Downloaded from bmj.com on 13 February 2006 A good death James Grogono, Ann Bowling, Matthew Thomas, Richard Day, Margaret Brown, Andrew Fowell and Ilora Finlay BMJ 2000;320;1205- doi:10.1136/bmj.320.7243.1205 Updated information and services can be found at: http://bmj.com/cgi/content/full/320/7243/1205 These include: References This article cites 6 articles, 5 of which can be accessed free at: http://bmj.com/cgi/content/full/320/7243/1205#BIBL 1 online articles that cite this article can be accessed at: http://bmj.com/cgi/content/full/320/7243/1205#otherarticles Rapid responses You can respond to this article at: http://bmj.com/cgi/eletter-submit/320/7243/1205 Email alerting Receive free email alerts when new articles cite this article - sign up in the box at service the top right corner of the article Notes To order reprints of this article go to: http://www.bmjjournals.com/cgi/reprintform To subscribe to BMJ go to: http://bmj.bmjjournals.com/subscriptions/subscribe.shtml  Downloaded from bmj.com on 13 February 2006 Letters Website: bmj.com Email: letters@bmj.com A good death out Britain, which is perhaps inevitable when the NHS largely relies on charity to provide them. This is not a value free commentary. I Sharing control in death: the role of an amicus mortis at the end of last year when speak as one who recently cared, alone, for “amicus mortis’’ my father died. My older brother, enviably in my 79 year old father, who died at home Editor—May I offer one further ingredient full retirement, played the main part. He from stomach cancer in a rural part of Eng- to a good death as discussed by Smith in his closed down his home in the Carolinas for a land. The individual health professionals editorial1—having an “amicus mortis,” a couple of months for the purpose. He involved (general practitioner and district friend at death. Most items on his list of endorses this view. nurse) were caring, kind people, doing their principles use the word control or imply it, The chosen person must have time and best within an extremely limited system. yet the very process of death entails losing love and prescribing power. Perhaps it is an The only specialised professional was a control. Control of strong drugs is especially unfair advantage to have a doctor-husband Macmillan nurse based at the hospital difficult for the one who is dying. An amicus or doctor-sons available, but prescribing whose role was to advise the very busy mortis makes it easy. I wrote the following power can easily be delegated, and the other district nurses on aspects of their workload within days of my wife’s death from cancer attributes are just as important. relating to terminal care; no help or four years ago, and it was read at her funeral. I fully agree with Smith that there is support was given by Marie Curie nurses “Towards the end I was given the nothing macabre or morbid about thinking because they were “scarce.” The outcome privilege of care. I don’t want to belittle the of death and planning your funeral through- was lack of preparation for the distressing role of the care team. None the less, I was the out life, but I suggest that it is equally impor- final stages, and great difficulty contacting lucky one in charge, especially at night, and tant to choose and cultivate your amicus doctors and nurses out of hours at times of my task was an easy one, aided by small mortis—and see to it that he (or she) doesn’t need. This, in turn, resulted in totally doses of morphine towards the end. die first. inadequate pain relief and lack of help “She had no pain, no distress, no loss of James Grogono consultant surgeon with incontinence and distress in the final dignity, no catheters, none of the things my South Buckinghamshire NHS Trust, 24 hours. High Wycombe, Buckinghamshire HP13 6PS patients in hospital have to put up with. In grog@connect-2.co.uk the last week our nightly family parties had to be in her room. Her last hours were 1 Smith R. A good death. BMJ 2000;320:129-30. tranquil.” (15 January.) The role of amicus mortis is life enhanc- ing, and there should thus be no shortage of Research on dying is scanty supply. I had a further small dose of being an Editor—It is true that no one can fully answer Smith’s question about the state of dying in Britain.1 The current fashion for Advice to authors evaluative studies of health service outcomes We prefer to receive all responses electronically, and the low priority given by grant bodies to sent either directly to our website or to the descriptive research (generally discarded as editorial office as email or on a disk. Processing needs assessment) have led to an emphasis your letter will be delayed unless it arrives in an instead on the costs and outcomes of electronic form. treatments for the living. Admittedly, surveys We are now posting all direct submissions to of terminal care and bereavement have been our website within 24 hours of receipt and our conducted.2 3 But these, by the very nature of intention is to post all other electronic survey design, can only tap the surface. Nei- submissions there as well. All responses will be ther has the more detailed body of research eligible for publication in the paper journal. carried out on selected groups of patients Responses should be under 400 words and (including those in hospices) led to the wider relate to articles published in the preceding profession or the public being fully month. They should include <5 references, in the informed about death. Vancouver style, including one to the BMJ article Smith’s principles of a good death are to which they relate. We welcome illustrations. timely and greatly welcomed.1 While infor- Please supply each author’s current mation about dying should be more widely appointment and full address, and a phone or available, and regarded as less taboo, an fax number or email address for the informed profession is also required to act corresponding author. We ask authors to declare in the best interests of patients: to provide any competing interest. Please send a stamped them with preparatory information to addressed envelope if you would like to know demand and initiate timely and appropriate Death’s embrace: Death and a Woman by Hans whether your letter has been accepted or rejected. Baldung Grien from the exhibition help. When people are fortunate enough to “Grunewald and his contemporaries: Letters will be edited and may be shortened. have access to a hospice or a good palliative paintings from the Kunstmuseum, Basel” in bmj.com care team they may be given adequate the National Gallery, London, until 21 May letters@bmj.com support, information, and preparation. But 2000 such services are patchily provided through- BMJ VOLUME 320 29 APRIL 2000 bmj.com 1205 Letters Downloaded from bmj.com on 13 February 2006 This is just one example of what can allow us to measure quality of death in 1 Smith R. A good death. BMJ 2000;321:129-30. (15 January.) happen in the face of a patchy infrastructure medical and geriatric practice. 2 Department of Public Affairs, University of Toronto. News for terminal care in this country—but it is Matthew Thomas consultant physician release. Toronto: DPA, 2000. (2 February.) 3 South Australian Parliamentary Select Committee on the one example too many. Richard Day consultant physician Law and Practice Relating to Death and Dying. Report. Department of Medicine for the Elderly, Adelaide: SAPSC, 1991. Ann Bowling professor of health services research Poole Hospital NHS Trust, Poole BH15 2JB 4 Pincombe J, Brown M, Ballantyne A, Thorne D, McCutch- CHIME/Department of Primary Care and eon H. Care of dying patients in the acute hospital: an Population Sciences, University College London, The results of this audit were first presented to the exploratory study. Progress in Palliative Care (in press). London NW1 2DA British Geriatric Society in Cork in April last year. 5 Pincombe J, Brown M, Ballantyne A, Thorne D, McCutch- eon H. Care of dying patients in the acute hospital setting. 1 Smith R. A good death. BMJ 2000;320:129-30. 1 Smith R. A good death. BMJ 2000;321:129-30. Report to the NHMRC. Magill: University of South Australia, (15 January.) (15 January.) 2000. 