patients in Sweden. http://www.socialstyrelsen.se/en/Statistics/

[105] Disease registers

Disease registers gather information on the natural history and/or on the management of single diseases. A new case is registered in the database every time a diagnosis of the target disease is made. Some conditions may occur several times in life (i.e. heart attack), thus a single person might be represented several times in the register. When appropriately designed, disease registers allow assessment of the utilisation and diffusion of different diagnostic strategies or technologies in the care of persons with the condition or even to explore variations in the outcomes of different diagnostic interventions (e.g. differences in the consecutive management).

The Swedish National Board of Health and Welfare maintain a number of registers including the drug register, the cause of mortality register and the registers containing the diagnoses of all hospitalized patients in Sweden. http://www.socialstyrelsen.se/en/Statistics/

British Heart Foundation's statistics website is an up-to-date source of statistics on the burden, prevention, treatment and causes of heart disease in the UK http://www.heartstats.org/homepage.asp

[109] Regulatory institutionsEMA

The European Medicines Agency EMA www.ema.europa.eu is responsible for the scientific evaluation of applications for European marketing authorisations for both human and veterinary medicines (centralised procedure).

–        Once a medicine has been granted a Community marketing authorisation by the European Commission, the EMA publishes a full scientific assessment report called a European Public Assessment Report (EPAR) http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/landing/epar_search.jsp&murl=menus/medicines/medicines.jsp&mid=WC0b01ac058001d125&jsenabled=true

–        All medicines for human and animal use derived from biotechnology and other high-tech processes must be approved via the centralised procedure. The same applies to all advanced-therapy medicines and human medicines intended for the treatment of HIV/AIDS, cancer, diabetes, neurodegenerative diseases, auto-immune and other immune dysfunctions, and viral diseases, as well as to all designated orphan medicines intended for the treatment of rare diseases.

–        EMA becomes involved in the assessment of medicines that do not require centralised procedure, in cases where they have been referred to the Agency due to a disagreement in authorisation or use of the medicine between two or more Member States, or due to some other issue that requires resolution in the interest of protecting public health.

–        EMA constantly monitors the safety of medicines through a pharmacovigilance network, and takes appropriate actions if adverse drug reaction reports suggest that the benefit-risk balance of a medicine has changed since it was authorised. 

–        EMA can be considered as the 'hub' of a European medicines network comprising over 40 national competent authorities in 30 EU and EEA-EFTA countries, the European Commission, the European Parliament and a number of other decentralised EU agencies.

–        In many countries over the counter medicines (OTC) are controlled by a regulatory agency. OTC drugs are usually regulated by active pharmaceutical ingredients (APIs), not final products.

FDA

The US Food and Drug Administration (FDA) http://www.fda.gov/default.htm is the federal agency responsible for ensuring that human and veterinary drugs, biological products, and medical devices are safe and effective; cosmetics are safe; and electronic products that emit radiation are safe. FDA also ensures that these products are honestly, accurately and informatively represented to the public.

–        Drug labeling refers to all of the printed information that accompanies a drug, including the label, the wrapping and the package insert. Food and Drug Administration (FDA) requires that drug labeling be balanced and not misleading. The label must be scientifically accurate and provide clear instruction to health care practitioners for prescription drugs and to consumers for over-the-counter drugs and supplements. Labeling regulations require that the statement of ingredients must include all ingredients, in the order in which they are used in the drug. These ingredients must also be identified by their established name.

Standardization and regulatory concerns of medical devices

The government of each European Member State is required to appoint a Competent Authority responsible for medical devices. The Competent Authority (CA) is a body with authority to act on behalf of the government of the Member State to ensure that the requirements of the Medical Device Directives are transposed into National Law and are applied. The CA reports to the Minister of Health in the Member State. The CA in one Member State does not have jurisdiction in any other Member State, but they do exchange information and try to reach common positions.

–        In UK the Medicines and Healthcare products Regulatory Agency (MHRA) acts as a CA, in Italy it is the Ministero Salute (Ministry of Health).

In the EU, all medical devices must be identified with the CE mark.

