eMJA: Jamrozik & Kolybaba, Are ethics committees retarding the improvement of health services in Australia?

Ethics

Are ethics committees retarding the improvement of health services in Australia?

Konrad Jamrozik and Marlene Kolybaba

Multicentre studies are now required to go through multiple systems of ethical approval, which adds to costs and delays

MJA 1999; 170: 26-28
For editorial comment, see Loblay and Chalmers

→ Other articles have cited this article

Introduction - Case study - Discussion - Conclusion - Acknowledgements - References - Authors' details
Register to be notified of new articles by e-mail - Current contents list - More articles on Ethics

Introduction

This article uses a case study to show that the system of institutional ethics committees (IECs) that has been created in Australia now stands to impede the identification of suboptimal practice and waste in the healthcare system, and that there is a grave danger that IECs will delay or even prevent identification of avoidable threats to health originating either inside or outside the medical system. It presents the case for simplifying the requirements for ethical review of multicentre investigations that require only access to medical records.

Case study -- review of prostate cancer management

The incidence of cancer of the prostate rose steeply in Australia in the first half of the 1990s, due partly to widespread use of the prostate-specific antigen test for case-finding.¹ The rapid increase in incidence, with little concomitant change in mortality,¹ had major consequences for the many more men who had to live with the diagnosis and for the health services that had to be provided to them.

Recognising this, a multidisciplinary group of investigators is documenting the presentation, investigation and primary management of all cases of prostate cancer diagnosed in Western Australia (WA) in 1992, the year before active case-finding began. Records from the State Cancer Registry, to which notification of cases is mandatory, indicate that 634 new diagnoses of prostate cancer were made in 1992. Linkage to official mortality statistics for WA, for which permission was granted by the Confidentiality of Health Information Committee of the Health Department of Western Australia, revealed that 265 of these men had died by the time that a retrospective review of medical records began in early 1997.

The protocol for the study was approved by both the Confidentiality of Health Information Committee and the Committee for Human Rights of the University of Western Australia, the latter being an IEC operating within the terms of the NHMRC statement on human experimentation² and supplementary note 1 to that statement.³ All but one of the 16 urologists in WA also gave permission for a research nurse to review relevant medical records, to abstract details of each case and to copy pertinent laboratory reports.

Information has now been collected for nearly all relevant cases, but only after the investigators had sought and gained ethical clearance from a further 29 IECs or their representatives. Each IEC was sent an initial letter of application, including a copy of the approved research proposal outlining the background, aims and objectives of and methods to be employed in the project, together with copies of the approval from the Committee for Human Rights and the Confidentiality of Health Information Committee. For two IECs, misplaced correspondence necessitated resubmission of the application. In most instances (62%), the hospital IEC approved the project once it had been provided with evidence that another Committee had reviewed and passed the protocol, but this process still took several weeks. Six IECs required multiple copies of an application in a form unique to their institution before the application would even be considered.

Four IECs initially took the view, despite the investigators having permission from the creators of the records to inspect them, that the written consent of each patient would be required before his record could be inspected. One committee has remained firm in this view, even when, in supplementary correspondence and in face-to-face discussion, it has been pointed out that this provision might introduce substantial selection bias into the project, that it was not clear who might legitimately give permission on behalf of a deceased patient, and that to ask the investigators to identify and trace next-of-kin was not only an unduly onerous request that would make the study logistically impracticable, but could also result in undue anxiety to the bereaved.

The Table shows that it took between two and 11 weeks to obtain rulings on applications to IECs in public hospitals, with the median time from application to receipt of a written response being 6.5 weeks and 75% responding within 8.6 weeks. In the private sector, the median time was five weeks, with 75% responding within 11 weeks. Although one private hospital responded within a week, correspondence with another continued for more than a year after the initial application was submitted before it finally approved access to records with all identifying information removed. This defeats the purpose of the study, as such an arrangement makes it impossible to relate treatment given to outcome of treatment for a particular patient.