2 Bowling A, Cartwright A. Life after a death. A study of the eld- erly widowed. London: Tavistock Publications, 1982. Principles of palliative care are yet to be 3 Seale CF, Cartwright A. The year before death. Aldershot: applied in acute hospitals Care pathway in Wales aims to improve Avebury, 1994. care of dying patients Editor—Smith states in his editorial, “There is a suspicion that for the majority who die Editor—People always have died and Quality of death can be measured outside in acute hospitals or nursing homes the always will die.1 Palliative care has now come hospices experience is bad.”1 Similar stories abound out of hospices and is accepted as a Editor—Henry Fielding said: “It hath often in Australia even with its well developed pal- mainstream specialty, influencing care be said that it is not death, but dying which is liative care services. However, palliative care across the NHS. In Wales a care pathway, terrible,” though striving for a good quality is accessible to only some terminally ill developed from the work of Ellershaw et al,2 of death for patients1 should not prevent us patients and usually those dying of cancer. is being introduced across the whole region saving life and treating disease. The “good dying” in hospitals still eludes in various care settings, including acute hos- Most scientific work on death applies to most in Western countries, as indicated by a pitals and nursing homes. death from cancer in hospices, which is recent bequest in Toronto, Canada.2 The aim is to improve care of dying greatly different from the experience of Research commissioned by the South patients by implementing agreed evidence death in geriatric wards—for example, in the Australian parliament in 1991 found that the based clinical guidelines facilitated through number of unexpected deaths. We audited majority of respondents considered public the care pathway. The Clinical Effectiveness the quality of expected deaths in acute inpa- hospitals to be unsatisfactory in providing Support Unit and the National Assembly of tient geriatric practice by measuring care for terminally ill patients.3 They felt Wales are supporting the process and whether five standards were attained: excluded from medical decision-making and evaluation. x Patients should be free from symptoms had problems in communicating with hospi- Results of the pilot study in Bangor have (pain, anxiety, dyspnoea, pressure sores) tal specialists. This and other anecdotal shown important changes in practice, with x Death should occur in familiar evidence prompted a group of researchers in improved analgesic prescribing. The avail- surroundings—that is, the patient should not South Australia to investigate the care of ter- ability of analgesics to control pain rose be moved within three days of death minally ill patients in acute hospitals. from 72% to 98% when the care pathway x Relatives should be aware that the patient Two studies have now been completed was implemented. The care pathway thus is dying and be present if desired observing the care of terminally ill patients anticipates potential problems and empow- x Necropsy should be requested if cause of during their last six days of life in medical ers carers and nurses to give timely and death is unclear wards in two acute hospitals.4 5 The findings effective interventions. x The patient’s general practitioner should indicate that there are barriers to the care The pathway ensures that the diagnosis be informed of the death (by telephone when patients are dying, including the strict of dying is not attached inappropriately, within 24 hours) and diagnosis (by discharge adherence to hospital routine. Inexperi- either too early or late. The relatives are summary within seven days). enced health professionals without an informed of anticipated events and retain The table shows the results. We identi- understanding of the philosophy of pallia- choices and control. fied issues around the time of death related tive care or the skills required often We agree with many of the principles of to symptom control and communication undertook the care. The data showed that it a good death3: indeed, they underpin the with general practitioners that needed could be a very isolating experience for pathway. We do not prescribe a lingering attention within our department. patients left alone in a side room, most of death, but all must be aware that the precise whom were unresponsive. The presence or moment of death is unpredictable and not absence of family members influenced the in our or anyone’s control. Results of audit of quality of death in acute amount of care received.4 5 The results of Andrew Fowell Macmillan consultant in palliative inpatient geriatric practice. Values are numbers these two studies suggest that the principles medicine (percentages) of patients unless stated otherwise of palliative care are yet to be included in the Ysbyty Gwynedd, Bangor LL57 2PW 1995 1998 culture of acute hospitals. It is as though the Ilora Finlay Marie Curie consultant in palliative hospital environment reflects the busyness medicine Standard (n=28) (n=16) University of Wales College of Medicine, Cardiff Patient was symptom free 19 (67) 11 (69) of everyday life in society, which still denies CF14 7XL Patient died in familiar 18/24 (75)* 15 (94) the naturalness and inevitability of death. 1 Smith R. A good death. BMJ 2000;320:129-30. surroundings Where to from here? The questions go (15 January.) Relatives were aware 21/26 (81)† 14 (88) beyond the boundaries of medicine and 2 Ellershaw J, Foster A, Murphy D, Shea T, Overill S. belong to the human race. While the sanctity Developing an integrated care pathway for the dying Necropsy performed if cause of 1/2‡ 1/2‡ patient. Eur J Palliat Care 1997;4: 203-7. death unclear of life is overridden so often by the culture of 3 Debate of the Age Health and Care Study Group. The GP informed within 24 hours 6 (21) 2 (13) war, and death is portrayed as a successful future of health and care of older people: the best is yet to come. London: Age Concern, 1999. GP informed within 7 days 6 (21) 6 (38) outcome, we will live with this paradox. There GP=general practitioner. are bigger questions still to be asked about *Four patients died within three days. death and dying in society before deaths in †Two patients did not have any relatives. ‡Reflects low rate of necropsy. hospital are attended with humanity and Older Americans hold on to compassion. I, for one, hope my last days are life dearly not alone in a side room in an acute hospital. Standards of quality of death are univer- But I welcome the debate both in medical Editor—The study of Salkeld et al in sally applicable, though the different journals and in all facets of society. Australia of the choices of older patients in emphasis between, say, medical and hospice Margaret Brown lecturer hypothetical illness scenarios is particularly practice needs to be recognised. We believe University of South Australia, Magill, pertinent to those of us dealing with patients that development of these standards will South Australia 5072 in nursing homes.1 We are struck by the 1206 BMJ VOLUME 320 29 APRIL 2000 bmj.com  Downloaded from bmj.com on 13 February 2006 Letters contrasting attitudes and expectations of the New models of journals must 1999 report, 99 percent of Costa Rica’s population elderly patients with whom we are in daily has access to contraception, along with 96 percent be sought of Haiti’s, 93 percent of Zimbabwe’s and 89 contact. In our experience, most older percent of Peru’s. Yet all of these countries have Americans hold on to life very dearly and Editor—Alderson and Roberts argue that fertility rates well above the level needed to replace usually opt for even noxious treatments, the nature of journals (the fact that in a those who die . . . If fertility levels in developing such as chemotherapy, to gain a few months commercial environment editors choose the countries depend first and foremost on the or years of life that is reduced in quality. more interesting articles on offer) is distort- number of children desired by parents, then the These attitudes are consistent with observa- ing science.1 Quite so. But the answer is not population control movement faces a wrenching tions in patients of 80 years or more who to attempt to change the nature of journals dilemma. Purely voluntary programs will do little (goodness knows, we need better communi- to reduce fertility. Only those population pro- were in hospital, many of them in poor cation) but to seek new models that will grams that override parental preferences through health.2 When they were asked to choose bribes, bullying, threats or outright coercion will between their current state of health or a separate the validation of science from the lower birthrates significantly.5 shorter life in excellent health, over two reporting of it. Electronic publishing (where selection We believe that a change in attitude is thirds were unwilling to exchange even 10% on economic grounds is no longer an issue) possible if we raise the global standard of of life expectancy for the benefit of excellent will enable us to do this, provided we don’t living for couples. This, however, will take health.2 get caught up in trying to make commer- both time and effort, and by definition it will Basic expectations also seem to differ not succeed if true entrapment exists. The cially based journals what they can never between elderly Australians and Americans. dilemma that King faces is, therefore, to let be—bias free. While many would attest to having had a fair us know whether his stated objective—a one Tim Albert trainer or good long life, we rarely hear any sugges- child family—justifies the stern, unpopular, Tim Albert Training, 284 High Street, Dorking, tion that one has lived overlong or ‘‘on bor- Surrey RH4 1QT totally coercive, and, in our view, unethical rowed time at the expense of younger tatraining@compuserve.com measures that will be required. people.” Patient choices are amenable to Giuseppe Benagiano director general influence and alteration.3 The choices of the 1 Alderson P, Roberts I. Should journals publish systematic Istituto Superiore di Sanità, Viale Regina Elena reviews that find no evidence to guide practice? Examples subjects in the study of Salkeld et al from injury research. BMJ 2000;320:376-7. (5 February.) 299, 00161 Rome, Italy conformed to attitudes and opinions of the Michele Ermini associate professor, obstetrics and investigators, who note in several places that gynaecology University of Rome (la Sapienza), Viale del all subjects had already “exceeded average life expectancy.” Certainly people are unwill- Population, consumption, and Policlinico 155, 00161 Rome ing to lose independence of function or entrapment 1 Editor’s choice. Benign uproar. BMJ 1999;319(7215). decision, but we cannot readily accept (even (9 October.) 2 King M. The US Department of State is policing the popu- for Australians) that 80% would rather be Raise living standards to reduce lation policy lockstep. BMJ 1999;319:998-1001. (9 October.) dead than suffer a hip fracture and population growth 3 Abbasi K. King in a maverick style. BMJ 1999;319:942. (9 October.) subsequent admission to a nursing home. (Is Editor—The BMJ marked the passing of 4 King M, Elliott C. To the point of farce: a Martian view of there any incidence of suicide or request for the 6 billion mark of the globe’s population the Hardinian taboo—the silence that surrounds popula- tion control. BMJ 1997;315:1441-3. death when such events are actually faced?) with a series of articles over which Maurice 5 Glendon MA, Haynes M. Look again: lots of parents want This prompts further consideration that King cast his shadow.1 In his article King the children. International Herald Tribune 1999 Oct 30. advance preferences or directives may not raised the spectre of an international * Charles Elliott and Maurice King will be ** conform with those when crisis is at hand. In conspiracy led by the United States which is arguing the evidence for the policing of the abstract discussion (or in living wills) tube aimed at preventing free discussion of the lockstep at greater length later in the year. feeding, having to breathe on a respirator, possibility that certain countries in the and even intravenous treatments are often developing world may have become abjured, but such prohibitions are rarely car- entrapped.2 To King, entrapment occurs Improve access to contraception to curb ried forward when the acute illness is faced. when a population exceeds the carrying population growth The attitudes of those already in nursing capacity of its ecosystem and is unable to Editor—Loefler is right in stating in his let- homes with hip fractures might be com- buy in extra food or to migrate elsewhere. ter that in terms of population the adverse pared with those of the subects in the study Personally, we do not believe in this “influence of the Pontiff reaches far beyond by Salkeld al. Our rehabilitation department conspiracy theory and find it significant that, Catholics.”1 The Pope says that contracep- in moving along this path, King has been tives are “anti-life” yet they are cost effective deals with over 80 patients recovering from abandoned even by Charles Elliott, his coau- in reducing the mortality of both mothers hip fracture annually; many remain for long thor for previous articles.3 and infants.2 3 He is one of the people who term care. While some of those receiving King’s argument has for years been influences our culture by portraying “any long term care are depressed, most soon predicated on the idea that entrapment quantitative concern for population as accommodate, adjust, and have a reasonable exists in Africa and Asia on a large scale and intrinsically coercive.”4 As a contributor to quality of life. that therefore a policy of one child for each the 9 October issue of the BMJ in which Randall McShine fellow of geriatrics family must be widely adopted.4 Assuming population was discussed, I think that our Gerson T Lesser assistant professor of geriatrics that his diagnosis is correct, King fails to glesser@jhha.org omission of a frontal attack on the Pope explain how he plans to achieve such drastic derived from a fear that this might be coun- Antonios Likourezos research associate measures in countries where couples want Department of Geriatrics and Adult Development, terproductive. This highlights his baleful Mount Sinai School of Medicine, and Jewish Home to have, on average, six or more children. influence. and Hospital, 120 West 106th Street, New York, China adopted such a policy for years Could the lack of any mention of popu- NY 10025, USA and, although it has had remarkable success, lation by Haines et al in their editorial on the it has been successful using measures that International Poverty and Health Network have often been considered coercive. India be another manifestation of this influence?5 1 Salkeld G, Cameron ID, Cumming RG, Easter S, Seymour J, Kurrle SE, et al. Quality of life related to fear of falling established a similar, although less draco- As the authors acknowledged, poverty has and hip fracture in older women: a time trade off study. nian, policy of two children per family and many dimensions and not all could be men- BMJ 2000;320:341-5. (5 February.) has not succeeded in enforcing it easily. tioned but population growth is the unrec- 2 Tsevat J, Dawson NV, Wu AW, Lynn J, Soukup JR, et al. Health values of hospitalized patients 80 years or older. Indeed, it has been pointed out that: ognised multiplier of most major problems JAMA 1998;279:371-5. in the world. 3 Malloy TR, Wigton RS, Meeske J, Tape G. The influence of In much of the Third World, contraceptives are treatment descriptions on advance medical directive deci- readily accessible yet fertility rates remain high. It is a vicious cycle: for those living in sions. J Am Geriatr Soc 1992;40:1255-60. According to the UN family planning agency’s poverty in rural areas it seems advantageous BMJ VOLUME 320 29 APRIL 2000 bmj.com 1207 Letters Downloaded from bmj.com on 13 February 2006 to have more not fewer children, given high War and the preparation for it are both indicates massive heterogeneity between the infant mortality and the “social security” the causes and the results of poverty.2 This is results of the studies included. In the factor (“every new mouth has two hands”). most obvious in countries such as Afghani- therapeutic trials the P value of 0.046 also Yet with more children the limited resources stan and in many countries in Africa, notably indicates heterogeneity. of any poor family are further partitioned, Angola, which have been in the midst of war Meta-analysts faced with such bringing smaller shares for all and exacer- for decades; at the same time these are heterogeneity have three choices: they may bating poverty. among some of the world’s poorest coun- ignore the heterogeneity and pool the There is an added environmental tries. Many poor countries spend more on results with a fixed effects model; they may dimension, highlighted by the equation I = their military than on health and education use a random effects model, which takes the PAT, in which I stands for the impact on the combined.3 We welcome last year’s cancella- heterogeneity into account; or they may environment; P is the population; A is per tion of the debts of the 25 poorest countries decide not to