The ISO standards for medical devices are covered by

–        ICS 11.100.20 standard for biological evaluation of medical devices http://www.iso.org/iso/products/standards/catalogue_ics_browse.htm?ICS1=11&ICS2=100&ICS3=20& and

–        ICS 11.040.01 standard for medical equipment http://www.iso.org/iso/iso_catalogue/catalogue_ics/catalogue_ics_browse.htm?ICS1=11&ICS2=040 .

The quality and risk management regarding the topic for regulatory purposes is convened by ISO 13485 and ISO 14971. Further standards are IEC 60601-1, for electrical devices (mains-powered as well as battery powered) and IEC 62304 for medical software. The US FDA also published a series of guidances for industry regarding this topic.

Packaging standards

Medical device packaging is highly regulated. Often medical devices and products are sterilized in the package. The sterility must be maintained throughout distribution to allow immediate use by physicians. A series of special packaging tests is used to measure the ability of the package to maintain sterility. Relevant standards include: ASTMD1585-Guide for Integrity Testing of Porous Medical Packages, ASTMF2097-Standard Guide for Design and Evaluation of Primary Flexible Packaging for Medical Products , EN 868 Packaging materials and systems for medical devices which are to be sterilized. General requirements and test methods, ISO 11607 Packaging for terminally sterilized medical devices, and others.

Medical Device Directive

http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:31993L0042:EN:HTML

The Medical Device Directive (Council Directive 93/42/EEC of 14 June 1993 concerning medical devices, OJ No L 169/1 of 1993-07-12)  is intended to harmonise the laws relating to medical devices within the European Union. The MD Directive is a 'New Approach' Directive and consequently in order for a manufacturer to legally place a medical device on the European market the requirements of the MD Directive have to be met. Manufacturers' products meeting 'harmonised standards'[2] have a presumption of conformity to the Directive. Products conforming with the MD Directive must have a CE mark applied. The Directive was most recently reviewed and amended by the 2007/47/EC and a number of changes were made. Compliance with the revised directive became mandatory on March 21, 2010.

The US Food and Drug Administration (FDA) http://www.fda.gov/default.htm is the federal agency responsible for ensuring that foods are safe, wholesome and sanitary; human and veterinary drugs, biological products, and medical devices are safe and effective; cosmetics are safe; and electronic products that emit radiation are safe. FDA also ensures that these products are honestly, accurately and informatively represented to the public. Some of the agency's specific responsibilities include:

Drugs

–        product approvals

–        OTC and prescription drug labeling

–        drug manufacturing standards

Medical Devices

–        premarket approval of new devices

–        manufacturing and performance standards

–        tracking reports of device malfunctioning and serious adverse reactions

Radiation-Emitting Electronic Products

–        radiation safety performance standards for microwave ovens, television receivers, diagnostic

–        x-ray equipment, cabinet x-ray systems (such as baggage x-rays at airports), laser products,

–        ultrasonic therapy equipment, mercury vapor lamps, and sunlamps

–        accrediting and inspecting mammography facilities  (Source: FDA web sites)

[111] Other sources

Name

Description

Link

AHRQ

The Agency for Healthcare Research and Quality (US Department of Health and Human Services)

http://www.ahrq.gov/

EDQM.

The European Directorate for the Quality of Medicines & HealthCare

http://www.edqm.eu/en/Homepage-628.html  

FDA

U.S. Food and Drug Administration

http://www.fda.gov/

Google

 

http://www.google.com/

NHS Evidence

Free access to  clinical and non-clinical health information and evidence, guidance and government policy

http://www.evidence.nhs.uk/default.aspx

SIGLE

OpenSIGLE - System for grey literature in Europe (until 2005)

http://opensigle.inist.fr/

TRIP database

Clinical search tool to identify  evidence for clinical practice

http://www.tripdatabase.com/

WHO

World Health Organization

http://www.who.int/en/

Wikipedia

Wikipedia (free, web-based, collaborative, encyclopedia)

http://en.wikipedia.org/wiki/Main_Page

Appendix 3. Shared methodologies

This appendix contains advice on methdologies used in HTA that are general or sharable across several domains. Even if the assessment process may be different in different domains and in the assessment of different technologies, the main phases of the process can be defined. The following phases may need to be gone through in the following order and to the extent that is necessary to find answers for each relevant issue:

1)       Search for systematic reviews (and in the future Core HTA information produced within EUnetHTA Collaboration), and update and adapt its information.

If this is not available

2)       Conduct a systematic review

If this is not feasible or necessary

3)       Conduct a review comprehensive enough to reach information saturation. For transparency, report carefully your sources and method.

If there is no relevant published information available

4)       Conduct a primary study: e.g. interview, modelling, trial

If there is no resources for that

5)       Consult health care professionals and content experts (proxy informants) for their opinions.

General principles for framework and approachScoping

A core HTA

The usual way to start an assessment process is to specify the questions of the review following the so-called PICO structure where the letters stands for

  • Population / patients with the disease of interest
  • Intervention(s), i.e. the technology under assessment
  • Comparison(s), that should serve as reference
  • Outcomes which encompass the endpoints for assessing effectiveness and safety

PICO structure is most likely useful in the domain of clinical effectiveness, safety and economics. Population, intervention and comparison will generally be the same, only the outcomes will be different. In some cases, however, a subpopulation may be of specific interest in a certain domain due to specific details of safety or efficacy/effectiveness.

Diffusion and translation models

Diffusion model (Leavitt 1965)

  • bases on a linear, unidirectional conception of causality
  • considers technology as an exogenous and independent entity
  • seen as a given object which stands outside or above the society, its organisations and actors
  • supposes that technology stays constant
  • sees technology be diffused and transferred from the innovator to different users

Translation model (Leavitt 1965):

  • sees technology as endogenous, as a part of the organisational and use process
  • technology can't be separated from the organisation and its users
  • technology does not stay constant during the implementation process
  • human activity is a part of the technology in question
  • asks "how many and what kind of resources (material entities, time, money, people, etc.) must be mobilised and organized in order to produce satisfactory results from a health technology."
  • technology does not causally affect the organisation and change its social structures
  • organisation and its work processes and social structures have to be organized so that good results can be produced from the technology.

General guidance to critical appraisal of published studies and other informationGeneral

Study types:

  1. Randomised controlled trials
  • randomised cross-over-trials
  • cluster randomised trials
  1. Quasi-experimential studies
  • non-randomised controlled studies
  • before-and-after study
  • interrupted time series
  1. Observational studies
  • cohort study
  • case control study
  • case series
  1. Qualitative studies
  • range of designs: interviews, direct observation, analysis of texts/documents, audio/video recorded speech or behavior, conversation analysis
  1. Economical studies
  • costing and budget impact analysis

Quality assessment of any study type is likely to consider the following (CRD Guidance 2009):

  • appropriateness of study design to the research objective
  • risk of bias
  • other issues related to study quality: choice of outcome measure, statistical issues, quality of reporting, quality of the intervention, generalisability.

Critical appraisal of HTAs

[to be added]

Critical appraisal of systematic reviews

[to be added]

Critical appraisal of guidelines

  • AGREE is an international collaboration improving the quality of clinical practice guidelines by establishing a shared framework for development, reporting and assessment http://www.agreecollaboration.org
  • GRADE Working Group recommendations for grading quality of evidence and strength of recommendations. http://www.gradeworkinggroup.org

Critical appraisal of trials

[to be added]

Critical appraisal of observational studies

There are several checklists or scales on quality available but no consensus about using those. The most appropriate are:

Critical appraisal of diagnostic accuracy studiesQUADAS quality assessment tool (Whiting 2003)

Mandatory items (as in Cochrane handbook)

  1. Was the spectrum of patients representative of the patients who will receive the test in practice?
  2. Is the reference test likely to correctly classify the target condition?
  3. Is the time period between reference test and index test short enough to be reasonably sure that the target condition did not change between the two tests?
  4. Did the whole sample, or random selection of the sample, receive verification using a reference standard of diagnosis (reference test)?
  5. Did patients receive the same reference test regardless of the index test result?
  6. Was the reference test independent of the index test i.e. the index test did not form part of the reference test?

Were the index test results interpreted without knowledge of the results of the reference test?

  1. Were the reference test results interpreted without knowledge of the results of the index test?
  2. Were the same clinical data available when the test results were interpreted as would be available when the test is used in practice?
  3. Were uninterpretable / intermediate test results reported?
  4. Were withdrawals from the study explained?

