Education And Debate

Are research ethics committees behaving unethically? Some suggestions for improving performance and accountability

BMJ 1996; 313 doi: http://dx.doi.org/10.1136/bmj.313.7069.1390 (Published 30 November 1996) Cite this as: BMJ 1996;313:1390
  1. Julian Savulescu, Sir Robert Menzies medical scholara,
  2. Iain Chalmers, director, UK Cochrane Centreb,
  3. Jennifer Blunt, former chair, Salford research ethics committeec
  1. a Clinical Ethics Project, Oxford Radcliffe Hospitals, Oxford OX3 9RP
  2. b NHS Research and Development Programme, Oxford OX2 7LG
  3. c Salford and Trafford Health Authority, Eccles M30 0NJ
  • Accepted 19 September 1996

The results of recent empirical investigations in research synthesis imply that research ethics committees are behaving unethically by endorsing new research which is unnecessary and by acquiescing in biased under-reporting of research which they have approved. The performance and accountability of research ethics committees would be improved if they required those proposing research to present systematic reviews of relevant previous research in support of their applications; to summarise the results of these reviews in the information prepared for potential participants; to register new controlled trials at inception; and to ensure that the results of these trials are made publicly available within a reasonable period of time after completion of data collection.

Properly designed, conducted, and reported research is essential to distinguish useful from useless or harmful forms of health care.1 Research ethics committees exist to ensure, firstly, that proposed research will not expose participants to unacceptable risks and practices; and, secondly, that the potential participants can evaluate the expected consequences of their involvement and decide for themselves whether to participate.2 3 Research ethics committees have a wider responsibility to promote the public interest by helping to ensure that relevant research is done.3 In this article we suggest that the results of recent research in a new and rapidly expanding sphere of empirical investigation—research synthesis4—has important implications for the work of research ethics committees.

Research synthesis is the aggregation and integration of the results of related primary studies with the purpose of drawing conclusions from the totality of the relevant evidence. As the body of primary research evidence expands, research synthesis has become essential; but there is an emerging recognition that the quality of most research syntheses leaves much to be desired.5 6 Research syntheses of poor quality can lead to beneficial effects of treatments being overlooked, as well as to promotion of treatments that are either ineffective or actually harmful.6 7 8 It is against this background that the importance of improving the quality of research synthesis has been recognised, not only within the research community,9 but increasingly by bodies responsible for funding scientific research and accrediting academic institutions.10

Reliable syntheses of the results of primary research require the adoption of methods that will reduce systematic and random errors (see box),11 and empirical investigation to avoid such errors is a rapidly growing field of scientific inquiry.4 In at least two important ways the results of this empirical research are relevant to research ethics committees.

Are research ethics committees satisfied that proposals for research have taken proper account of the results of existing research?

In 1981 Baum and his colleagues reported the results of a systematic review of controlled trials assessing the effects of prophylactic antibiotics on wound infection and mortality after colon surgery.12 Cumulative synthesis of the results of these trials showed that strong evidence of the effectiveness of antibiotic prophylaxis in reducing morbidity and mortality existed by the mid-1970s. However, when Baum and his colleagues published their systematic review five years later researchers were continuing to invite patients to participate in such trials, and reports of trials involving comparison groups given no active treatment continued to appear throughout the 1980s (fig 1).13 Research addressing previously answered questions either denies participants effective treatment or places them at risk for no benefit, or both.14

Sources of bias and methods of protecting against bias11

  • Problem formulation

  • Is the research clearly focused?

  • Study identification

Is the search for relevant studies thorough?

  • Study selection

Are the inclusion criteria appropriate?

  • Appraisal of studies

Is the validity of included studies adequately assessed?

  • Data collection

Is missing information obtained from investigators?

  • Data synthesis

How sensitive are the results to changes in the way the review is done?

  • Interpretation of results

Do the conclusions flow from the evidence that is reviewed?

Are recommendations linked to the strength of the evidence?

Are judgments about preferences (values) explicit?

If there is “no evidence of effect” is caution taken not to interpret this as “evidence of no effect.”

Are subgroup analyses interpreted cautiously?