pool the results. In this capita affluence (with intrinsic effluence); announced by Gordon Brown, the chancel- instance the authors chose to use a fixed and T is the per capita impact of technology. lor of the exchequer, who also specified that effects model despite the heterogeneity. In (The T factor is lower when the technology the money saved should not be spent on consequence the confidence intervals of the is greener.) Science can and will deliver arms. Even more recently, the retiring presi- pooled estimates are narrow and signifi- reductions in the T factor but there are dent of the International Monetary Fund, cance is imputed. A random model might absolute constraints. Greener technologies Michel Camdessus, called on the developed well lead to a summary estimate that does are expensive and are often beyond the world to stop exporting arms to poorer not reach significance. The data extracted means of pre-industrialised countries. countries. from the individual trials have not been Everyone wants (and poorer people The adverse effects that spending on published in bmj.com so it is not possible to deserve and should be actively helped to arms has on health care are evident also in check this. obtain) more A factor (affluence). Therefore developed countries: the NHS underwent its The authors have carried out a sensitiv- the P factor (population) should not be left annual winter turmoil and there has recently ity analysis of the therapeutic trials and out of the equation by those who wish to been a study of the regional inequalities that noted that the single regimen trials alleviate pollution (which hits the poorest exist in health care in the United Kingdom.4 measured only short term outcomes and people hardest) and ill health related to pov- Of the members of the North Atlantic were of lower methodological quality. In erty. Population should not be seen as a con- Treaty Organisation and the European contrast, the two trials of triple treatment stant to which we should forever try to Union, the United Kingdom is among those measured outcomes at 12 months and were adjust. Human numbers can be stabilised, countries that spend the most on their mili- of higher quality; these trials showed a much tary, and it is likely to remain so if plans to smaller summary estimate of eradication, non-coercively. We are failing to provide the spend £34bn on the Eurofighter and £15bn which barely reaches significance (odds ratio contraceptive services most women now on new aircraft carriers go ahead.3 The 1.4, 95% confidence interval 1.0 to 2.3). want.2 3 United Kingdom is also among the coun- There is clinical as well as statistical At the end of their article Haines et al tries that spend the least on health care. Yet heterogeneity between these two groups of had seven excellent recommendations. Lord Robertson at NATO, Javier Solana at studies, so a summary estimate that com- Knowing the authors, I am confident that the European Union, and sources leaking bines both is of doubtful meaning. Thus in they will be pleased to agree one more: mak- information from the Ministry of Defence my opinion the best option is not to pool the ing certain that every woman in the world are all pressing for increases in the defence triple treatment and single treatment trials. wishing to control her fertility by the use of budget. We hope that they do not find them- For this reason the conclusion that contraception can actually obtain it for her- selves in an overstretched NHS accident and “eradication of H pylori was associated with self or her partner. emergency department late on a winter’s an almost twofold increase in dyspeptic John Guillebaud professor of family planning and night—or do we? symptoms among patients referred to reproductive health Robin Stott chairman specialist clinics” is misleading because it is Department of Obstetrics and Gynaecology, University College London Medical School, Douglas Holdstock honorary secretary based on a summary estimate that makes no London W1P ILB Medact, London N19 4DJ clinical sense and is statistically questionable. j.guillebaud@lineone.net info@medact.org Until a systematic review is carried out with a wider search strategy and more robust statis- 1 Haines A, Heath I, Smith R. Joining together to combat 1 Loefler I. Human population growth: population issue is poverty. BMJ 2000;320:1-2. (1 January.) tical analysis I do not think this meta- not entirely satisfactory. BMJ 2000;320:443. (12 February.) 2 The Hague Appeal for Peace. The Hague agenda for analysis should influence guidelines or 2 Guillebaud J. After Cairo. Br J Obstet Gynaecol 1995;102: peace and justice for the 21st century. www.haguepeace. org/ (accessed 12 April 2000). clinical practice. 436-8. 3 Correspondence. After Cairo. Br J Obstet Gynaecol 1996; 3 Sivard RL. World military and social expenditures 1996. Chris Cates general practitioner 103:91-3. Washington, DC: World Priorities, 1996:48-51. Manor View Practice, Bushey, Hertfordshire 4 Potts M. The population policy pendulum. BMJ 1999;319: 4 Yamey G. Study shows growing inequalities in health in Britain. BMJ 1999;319:1453. (4 December.) WD2 2NN 933-4. (9 October.) chriscates@email.msn.com 5 Haines A, Heath I, Smith R. Joining together to combat poverty. BMJ 2000;320:1-2. (1 January.) Competing interests: None declared. Will eradication of Helicobacter 1 Jaakkimainen RL, Boyle E, Tudiver F. Is Helicobacter pylori associated with non-ulcer dyspepsia and will eradication Eradicating war is essential to pylori improve symptoms of improve symptoms? A meta-analysis. BMJ 1999;319: eliminate poverty and improve non-ulcer dyspepsia? 1040-4. (16 October.) health Studies included in meta-analysis had This meta-analysis is potentially Editor—We welcome the editorial on the heterogenous, not homogenous, results misleading International Poverty and Health Network Editor—Jaakkimainen et al’s meta-analysis Editor—Jaakkimainen et al carried out a and will encourage Medact to participate in concludes that an improvement in dyspeptic meta-analysis to determine whether eradica- the work of the network.1 symptoms occurred among patients with tion of Helicobacter pylori will improve symp- Medact is an organisation of health pro- non-ulcer dyspepsia in whom Helicobacter toms associated with non-ulcer dyspepsia.1 fessionals who are concerned about major pylori was eradicated.1 Unfortunately, there is Their conclusions differ from those that we threats to health, such as violent conflict, a small but crucial problem at the heart of have reached in a systematic review address- poverty, and environmental degradation. To the analysis. The authors report that the ing the same question; we carried out our the editorial’s otherwise comprehensive list summary estimates are statistically homog- review for the United Kingdom health tech- of objectives we would like to add one which enous, but this is incorrect. In the observa- nology assessment programme. We are con- we regard as vital: the elimination of war. tional studies the P value of < 0.001 cerned that the authors’ paper may provide 1208 BMJ VOLUME 320 29 APRIL 2000 bmj.com  Downloaded from bmj.com on 13 February 2006 Letters a misleading impression of the effect of H 1 Jaakkimainen RL, Boyle E, Tudiver F. Is Helicobacter pylori unknown reasons.4 An additional negative associated with non-ulcer dyspepsia and will eradication pylori eradication treatment on symptoms of improve symptoms? A meta-analysis. BMJ 1999;319: study has recently been published.5 non-ulcer dyspepsia. Our review was con- 1040-4. (16 October.) We have conducted a meta-analysis 2 Baron JH, Barr J, Batten J, Sidebotham R, Spencer J. Acid, ducted using a protocol peer reviewed by pepsin, and mucus secretion in patients with gastric and using the four studies that should be the Cochrane Collaboration and will be sub- duodenal ulcer before and after colloidal bismuth included in an up to date review (figure). We mitted to the Cochrane Library. subcitrate (DeNol). Gut 1986;27:486-90. did not find significant symptomatic 3 McColl K, Murray L, El-Omar E, Dickson A, El-Nujumi A, We have identified several problems with Wirz A, et al. Symptomatic benefit from eradicating improvement in the group assigned to Jaakkimainen et al’s meta-analysis. Firstly, Helicobacter pylori infection in patients with nonulcer receive eradication treatment compared dyspepsia. N Engl J Med 1998;339:1869-74. the search strategy is substantially incom- 4 Fischer M. Why being a reviewer in the Cochrane Collabo- with the control group.3–5 We are