Additional items

  1. If a cut-off value has been used, was it established before the study was started (pre-specified cut-off value)?
  2. Is the technology of the index test likely to have changed since the study was carried out?
  3. Did the study provide a clear definition of what was considered to be a "positive" test result?
  4. Was treatment started after the index test was carried out but before the reference test was performed?
  5. Was treatment started after the reference test was carried out but before the index test was performed?
  6. Were data on observer variation reported?
  7. Were data on instrument variation reported?
  8. Were data presented for appropriate patient sub-groups?
  9. Was an appropriate sample size included?
  10. Were objectives pre-specified?

STARD - Standard for the reporting of diagnostic accuracy studies (Bossuyt et al., 2003)

Critical appraisal of modelling studies

International Society for Pharmacoeconomics and Outcomes Research (ISPOR) has published a useful article describing the basic guidelines for conducting and reporting modelling studies Weinstein 2003). It can be used also as guidance for using and critically appraising modelling studies. Furthermore, ISPOR is developing more specific guidelines on different modelling methods.

References:

Weinstein et al. Principles of Good Practice for Decision Analytic Modeling in Health-Care Evaluation: Report of the ISPOR Task Force on Good Research Practices - Modeling Studies. Value in Health 2003;6:9-17. http://www.ispor.org/workpaper/healthscience/TFModeling.asp

Critical appraisal of economic evaluation

There are several methodological characteristics to consider, when assessing the quality of an economic evaluation. Several checklists have been published for reporting an economic evaluation, but also to help in identifying the strengths and weaknesses of different studies (e.g. Drummond 1996, Drummond 2005). An example of a checklist (by Drummond 2005) is:

  1. Was a well-defined question posed in answerable form?
  2. Was a comprehensive description of the competing alternatives given?
  3. Was the effectiveness of the programmes or services established?
  4. Were all the important and relevant costs and consequences for each alternative identified?
  5. Were costs and consequences measured accurately in appropriate physical units?
  6. Were costs and consequences valued credibly?
  7. Were costs and consequences adjusted for differential timing?
  8. Was an incremental analysis of costs and consequences of alternatives performed?
  9. Was allowance made for uncertainty in the estimates of costs and consequences?
  10. Did the presentation and discussion of study results include all issues of concern to users?

References:

Drummond M, Manca A, Sculpher M. Increasing the generalizability of economic evaluations: Recommendations for the design, analysis, and reporting of studies. International Journal of Technology Assessment in Health Care 2005;21: 165-71.

Drummond MF, Jefferson TO, on behalf of the British Medical Journal Economic Evaluation Working Party. Guidelines for authors and peer reviewers of economic submissions to the BMJ. British Medical Journal 1996;313:275-283

Critical appraisal of qualitative studies

Examples of quality assessment instruments:

Quality assessment of routine collected statistics and administrative data

Routine collected administrative data (e.g. DRG, discharge databases, reimbursement claims databases) can be useful too, when available. For example sickness funds collect great amounts of information which could be used to analyse utilisation of technology etc. However, analysis of this kind of data might be very time consuming, since data need to be “prepared” before analysis. By definition, these data has been collected for other purposes than research and they cannot be used to answer scientific questions without previous processing. This might not be feasible in the context of an HTA project, due to resource constraints.

The use of routine collected statistics has several limitations. The reliability of the diagnosis varies and usually it is not possible to differentiate between different stages of the disease. Even the validity of the coding of causes of death may be variable, and in some countries it is known to be very limited.

Own analysis of administrative data often requires authorization from the data owner, which in some countries might be difficult to obtain due to issues of privacy protection and confidentiality.

Critical appraisal of register data

ISPOR is developing guidelines for patient registry data: http://www.ispor.org/sigs/PR_analysis_data_mgt.asp

General guidance for conducting researchGuidance for conducting a systematic review

The principles on how to conduct a systematic review are nowadays widely agreed upon and most of the methodologies published by different organisations vary only in details. The most widely used methodology applies the principles and instructions of the Cochrane Collaboration (Cochrane Handbook 2005). The same principles and procedures can be applied to systematic reviews of non-randomised comparative study designs or observational studies (CCOHTA 2001).