Fig 1
Fig 1

Reduction of perioperative deaths by antibiotic prophylaxis for colorectal surgery.13

Nearly a decade ago Freedman noted that participants in research must not be denied access to effective care.15 Over the past five years an increasing number of cumulative syntheses of the results of successive clinical trials (sometimes referred to as cumulative meta-analyses)13 16 have called into question decisions made by research ethics committees to endorse proposals for new placebo controlled research when existing evidence shows that an active form of care is better than placebo. Rothman and Michels, for example, have called attention to recent trials of new secondary treatments for rheumatoid arthritis, new antidepressants, new antiemetics, new antihypertensives and new drugs for congestive cardiac failure in which they suggest that the results of previous research made the use of placebos unethical.17

Further research, including placebo controlled trials, may be warranted when all the relevant outcomes of an intervention (including side effects) have not been sufficiently evaluated. The justifications for such further research involve value judgements about the degree of certainty necessary for evidence to be convincing, what constitutes a relevant outcome, the place of research in changing clinical practice, and the entitlements of patients to treatment when resources are limited.

In cases in which further research is deemed to be justified, however, potential participants should still be adequately informed of the existing results of earlier research. In 1992 researchers invited women to participate in a placebo controlled trial of prophylactic antibiotics at the time of caesarean section. After endorsement by the research ethics committee the following information was presented by the investigators to potential participants: “Some preliminary studies suggest that giving an antibiotic at the time of caesarean section may reduce the chances of having an infection”18 (our emphases). In fact, strong evidence had existed since the mid-1970s that prophylactic antibiotics at the time of caesarean section do reduce the rate of serious postoperative infection and this evidence had been presented in systematic reviews of over 60 relevant trials in 1989 and 1991.19 20 Potential participants could and should have been provided with a proper synthesis of this “preliminary” evidence.

At present, US Food and Drug Administration regulations require that trials of new treatments use placebo controls, even when effective treatments exist. The administration has also acknowledged that there are some important deficiencies in the quality of informed consent in clinical trials.21 22 A deficiency which the administration may not have highlighted, however, is that patients are unaware that they are being denied effective forms of care solely to comply with its policies. These coercive requirements are in contravention of international ethical guidelines,2 23 24 and the scientific justification is in many cases dubious.25 26 Many placebo controlled trials could and should be using treated controls.27

Are research ethics committees satisfied that the results of research they have endorsed are being made publicly accessible?

Over the past 10 years evidence has accumulated showing that the results of a significant proportion of controlled trials are never made properly accessible and that research which has yielded disappointing results tends to be under-reported.28 Studies which have yielded relatively dramatic estimates of beneficial effects are more likely to be submitted for publication,28 more likely to be reported in print,28 more likely to be published as full reports,29 more likely to be published in journals that are widely read,30 31 and more likely to be cited in reports of subsequent, related trials.32

The fate of scientific data is no idle issue. Reliable research syntheses depend on ensuring that as much as possible of the potentially relevant evidence is taken into account in systematic reviews. Simes first drew attention to the way that biased under-reporting of research can lead to misleading conclusions about the value of treatments,30 and Egger and Davey Smith have recently suggested that this phenomenon may explain the discrepancy between the results of a review of small, early trials of magnesium administration in the acute treatment of myocardial infarction and the results of a trial involving nearly 60 000 patients which was designed to test the promising hypothesis derived from the review.33

Contrary to a widely held assumption, it is investigators and some research funders—and not necessarily journal editors—who are primarily responsible for under-reporting of research,28 and there is evidence34 that some investigators are unconcerned about this.35 36 Prevention of this form of scientific misconduct must therefore involve the bodies to which investigators are answerable, in particular research ethics committees and research funding organisations, both public and private37 38 39 40 as well as professional bodies.41

The consequences of under-reporting of research are that patients are being expected to accept the harmful side effects of ineffective forms of care; accept advice about the effects of health care which is based on evidence that is less complete than it should be; participate in research which has been conceptualised and designed on the basis of unnecessarily incomplete information; and contribute to research (both as indirect funders and as patients) which may not be published if the results come as a disappointment or an embarrassment to the investigators or sponsors.42

Five tasks for research ethics committees

To protect the interests of participants in research and promote the public interest in general, research ethics committees must play their part in reducing unnecessary, sometimes harmful, research and in ensuring that the results of necessary, well conducted research are made publicly available.43 What might they do to reduce the problems we have discussed?