convinced, plete, with only one electronic database ration is a better deal than working outside it! [electronic therefore, that both meta-analyses presented response to Jaakkimainen et al. Is Helicobacter pylori being searched, no text words used, no “grey associated with non-ulcer dyspepsia and will eradication by Jaakkimainen et al are flawed and should literature” included (this literature is impor- improve symptoms? A meta-analysis.] bmj.com 1999; 319 be disregarded when doctors are deciding tant for obtaining papers in press in a fast (www.bmj.com/cgi/eletters/319/7216/1040#EL4). whether to treat H pylori infection in patients moving field), and non-English language with non-ulcer dyspepsia. papers excluded. Meta-analysis included unreliable studies D Pantoflickova postdoctoral fellow, University of Secondly, we believe that there is a Editor—In their two meta-analyses Jaakki- Lausanne and Geneva potential problem with the selection of trials. mainen et al concluded, firstly, that “people A L Blum head, division of gastroenterology alblum@hola.hospvd.ch Single treatment trials will not have eradi- infected with Helicobacter pylori are about Centre Hospitalier Universitaire Vaudois, cated H pylori adequately, and confounding one and a half to twice as likely to have non- CH-1011 Lausanne, Switzerland may have arisen where the treatment has ulcer dyspepsia compared to controls” and, N J Talley head, division of gastroenterology benefits on dyspepsia other than H pylori secondly, that “eradicating H pylori results in Department of Medicine, University of Sydney, eradication (for example, bismuth and an almost twofold improvement in dyspep- Clinical Sciences Building, Nepan Hospital, erythromycin).2 PO Box 63, Penrith, NSW 2751, Australia tic symptoms.”1 These conclusions are based Thirdly, we are concerned by the on flawed analyses. H R Koelz head Division of Gastroenterology, Triemli Hospital, exclusion of the trial by McColl et al3 and In addition to the critique given in CH-8063 Zurich, Switzerland wonder whether this was because the trial several responses about the paper in did not exclude all patients with reflux-like bmj.com we would like to raise the following Competing interests: None declared. symptoms. A list of excluded studies, and points. In the first meta-analysis, of 23 epide- reasons for exclusions, should be available in miological studies, Jaakkimainen et al exam- 1 Jaakkimainen RL, Boyle E, Tudiver F. Is Helicobacter pylori bmj.com. ined the association between H pylori associated with non-ulcer dyspepsia and will eradication Fourthly, as discussed by Cates in his infection and non-ulcer dyspepsia. Unfortu- improve symptoms? A meta-analysis. BMJ 1999;319: 1040-4. (16 October.) response in bmj.com [published above], the nately, they failed to eliminate unreliable 2 Blum AL, Talley NJ, O’Morain C, van Zanten SV, Labenz J, misinterpretation of odds ratios as effect studies (for example, those in which organic Stolte M, et al. Lack of effect of treating Helicobacter pylori sizes, the handling of tests for heterogeneity, infection in patients with nonulcer dyspepsia. N Engl J Med causes of dyspepsia were not excluded by 1998;339:1875-81. and the application of a fixed effects model endoscopy (eight studies), those in which the 3 Talley NJ, Janssens J, Lauritsen K, Racz I, Bolling- are all potential flaws in the analysis. Even on symptom profile of non-ulcer dyspepsia was Sternevald E on behalf of the Optimal Regimen Cures Helicobacter Induced Dyspepsia (ORCHID) Study Group. the basis of the studies presented here it is not defined (five), and those in which the Eradication of Helicobacter pylori in functional dyspepsia: simply not true to say that “eradication of H patient and control populations were not randomised double blind placebo controlled trial with 12 months’ follow up. BMJ 1999;318:833-7. pylori is associated with an almost twofold matched for age (12). Lack of age matching 4 McColl K, Murray L, El-Omar E, Dickson A, El-Nujumi A, improvement in dyspeptic symptoms.” is particularly important because of the Wirz A, et al. Symptomatic benefit from eradicating Fifthly, the quantitative estimate that is Helicobacter pylori infection in patients with nonulcer direct relation between advancing age and dyspepsia. N Engl J Med 1998;339:1869-74. required by clinicians and researchers plan- prevalence of H pylori infection. 5 Talley NJ, Vakil N, Ballard ED 2nd, Fennerty MB. Absence ning trials is a measure of the likely absolute In the second meta-analysis, of five con- of benefit of eradicating Helicobacter pylori in patients with nonulcer dyspepsia. N Engl J Med 1999;341:1106-11. benefit (or number needed to treat) of trolled clinical trials, the authors examined H pylori eradication in non-ulcer dyspepsia. the effect of treatment of H pylori infection Nowhere in this paper are any figures on dyspeptic symptoms. Here, several More studies should have been included quoted that could be used to determine criteria should have been applied for the in meta-analysis what this might be. selection of the studies: an appropriate defi- Editor—Studies such as Jaakkimainen et The benefit from H pylori eradication nition of non-ulcer dyspepsia; careful blind- al’s meta-analysis on Helicobacter pylori infec- may be modest, and any application in clini- ing; validated outcome measures of cure of tion and non ulcer dyspepsia1 will help to cal practice would require careful considera- the infection and relief of symptoms; clarify currently controversial issues. Unfor- tion and a supportive cost effectiveness adequate follow up of at least six months; tunately, the authors’ conclusions regarding analysis in comparison with alternative and calculation of a study sample size that is the efficacy of H pylori eradication for treatments. We would like to extend Fischer’s adequate to detect the predefined therapeu- non-ulcer dyspepsia may be questionable as comments in bmj.com4 and emphasise how tic gain. Only two of the five studies fulfil only five treatment studies were included–a contact with the relevant Cochrane review these criteria.2 3 A high quality study, source of bias recognised by the authors in group is helpful in ensuring quality in all published in 1998, was not included for their discussion. phases of systematic reviews. Several key studies were omitted despite Brendan Delaney senior lecturer apparently meeting the inclusion criteria— Department of Primary Care and General Practice, Blum et al2 namely, being randomised control trials with University of Birmingham, Birmingham B15 2TT b.c.delaney@bham.ac.uk Talley et al3 accepted definitions of dyspepsia and Paul Moayyedi MRC training fellow in health services non-ulcer dyspepsia, using accepted and McColl et al4 research effective eradication regimens, and having Shelly Soo clinical research fellow Talley et al5 symptoms of dyspepsia as a defined Centre for Digestive Diseases, University of Leeds, Leeds LS2 9JT outcome measure. Summary odds ratio (95% CI 1.01 to 1.72) McColl et al randomised 160 patients to Jon Deeks director Systematic Review Programme, ICRF/NHS Centre 0.1 1 10 omeprazole, amoxycillin, and metronidazole for Statistics in Medicine, Oxford OX3 7LF Odds ratios and summary odds ratio for proportion and 158 to placebo. The authors used a vali- David Forman coordinating editor of patients with complete or almost complete relief dated Glasgow dyspepsia severity score and Cochrane Upper GI and Pancreatic Diseases of dyspeptic symptoms 6-12 months after treatment at one year found a significant benefit in Review Group, Leeds LS2 9LN in eradication trials resolution of symptoms for those who had Competing interests: None declared. become H pylori negative (21% v 7% for BMJ VOLUME 320 29 APRIL 2000 bmj.com 1209 Letters Downloaded from bmj.com on 13 February 2006 those who remained infected).2 Similarly, an estimate,1 the appropriateness of its use has Use of unlabelled and off earlier study by Gilvarry et al reported a sig- been debated.2 nificant reduction in symptoms in patients Inclusion of more studies will indeed licence drugs in children successfully treated with bismuth, tetracy- produce a more stable estimate. What is not cline, and metronidazole compared with obvious in our paper is that we reviewed Use of unlicensed drugs may be bismuth and placebo (symptom score 14.2 more studies than are referenced, including recommended in guidelines and 9.2 at inclusion and at one year follow McColl et al’s trial.3 We calculated a similar Editor—Conroy et al report the wide- up respectively).3 estimate (not published) to that of Pantof- spread use of drugs that are either not A contradictory, and equally valid, study lickova et al when we compared improve- licensed for use in children or are by Talley et al was also not included for ment in symptoms in groups receiving prescribed outside the terms of their analysis. In that study, 278 subjects were ran- eradication treatment with that in a control product licence (off label prescribing) in domised to triple treatment that included a group. Our pooled estimate compared children admitted to hospital. 