In the process of conducting a systematic review it is wise to check whether others have already published systematic reviews covering the same intervention. There may be one on the topic of interest, which is sufficiently comprehensive, satisfies the requirements on methodological quality, and meets the research questions.

Reviews should follow an a priori defined protocol, and possible deviations from the protocol should be described.The literature search should be comprehensive and transparent.Restrictions in locating studies should be given (year, language, population etc).The process of identifying the relevant information should be explicitly described including the search strategy (which may be presented separately as appendix) and the inclusion and exclusion criteria. The selection process is iterative and includes two or more phases: first the selection of potentially relevant titles and / or abstracts, then the selection of full papers that are considered as potentially relevant, then the selection of the final set of studies. Using a flow chart to display the process of study selection as recommended by the QUOROM group(Moher 1999) has been proven to be very informative. The flow chart describes the number of citations and studies identified at each step as well as the number excluded, with reasons for exclusion at each stage ending at the final number and types of studies included.

Since some of the decisions in the selection process will always remain subjective, engaging two or more reviewers in this process will improve transparency and objectivity, especially when the report presents information about the degree of agreement or disagreement among reviewers and how disagreement was solved (e.g. consensus after consulting a third researcher).

Comprehensive and informative evidence tables about the methodology and the content of the individual studies are the best guarantor for transparency and reliability. They should allow a judgement of the similarities and differences of the included studies and should provide the basis for the conclusions of the review.

References:

The Cochrane Handbook for Systematic Reviews (Cochrane Handbook 2005)

CRD guidance for systematic reviews (Khan 2001),

CONSORT statement (Altman 2001, Piaggio 2006, Moher 2001)

QUOROM statement for reporting of systematic reviews (Moher 1999, Clarke 2000)

Checklist for HTA reports by INAHTA (Hailey 2003)

GRADE Working Group recommendations for grading quality of evidence and strength of recommendations Atkins 2004, Guyatt 2006).

(Moher 1999)

Guidance for modelling

International Society for Pharmacoeconomics and Outcomes Research (ISPOR) has published an article describing the basic guidelines for conducting and reporting modelling studies. ISPOR is also developing more specific guidelines on different modelling methods.

References

Weinstein et al. Principles of Good Practice for Decision Analytic Modeling in Health-Care Evaluation: Report of the ISPOR Task Force on Good Research Practices - Modeling Studies. Value in Health 2003;6:9-17. http://www.ispor.org/workpaper/healthscience/TFModeling.asp

Guidance for conducting a register study

[to be added]

Guidance for conducting survey (questionnaire, interview)

[to be added]

General guidance for synthesis

GRADE is s systematic approach for grading evidence  http://www.gradeworkinggroup.org

Meta analyses of accuracy studiesNo heterogeneity

A forest plot of sensitivity versus specificity with 95 % confidence intervals can be used whenever the results from two or more comparable studies are included in the review. Forest plot illustrates the range of results, enables the reader to assess heterogeneity, and possible trade-off between sensitivity and specificity, and may show the summary estimate where pooling is appropriate.

Another option is to plot pairs of sensitivity and 1 - specificity from original studies on a ROC plane. If sensitivity or specificity is constant or if there is linear relationship between them, simple summary measures for sensitivity, specificity, or likelihood are adequate.

When pooling pairs of sensitivity and specificity, the statistical model used depends on the studies selected. Fixed effect model assumes the studies to represent a random sample of one large common study. The differences between study outcomes are considered to be the result of random error. The model weights individual studies based on the inverse variance of the accuracy or the number of participants. Random effects model assumes the differences between studies to be due to real differences between the study populations and procedures. A more complex mathematical model is used to weight studies. Separate estimates of mean sensitivity and specificity underestimate test accuracy.

Heterogeneity present

When forest plot and heterogeneity testing (see below) shows that there is significant heterogeneity in sensitivities and specificities across studies, it is not appropriate to report the pooled values of sensitivity and specificity as a summary estimate. Instead, further analysis of the heterogeneity detected is needed, and it starts with examining of threshold effect. Threshold effect can be seen in forest plot if there is an inverse relationship between sensitivity and specificity. If this is not apparent the results should be plotted to a ROC plane to examine the threshold effect further.