  1. Require systematic reviews of existing research before approving research

    Research ethics committees should insist that proposals for research to assess the effects of health care be supported by scientifically defensible reviews of the results of relevant existing research.4 Applications for ethical approval should include a section entitled “Systematic review of relevant existing research.” This should show that the proposed research is necessary to address relevant uncertainty about the effects of one or more forms of health care; that it properly incorporates lessons from previous research; and that it would not entail withholding from some participants forms of care which are known to be effective. These steps would complement those which have already been taken by some national drug licensing organisations.27 44

    While we do not propose that research ethics committees should conduct the kind of systematic reviews which we have described, we do suggest that committee members should have the capacity to evaluate systematic reviews submitted in support of applications for research. An appropriate starting place is Oxman's checklist (see box)11 and the growing literature on this topic.4 45 46 These skills are not beyond the reach of lay people, as shown by the encouraging experience with lay participants in training workshops organised by the critical appraisal skills programme (S Oliver, R Milne, unpublished report to the King's Fund Development Centre, November 1995).47

  2. Require that a summary of relevant systematic reviews be made available to potential participants

    Participants in research have a right to know why research is being performed and its potential significance, as well as its potential harm. This requires that the information presented to patients includes a summary of the results of existing research, indicating not only the possible risks but also the possible benefits.

  3. Require registration of clinical trials at inception as a condition of approval

    Research ethics committees should support moves to require prospective registration of controlled trials at inception.38 48 49 Not only would such registration make it easier to identify irregularities in trial design, analysis, reporting, and publication or non-publication, but it should also help to prevent unknowing duplication of efforts within the research community. Within the United Kingdom research ethics committees should require registration of all clinical trials in the national research register.50

  4. Require a commitment to ensure that research results are made publicly accessible as a condition of approval

    Research ethics committees should restrict approval of proposals for research to studies in which the principal investigators confirm in writing that adequate reports of the research will be made publicly accessible within a reasonable period of time (say 12 months) after the completion of data collection. Researchers who will not make this commitment should be required to make it explicit in the information and consent form inviting participation that the results of the research in which participation is invited may not be made publicly available.

    Calls are increasing for the results of research to be made more accessible.51 52 Indeed, the Council for International Organisations of Medical Sciences has stated that researchers have an obligation to inform participants of the results of research when those results have implications for their health.53 Furthermore, some consumer groups are being actively encouraged to seek reports of the research to which their members have contributed.54

  5. Audit the reporting of results of research previously approved

    In several countries acceptance is growing for the idea that research ethics committees have responsibilities that continue after approval of research.39 55 56 57 58 59 60 61 Whatever the arguments for and against some of these proposed responsibilities, we believe that one of these responsibilities should be regular audit of the reporting of the results of research which research ethics committees have approved. Indeed, some research ethics committees do already regard this as their responsibility.62

    Since the resources of research ethics committees are limited57 60 audit of reporting could be selective, using a random sample of approved research projects to assess whether the results of research had been made publicly accessible. If researchers had failed to produce a publicly accessible account of the results of their research ethics committees might consider withholding approval of future research by the investigator(s) concerned until the results of the unreported research had been made available.

Conclusions

Research ethics committees are uniquely important institutions for at least two reasons. Firstly, they are the only regulatory point through which all proposed clinical research is likely to pass. Secondly, unlike other players who influence the research industry, they are unlikely to have strong vested interests in seeing particular results from research.

Summary of recommendations

Research ethics committees should:

  1. Require systematic reviews of existing research before approving research

  2. Require that a summary of relevant systematic reviews be made available to potential participants

  3. Require registration of clinical trials at inception as a condition of approval

  4. Require a commitment by investigators to make the results publicly accessible as a condition of approval

  5. Audit the reporting of results of research previously approved by them

There have been calls now for over 15 years for greater accountability of research ethics committees to justify their decisions,63 64 65 66 67 68 and especially their decisions actively to thwart what has subsequently turned out to be well designed, beneficial research.63 69 In medical ethics debate continues about whether there is a morally relevant difference between acts and omissions, or between what people do and what people allow to happen.70 71 72 73 74

As the most independent bodies regulating the practice of research, we believe that research ethics committees should be held accountable if, in the light of present understanding of the importance and principles of research synthesis, they continue to allow two forms of scientific malpractice to occur: the execution of unnecessary, sometimes harmful, research and the failure to ensure that the results of research are publicly accessible. Although our proposals have some resource and training implications, they are morally required. Indeed, unless research ethics committees can meet these requirements, we find it difficult to understand how their continued existence can be justified.

We are grateful to Doug Altman, Peter Davey, Claire Foster, Peter Gotzsche, David Henry, Suzanne Hill, Les Irwig, Geir Jacobsen, Eric Meslin, Michael Minton, David Moher, John Pearn, Fiona Smaill, Alastair Watson, and an anonymous referee for helpful comments on earlier drafts.

Footnotes

  • Funding JS is supported by the Sir Robert Menzies Foundation and the Oxford Radcliffe Hospital Trust; IC is supported by the NHS Research and Development Programme.

  • Conflict of interest None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.