1 Although it proton pump inhibitor or to placebo; symp- groups in which Helicobacter pylori had and is not illegal to use medicines in this way, the tom scores at one year did not differ between had not been eradicated. For this reason, the responsibility for any adverse events the groups, but an improvement in symp- inclusion criteria for our paper were limited becomes the clinician’s or the pharmacist’s toms with resolution of chronic gastritis was to studies that provided data allowing calcu- rather than the manufacturer’s. However, reported.4 lation of an odds ratio in relation to eradica- much unlicensed use may be recommended Although these studies are not perfect tion, not just treatment. in local or national guidelines. with regard to assessment of compliance, The literature search was conducted As part of our trust’s response to the use description of the randomisation process, through March 1999 and did not include of unlicensed drugs in children, I reviewed and even presentation (as referred to in the studies published after this. The search all drugs recommended in our local meta-analysis), their design is adequate and began with quite broad criteria for including paediatric medical guidelines. These con- their findings significant. Their lack of inclu- studies, including observational studies and tained 69 guidelines for acute management sion in the meta-analysis could affect its non-randomised trials. We did not include and elective investigation of children. The findings and conclusions. The jury in the the qualitative review of all studies in our guidelines recommended 86 drugs, but only trial of H pylori infection is still out; the paper, or all the summary estimates. We 47 (55%) were licensed for use in children. evidence put forward should include all rel- mentioned in our discussion the point that A further 14 drugs were licensed only for evant information. Delaney et al make about our search children above a certain age or weight, Deirdre McNamara specialist registrar 24 were unlicensed or off label, and the sta- M Buckley consultant gastroenterologist strategy. Studies obtained from the “grey lit- C O’Morain consultant gastroenterologist erature” rely on the cooperation of editors. tus of one drug (methylcellulose) was Department of Gastroenterology, Adelaide and This may introduce another selection bias. unknown. Five drugs used for investigations Meath Hospitals, Tallaght, Dublin 24 Changing clinical practice for a com- were not licensed or the licence was Annemarie.Murphy@AMNCH.ie restricted. National guidelines also recom- mon condition with multiple therapeutic Competing interests: None declared. strategies requires a broad look at the litera- mend drugs that are unlicensed for use in ture and a full understanding of the children. 1 Jaakkimainen LR, Boyle E, Tudiver F. Is Helicobacter pylori consequences of treatment. For non-ulcer The British Thoracic Society guidelines associated with non-ulcer dyspepsia and will eradication improve symptoms? A meta-analysis. BMJ 1999;319: dyspepsia, this includes fully examining acid for treating tuberculosis recommend that 1040-4. (16 October.) pyrazinamide is given routinely, although 2 McColl KE, Murray L, El-Omar E, Dickson A, El-Nujumi A, suppression treatment, motility agents, and this drug is not licensed for use in children.2 Wirz A, et al. Symptomatic benefit from eradicating lifestyle changes along with eradication of H Helicobacter pylori infection in patients with nonulcer Primaquine is recommended by national dyspepsia. N Engl J Med 1998;339:1869-74. pylori. We do not know why Delaney et al guidelines for use in vivax malaria, although 3 Gilvarry J, Buckley MJ, Beattie S, Hamilton H, O’Morain would prefer presenting number needed to CA. Eradication of Helicobacter pylori infection affects it is unlicensed.3 treat as the pooled measure. We hesitate to symptoms in non-ulcer dyspepsia. Scand J Gastroenterol Paediatric guidelines (both local and 1997;32:535-40. provide this number, given concerns about 4 Talley NJ, Janssens J, Lauritsen K, Rácz I, Bolling- national) need to acknowledge the licensed pooling studies with variations in the Sternevald E on behalf of the Optimal Regimen Cures status of the drugs they recommend. Helicobacter Induced Dyspepsia (ORCHID) Study Group. background level of risk related to different Linking guidelines to the Royal College of Eradication of Helicobacter pylori in functional dyspepsia: entry criteria and clinical settings.4 randomised double blind placebo controlled trial with 12 Paediatrics and Child Health’s formulary months’ follow up. BMJ 1999;318:833-7. Liisa Jaakkimainen assistant professor (Medicines for Children4) might facilitate this. Department of Family and Community Medicine, University of Toronto, Sunnybrook and Women’s F Andrew I Riordan consultant paediatrician Authors’ reply Department of Child Health, Birmingham College Health Sciences Centre, Toronto, Ontario, Editor—As two letters point out here, Canada Heartlands Hospital, Birmingham B9 5SS a.riordan@kippers-korner.demon.co.uk heterogeneity exists with the summary liisa.jaakkimainen@ices.on.ca estimate of the association studies in our Eleanor Boyle doctorate candidate 1 Conroy S, Choonara I, Impicciatore P, Mohn A, Arnell H, meta-analysis. For this reason we undertook Inner City Health Research, St Michael’s Hospital, Rane AR, et al. Survey of unlicensed and off label drug use sensitivity analyses to explain the robustness Toronto in paediatric wards in European countries. BMJ Fred Tudiver director 2000;320:79-82. (8 January.) of the estimate and the reasons for statistical 2 Joint Tuberculosis Committee of the British Thoracic heterogeneity. Heterogeneity improved in Department of Family Medicine, Center for Society. Chemotherapy and management of tuberculosis Evidence-Based Practice, State University of in the United Kingdom: recommendations 1998. Thorax the subgroup analyses (study design, control New York Health Science Center at Syracuse, 1998;53:536-48. of confounders, and study quality), and the Syracuse, NY, USA 3 Benign malarias (treatment). British National Formulary 38, summary estimate remained stable. Pantof- September 1998:283-4. 4 Royal College of Paediatrics and Child Health. Medicines lickova et al think that studies should be Competing interests: None declared. for children. London: RCPCH Publishing, 1999. eliminated if dyspepsia is not defined or organic causes not excluded by endoscopy. 1 Berlin JA, Laird NM, Sacks HS, Chalmers TC. A compari- When we pooled 11 studies meeting these son of statistical methods for combining event rates from A European paediatric rule is needed to criteria the summary odds ratio was 2.0 clinical trials. Stat Med 1989;8:141-51. protect children 2 Petitti DB. Meta-analysis, decision analysis and cost-effectiveness (95% confidence interval 1.6 to 2.4). analysis. Methods for quantitative synthesis in medicine. Oxford: Editor—Drugs for the paediatric rheumatic The summary estimate for the five eradi- Oxford University Press, 1994:90-114. diseases are now used in new dosages, 3 McColl K, Murray L, El-Omar E, Dickson A, El-Nujumi A, cation trials is not generalisable. We base this Wirz A, et al. Symptomatic benefit from eradicating through new routes of administration, or in argument mainly on the lack of robustness Helicobacter pylori in patients with nonulcer dyspepsia. new combinations, but data on safety and of the estimate when we performed the sen- N Engl J Med 1998;339:1869-74. efficacy tend to be from small, uncontrolled 4 Smeeth L, Haines A, Ebrahim S. Numbers needed to treat sitivity analysis. Even though a random effect derived from meta-analyses—sometimes informative, usu- series. Moreover, most if not all of the drugs model may produce a more conservative ally misleading. BMJ 1999;318:1548-51. for the paediatric rheumatic diseases are 1210 BMJ VOLUME 320 29 APRIL 2000 bmj.com  Downloaded from bmj.com on 13 February 2006 Letters prescribed outside the terms of their 1 Conroy S, Choonara I, Impicciatore P, Mohn A, Arnell H, dane but deadly everyday problems such as Rane AR, et al. Survey of unlicensed and off label drug use product licence (used off label) in most in paediatric wards in European countries. BMJ chronic and infectious diseases and environ- European countries, as described by Conroy 2000;320:79-82. (8 January.) mental insults. 2 European Medicines Evaluation Agency. Report on the et al.1 experts round table on the difficulties related to the use of new In short, the proponents of antibioter- Childhood chronic illnesses with high medicinal products in children held on 18 December 1997. rorism programmes have it backwards. London: EMEA, 1998. (27164/98 Rev 1.) Instead of pumping more resources into ill morbidity should be the target of intense 3 European Medicines Evaluation Agency. Note for guidance research aimed at ameliorating or curing the on clinical investigation of medicinal products in children. Lon- advised and risky antibioterrorism pro- disease. Yet securing funding for trials in don: EMEA, 1997. grammes, we should build national and 4 Connor JD. A look at the future of pediatric therapeutics: children is difficult: the pharmaceutical an investigator’s perspective of the new pediatric rule. Pedi- international public health systems that can industry has little interest in funding these atrics 1999;104:610-3. adequately reduce, detect, and respond to 5 Food and Drug Administration. Regulations requiring trials, the potential market is small, the manufacturers to assess the safety and effectiveness of new drugs natural disease outbreaks and industrial enrolment period is long, and most child- and biologic products in pediatric patients (21 CFR Parts 201, chemical spills. Then, in the unlikely event of 312, 314, and 601). Washington, DC: Federal Register, a bioterrorist attack, these systems will be hood chronic conditions are rare illnesses. 1998:63. To address these problems, in 1996 a available to manage the challenge. European network called the Paediatric Hillel W Cohen assistant professor hicohen@aecom.yu.edu Rheumatology International Trials Organis- Victor W Sidel distinguished university professor ation was founded to facilitate and coordi- Prescriptions on bioterrorism Department of Epidemiology and Social Medicine, nate the development, conduct, and report- have it backwards Albert Einstein College of Medicine and ing of clinical trials with or without the Montefiore Medical Center, New York 10461, USA support of pharmaceutical companies. Editor—Rosen makes several factual mis- Robert M Gould chairperson Thirty two countries now belong to the statements in his editorial on bioterrorism.1 Peace Committee, Physicians for Social For example, the vaccine for anthrax has Responsibility, San Francisco, CA 94704, USA organisation. In 1997 the organisation obtained a been shown to be effective only for the cuta- 1 Rosen P. Coping with bioterrorism is difficult, but may three year grant from the European Union neous form of anthrax and not for the inha- help us respond to new epidemics. BMJ 2000;320:71-2. for a randomised controlled clinical trial of lation form used in weapons.2 He states that (8 January.) anthrax vaccine is not being produced, but 2 Sidel VW, Nass M, Ensign T. The anthrax dilemma. Med medium and high dose parenteral meth- and Global Survival 1998;5:97-104. otrexate in children with juvenile chronic the Pentagon has embarked on a massive 3 Myers LM. Criticizing Pentagon, panel calls for suspension effort to produce this vaccine and to of military’s anthrax shots. New York Times 2000 Feb 18: arthritis that does not respond to standard A14. doses. This trial is built on the current stand- inoculate all US troops on active duty. This 4 Clinton WJ. Remarks by the President on keeping America ard of care where the costs of insurance cov- programme is now under sharp attack in secure for the 21st century. www.whitehouse.gov/WH/ New/html/19990122-7214.html (accessed 18 Aug 1999). erage, medications, clinic visits, and labora- Congress for possible adverse effects and 5 Cohen HW, Gould RM, Sidel VW. Bioterrorism initiatives: tory tests are paid by the usual method of lack of scientific justification.3 The statement public health in reverse? Am J Public Health 1999;89: regarding bioterrorism that “. . . it is hard to 1629-31. medical reimbursement—that is, through a raise money to defend against a problem national health system or medical insurance that has such a low incidence”1 seems system. Interestingly, approval by an ethics reasonable, but is just not true. Last year, the committee has been denied in four coun- Pathogen causing infection United States allocated $10bn (£6.25bn) for tries because of lack of insurance coverage, because of lack of support by a pharmaceu- anti-terrorism, with a substantial portion for related to body piercing should bioterrorism.4 In contrast, food borne be determined tical company, and because drugs were not diseases, which have high morbidity and given free of charge. mortality, have received far less attention Editor—We accept Ferguson’s contention The problems faced by the organisation, and fewer resources. that wearing body jewellery is essentially a and surely by other paediatric groups, in Rosen does, however, make some personal experience,1 but such jewellery dealing with pharmaceutical companies and important observations. Bioterrorist attacks does have clinical implications.2 Maybe ethics committees underline the difficulties have indeed been rare. The incident in 1984 Ferguson would not be astonished at having of designing and conducting clinical trials in Oregon with no fatalities and the two been asked to remove his ornaments if he for children, especially if the issue is to seek chemical attacks in Japan in 1994-5 men- appreciated that x rays cannot penetrate approval from regulatory agencies. tioned in the editorial are the only ones steel and that the ornaments may have been As stated by Conroy et al,1 the European documented. creating diagnostic difficulties. This would be Medicines Evaluation Agency2 issued guid- We agree with Rosen that: “It has also especially true if computed tomography was ance on the clinical investigation of medicinal become apparent that the management of being performed. Requests to remove body products in children that simply encourages any biological attack must depend on jewellery may be clinically necessary and companies to investigate drugs in children systems already in place for managing new should be seen as part of the process of sav- when clinically appropriate.3 The same diseases or new epidemics of old diseases. ing a life. difficulties led the Food and Drug Adminis- Unfortunately, US public health surveil- In this era of evidence based medicine, tration to issue the “paediatric rule,”4 5 by lance systems are not modern, and there with doctors being urged to reduce use of which manufacturers of products likely to be has been little thought about how an antibiotics, we were surprised to read the used in children have to test those products in epidemic might be recognised quickly. Most author’s recommendation of the blind the relevant paediatric population. state public health departments are under- treatment of infections related to body Like Conroy et al, we urge the European funded and do not have the staff to investi- piercing. Ferguson recommends the use of Union and the European Medicines Evalua- gate anything more than a recognised flucloxacillin on the grounds that these tion Agency to issue a similar paediatric rule epidemic.”1 infections are likely to be caused by Staphy- for the European Community to assure chil- Sadly, Rosen’s prescription to encourage lococcus aureus, without any supporting dren and their families the same rights as antibioterrorism programmes is likely to evidence. The microbiology of infections adults to receive drugs that have been fully make this deplorable situation worse. Some related to body piercing has not been tested. programmes, such as secret research spon- determined, although the bacteria associ- sored by the military that could trigger a ated with surgical sutures and staples have Nicolino Ruperto first level medical director nruperto@smatteo.pv.it new arms race in chemical and biological been. These are a mixed flora, with S epider- warfare agents, are inherently dangerous.5 midis the commonest organism; the bacteria Alberto Martini professor of paediatrics For the Paediatric Rheumatology International Diverting resources and attention to the are enveloped in a complex biofilm, Trials Organisation, Paediatric Clinic, IRCCS S “unusual and infrequent event”1 of bioter- which is thought to protect the organism Matteo, University of Pavia, 27100 Pavia, Italy rorism increases vulnerability to the mun- from host defence factors and to account BMJ VOLUME 320 29 APRIL 2000 bmj.com 1211 Letters Downloaded from bmj.com on 13 February 2006 for their persistence on suture surfaces with the cost of leaflets. In our own study of by the government. Its crucial role in ensur- until they are removed with the sutures.3 patients with cancer, 3 maintaining a compu- ing quality training is now threatened. Genital piercing is more likely to be ter system (and replacing it after four years) In 1997 the committee’s outgoing chair- infected with bacteria from the periurethral was cheaper than giving full access to the man, Denis Pereira Gray, opened discussions microflora. Escherichia coli is the commonest expensive cancer booklets (typically £3.00 with the Departments of Health over remu- bacterium to cause urinary tract infections,4 each). neration for the three officers. A proper rate but it cannot be assumed to remain the Although Graham et al found no differ- was set for these posts, and the minister of infecting pathogen in the presence of a for- ence in knowledge, they did note a health gave support in principle. Meanwhile, eign body. Intraoral piercings are likely to be reduction in anxiety in the intervention ad hoc arrangements ensure that some infected with oral commensals. group. Is the possible reduction in anxiety financial recompense is available to the We believe that body jewellery associ- worth 26 pence per woman? The marginal officers from other, inappropriate sources. ated with infection should be removed and benefits are attenuated by the fact that, as The activities of the Specialist Training an antibiotic decided on after signs of local Graham et al and Wyatt point out, this Authority and General Medical Council are and systemic dissemination have been group in the Aberdeen population had a supported by charges on doctors. The joint examined. good baseline knowledge of prenatal tests, committee has no power or inclination to Rakesh Khanna specialist registrar in accident and so only minor improvements in knowledge charge for its certificates. emergency may be achievable in that context. More For many years the government has Manor Hospital, Walsall WS2 9PS supported standard setting by medical royal than half of their sample (55%) came from Rakesh@ChandKhanna.freeserve.co.uk affluent areas. Evaluation of the system in an colleges in the form of a grant in aid. In S Sathish Kumar house officer in cardiology 1975 the Royal College of General Practi- Tameside General Hospital, Ashton-under-Lyne area of high deprivation, where non-print OL6 9RW media may have benefits over the written tioners decided to transfer all general prac- word among a population with lower tice grant in aid to the joint committee and literacy and knowledge levels, may show dif- to supplement it with a matching sum. The 1 Ferguson H. Body piercing. BMJ 1999;319:1627-9. (18-25 ferent outcomes. grant in aid for general practice compares December.) 2 Khanna R, Kumar SS, Raju BS, Kumar AV. Body piercing Further evaluation, including economic unfavourably with that for other disciplines. in the accident and emergency department. J Accid Emerg analysis, of the use of the system in an area In 1995 (latest available figures) standard Med 1999;16:418-21. setting costs for the United Kingdom’s 3 Gristina AG, Price JL, Hobgood CD, Webb LX, Costerton with higher levels of deprivation is worth- JW. Bacterial colonization of percutaneous sutures. Surgery while and would also show the feasibility of 33 000 general practitioners were sup- 1985;98:12-9. “technology transfer” for such systems. ported by a grant in aid of £58 725—in stark 4 Faro S, Fenner DE. Urinary tract infections. Clin Obstet Gynecol 1998;41:744-54. However, the simple costing exercise in this contrast to the £48 649 and £38 532 for letter may show that Wyatt is wrong and that 2900 career grade pathologists and 750 the system should continue in routine use in ophthalmologists respectively. Aberdeen. A 1998 government review of grant in Patient information systems are aid praised the committee’s efficiency and Ray Jones senior lecturer in health informatics frugality.1 Its financial problems now not more expensive than r.b.jones@udcf.gla.ac.uk threaten its core regulatory functions. Neil Craig lecturer in health economics leaflets Department of Public Health, University of Despite the voluntary return in 1998 of Glasgow, Glasgow G12 8RZ unused funds because of an unfilled post, Editor—Graham et al, who compared a the Department of Health requires a reduc- touch screen system with leaflets for provid- tion of 7.5% in budgets for the year 2000-1 ing women with information on prenatal 1 Graham W, Smith P, Kamal A, Fitzmaurice A, Smith N, (or over 10% if pension contributions are tests,1 and Wyatt in his commentary2 may be Hamilton N. Randomised controlled trial comparing effectiveness of touch screen system with leaflet for provid- taken into account). Consequently, with too conservative in their conclusions. ing women with information on prenatal tests. BMJ cramped premises, barely sufficient staffing, Graham et al mention the cost of develop- 2000;320:155-59. (15 January.) and increasing statutory responsibilities, the ing the system and Wyatt says because there 2 Wyatt J. Commentary: Evaluating electronic consumer health material. BMJ 2000;320:159-60. (15 January.) committee is faced with reducing its is limited evidence of benefit for these 3 Jones R, Pearson J, McGregor S, Cawsey AJ, Barrett A, services.This government demands quality expensive tools over well designed leaflets Craig N, et al. Randomised trial of personalised computer based information for cancer patients. BMJ in the NHS and looks to its competent they should be used only in the context of 1999;319:1241-7. authorities to mediate and promote it. rigorous research studies. However, the Ironically, this very quality is now endan- study did not include an economic com- gered by the government’s shortsighted parison of the use of the computer and leaf- parsimony. lets. A full economic analysis would compare the estimated development costs Joint committee on Brian Keighley chairman, Joint Committee on Postgraduate Training for General Practice of £25 000 and the subsequent mainte- postgraduate training for GPs John Chisholm chairman, General Practitioners nance costs with the costs of the alternative. needs more money for its work Committee, BMA It would then compare the costs with the Mike Pringle chairman of council, Royal College of marginal benefits during the period that the Editor—Since 1976 the Joint Committee General Practitioners on Postgraduate Training for General Joint Committee on Postgraduate Training for system is in use. General Practice, London SW7 1PU If the system is readily transferable to Practice has presided over the standards of another site, development costs at that vocational training in general practice in the 1 NHS Executive. Review of the central funding of the medical other site would not be anywhere near United Kingdom. Despite the committee royal colleges in respect of training related activities and clinical £25 000. Given the high volume of use that having a relatively low profile compared audit. Leeds: NHSE, 1998. could be achieved in antenatal care, the cost with the two other competent authorities in per patient could be small. For example, use the United Kingdom (the General Medical of a computer with touch screen and a Council and Specialist Training Authority) printer costing £2000 capital with 5% its influence extends well beyond general maintenance in years 2-4 might have a four practice. It is responsible for approving year cost (without discounting) of £2600. training posts for intending general practi- Aberdeen Maternity Hospital had 4734 tioners in hospitals and primary care and deliveries in 1997, a mean of 13 a day.1 If, has tried to improve training for senior over four years, 10 000 women (53%) used house officers in all disciplines. the system, the attributable cost per woman Unfortunately, its low profile has been Correspondence submitted electronically is 26 pence. This may compare favourably reflected in the value and support afforded it is available on our website 1212 BMJ VOLUME 320 29 APRIL 2000 bmj.com