Table, as mentioned above

In the event, at least a tenth of the resources committed to the investigation have been used in preparing applications to and dealing with correspondence from IECs.

Discussion

This case study highlights many of the issues mentioned in a recent review of the system of IECs in Australia, chaired by Professor Don Chalmers:⁴ the number of IECs has grown very rapidly in recent years; there is a lack of uniformity in application processes and decisions of different committees; and Australia's present system of ethical overview does not cater well for multicentre investigations. Similar problems have been reported overseas,⁵ as have long delays and high costs in gaining ethical approval for multicentre studies.⁶

The report from the Review Committee identifies two guiding principles for our system of IECs, namely that the prime responsibility of an IEC is to protect the interests of the subject of the research, and that the responsibility for ethical conduct of the research is vested in the investigator and ultimately in the institution to which he or she is affiliated.⁴ It does not consider whether an IEC also has a duty in relation to the collective interests of the community that stands to benefit from the results of research.

Distinguishing between retrospective and prospective studies

The primacy of the interests of an individual participant in a research project seems simple enough, especially when human experimentation is proposed, but, when what is under consideration is only a systematic review of records, how should an IEC balance the likely desire of individuals for confidentiality and privacy against their likely desire for confidence that the system of healthcare they are using, and for which they are paying, is running effectively, efficiently and safely? In our experience, patients see follow-up surveys of their long-term outcome after an admission to hospital as important and they support them strongly. Response rates of 80% or more are readily achieved in such studies, and many participants add complimentary remarks to postal questionnaires to the effect that they perceive such activities as evidence of a commitment to maintaining excellence in the healthcare system. Positive comments are many times more frequent than negative ones.

Retrospective reviews of medical records are even less intrusive than postal surveys and are a cheap, relatively rapid and efficient way of gaining a comprehensive view of the health system's response to a particular medical problem. Clearly, they involve use of medical notes for other than the primary purpose for which they were created, but, equally clearly, systematic reviews of this kind are qualitatively different from other research in which participants are asked to undergo additional or novel tests or treatments.

For example, in an appropriately designed Phase III trial of a new treatment, there is at least an 80% chance that patients receiving one of the two treatments will have a significantly worse outcome than those randomised to receive the other. It is therefore entirely proper that such studies are subjected to rigorous ethical scrutiny, although, as a recent commentator points out,⁷ the requirement that the IEC of every hospital participating in a multicentre trial review the same protocol seems needlessly wasteful of time and energy.

Contrast such a prospective trial with an investigation of the same drug after it has been on the market for some time when the possibility of a rare teratogenic effect is raised. Under the present Australian system of ethical oversight, investigators collecting data to confirm or refute this association would have to submit a detailed application to the IEC of every health institution with which the mothers of affected and unaffected control children had had contact within the year before the child's birth. The example of Debendox shows that, faced with the threat of litigation and unable to wait for independent investigators to complete a cumbersome and time-consuming series of applications to IECs before even beginning their research, the manufacturer of what may be a safe and useful drug may feel that there is no option but to withdraw the product from the market.

In Australia one peer-reviewed case-control study of a possible life-threatening side effect of a drug used in the treatment of asthma was rejected by one IEC, and the investigators estimate that it cost them $20 000 and a year's work to gain approval from 14 other IECs.⁸ Thus, significant amounts of public money from government agencies or charitable organisations are wasted by IECs requiring that generally innocuous retrospective studies go through multiple ethical reviews.

Investigators conducting retrospective studies of all types -- be they of alleged side effects of pharmaceutical agents, the long-term consequences of occupational exposures or military service, or simply descriptions of how particular presenting complaints were investigated and treated by the medical system at a given time -- are acutely aware of the privilege of using existing medical records for their work as well as of the need to observe the highest ethical and scientific standards in their enquiry. Such studies are motivated as much by the public interest as by scientific curiosity, but it is obviously also in the investigators' own interests to ensure that the privileges extended to them are respected and not abused. As supplementary note 6 of the NHMRC statement on human experimentation⁹ makes clear, in considering applications for approval of research of this kind IECs need to find a balance between protecting the interests of individual patients and not jeopardising the validity and hence broader utility of the results of the research.