Paired estimates of sensitivity and 1 - specificity in original studies are plotted in a ROC plane. Regression model is used to fit the SROC curve (Moses 1993). If the SROC curve is symmetrical around the line where sensitivity equals specificity, the studies share one common DOR, and any variability is due to differences in the test threshold. In statistical terms, if in the model the slope b (estimated regression coefficient) is not statistically significant and approaches zero, The SROC will be symmetrical.

Spearman's test for a nonparametric distribution has also been used to test for a threshold effect. Using this method, the correlation between sensitivity and 1-specificity for each study is measured and a Spearman rank correlation coefficient > 0.6 is used to confirm variation across studies due to a threshold effect (Moses 1993). If the correlation is poor (Spearman rank correlation coefficient < 0.6) the variation between studies is attributed to other differences. This is a crude measure and is not generally recommended.

Threshold effect only

If there is symmetry in the SROC curve, DOR is constant regardless of the diagnostic threshold, and any variability in the paired sensitivity and specificity between different studies is due to differences in the test threshold. In this case, SROC curve represents the most informative synthesis of evidence about test accuracy and the pooled DOR is a useful single summary measure.

SROC curve does not provide one summary estimate of sensitivity and specificity but it allows assessment of their interdependence. Summary DOR (SDOR) of the test and a comparator test can be presented with 95 % CI:s to compare differences in diagnostic performance. The area under SROC curve and its 95% confidence interval provides a global summary of overall test accuracy. The point on the curve where sensitivity equals specificity, the Q* statistics, can also be used as a summary measure of the accuracy of the test. These summary measures can also be used to compare the accuracy of two test strategies. Software for diagnostic meta-analysis include Meta-Test, Meta-Disc, Stata and SAS.

Heterogeneity that is more than just threshold effect

If the slope b in the SROC model is statistically significant, the SROC will be asymmetrical and the DOR changes along the threshold. In such cases advanced methods for fitting the SROC is used. Advanced methods to pool are indicated if heterogeneity in the results can be attributed to known sources of variation (see above Chapter Assessing heterogeneity). Otherwise the interpretation of the summary estimate is not possible (Lijmer 2002).

Possible sources of variation include

  1. Chance
  2. Different threshold
  3. Different study designs, methods, biases: different reference standard, different versions of the technology
  4. Variation by clinical subgroups in terms of age, severity or stage of disease, prevalence of the target condition, differential diagnoses, and setting
  5. Unexplained heterogeneity

If differences in the results can not be attributed to these known sources of heterogeneity, then pooling of the results to one summary estimate should not be attempted, because its interpretation will be impossible (Lijmer 2002).

Methods to test for heterogeneity (Medical Services Advisory Committee 2005):

  1. Plot the sensitivity and specificity from each study with their 96% confidence interval in a table and/or forest plot to illustrate the range of estimates and identify outliers.
  2. If sufficient data are available, plot the paired sensitivity and 1-specificity results for each study on the ROC plane to detect heterogeneity and identify outliers. A small number of studies will limit the power of regression to detect heterogeneity.
  3. Use a chi-square test for heterogeneity (Cochran's Q test) or Fischer's exact test for small studies to test the hypothesis that there is no statistically significant difference in the sensitivity and specificity reported.

Advanced models enable incorporation of covariates, e.g. population subtype in the meta-regression analysis. Poor reporting of primary studies may though lead to biased estimates. The two main advanced models are hierarchical SROC and bivariate meta-regression, and they are mathematically identical (Harbord 2007). Syntax to run these models in SAS, STATA, WINBUGS, S-PLUS and R are or will be available. Hierarchical SROC (HSROC) produces informative summary measures with confidence ellipses (Reitsma 2005).  Model is infrequently used, probably due to complex fitting.

More reading: (Deeks 2001, Deville 2002, Kester 2000, Irwig 1995)

General guidance for interpretation

[to be added]

General guidance for reportingTransferability

To allow transfer of data across countries, HTAs have to be sufficiently transparent and distinguish between evidence (“facts”) and judgements (including values and preferences). Value judgements and preferences (of individuals or of health care systems) have to be labelled as well as the anticipated influence in transferring the result from one health care system to another. There will be situations where only the body of evidence of an HTA can be used, but the data need to be interpreted in the context of the health care system and the prevailing values.