Need for clear ethical guidelines on retrospective research

The focus of IECs is on safeguarding the interests of subjects associated with single institutions and of the single institutions themselves. Supplementary note 1 of the NHMRC statement on human experimentation³ includes no requirement that the membership of an IEC include someone with training or experience in a population-based discipline, and many IECs appear to struggle with the need for a whole-of-population study to include every single case of interest from a defined community, regardless of where that case was diagnosed or treated. Sometimes this is motivated by a reluctance to make "our data" available to outsiders, but more often it seems related to a failure to grasp issues surrounding selection bias and statistical power in studying infrequent but important events.

In retrospective research, any harm that might have been done has already occurred, and the challenge is to strike a balance between the risks to privacy and confidentiality associated with a review of existing records and the potential benefits to existing patients, future patients and the public in general.¹⁰ If we are not to have regional or whole-of-State ethics committees, IECs should distinguish between retrospective studies based entirely on records, retrospective studies involving contact with patients or their families (and thus collection of new information) and prospective studies, and develop mechanisms for expeditious consideration of at least the first of these. For multicentre studies requiring only access to records, the combination of an established research investigator, clearance from one IEC constituted according to guidelines set down by the NHMRC and permission from the relevant professional specialist body might be the minimum required for approval-in-principle by Chairs of other IECs.

Conclusion

Currently, the healthcare system has embarked upon a quest for effectiveness and efficiency embodied by the "Golden Fleece" of evidence-based medicine. Measuring our progress towards that goal requires systematic review of past and present practice, which entails applying medical records to a purpose other than that for which they were primarily, but not exclusively, created. This immediately raises complex ethical questions, answers to which are rarely absolute. Yet, IECs should not tie down such studies with red tape, nor needlessly impede the interests of past and future patients and of the broader community.

Acknowledgements

The study of prostate cancer mentioned in this report is supported by the Health Department of Western Australia and the Cancer Foundation of Western Australia. Professor D'Arcy Holman kindly provided comments on an early draft of the manuscript.

References

Threlfall T, Whitford M, Thompson J. Cancer incidence and mortality in Western Australia 1992-94. Perth: Health Department of Western Australia, 1996.
National Health and Medical Research Council. Statement on human experimentation. Canberra: NHMRC, 1992.
National Health and Medical Research Council. Statement on human experimentation: supplementary note 1 -- institutional ethics committees. Canberra: NHMRC, 1992.
Review Committee (Chair: Professor D Chalmers). Report of the review of the role and functioning of institutional ethics committees. Canberra: Australian Government Publishing Service, 1996.
White AE. Research ethics committees at work: the experience of one multi-location study. J Med Ethics 1996; 22: 352-355.
Ahmed AH, Nicholson KG. Delays and diversity in the practice of local research ethics committees. J Med Ethics 1996; 22: 263-266.
Beran RG. Should there be an accredited ethics committee system for centralised review of multicentre clinical research? Med J Aust 1998; 168: 174-176.
Smith MA, Jalaludin B, Leeder SR, Smith WT. Isn't one institutional ethics committee's approval enough? Med J Aust 1994; 160: 662.
National Health and Medical Research Council. Statement on human experimentation: supplementary note 6 -- epidemiological research. Canberra: NHMRC, 1992.
Emson HE. Minimal breaches of confidentiality in health care research: a Canadian perspective. J Med Ethics 1994; 20: 165-168.

Authors' details

Konrad Jamrozik, MB BS, DPhil, FAFPHM, Associate Professor in Public Health.
Marlene Kolybaba, BSc, MPH, Research Officer.

Reprints will not be available from the authors.
Correspondence: Associate Professor Konrad Jamrozik, Department of Public Health, University of Western Australia, Nedlands, WA 6907.
Email: konradATdph.uwa.edu.au

Other articles have cited this article: