Ethics review of multisite studies: the difficult case of community-based Indigenous health research

Ethics review of multisite studies involving human participants has long been a challenge for researchers in Australia1-3 and overseas.4,5 During the past decade, a series of reports, primarily from the United States6-13 and United Kingdom,14-16 have documented negative experiences with repeated reviews of the same study by ethics committees at multiple sites. Researchers’ concerns include cost, inefficiency, tardiness, duplication, and the variable and inconsistent nature of ethical concerns identified across committees.

Following the UK’s lead in this area, Australia is moving to address the problem by introducing procedures for a single locus of review for multisite studies that would have authority across all or most sites.17,18 Centralised ethics review is a logical strategy for bringing efficiencies to the oversight of multisite studies. It is relatively easy to envision how this approach will work for some types of research — for example, a clinical trial recruiting patients at tertiary care hospitals in several states. For other types of research, it is less clear. In community-based population health research, for example, moves towards centralisation will force difficult trade-offs that could limit the role of local concerns and community input in the review process.

Multisite Indigenous health research looms as a particularly vexing challenge. Current national ethics guidelines stress the need to scrutinise researchers’ engagement with Indigenous communities; the regulatory goal is to ensure that the design and conduct of the research is sensitive to the interests and concerns of those communities19 (although there remains considerable debate and controversy regarding how best to achieve effective engagement20-23). Centralisation of ethics review of multisite studies will not displace those responsibilities. Hence, the burdens and barriers imposed by ethics review of multisite Indigenous health research could remain substantial.

Our recent experience during the National Indigenous Eye Health Survey (NIEHS) shows this mixture of forces at work. Here, we describe the course of ethics review and community engagement associated with the NIEHS. We also discuss pending reforms in the oversight of multisite studies and, reflecting on our experience, outline several ideas to guide policy reforms aimed at achieving sound ethics oversight of multisite Indigenous health research.

The NIEHS is a multistage randomised cluster study. It was designed to assess the prevalence and principle causes of vision impairment, the utilisation of eye care services, the barriers to eye health, and the impact of vision impairment among Indigenous Australians. The study methods are summarised in Box 1 and described in detail elsewhere.24

To conduct the NIEHS, we sought ethics approval and engaged in community consultation in all Australian states and territories during 2007 and 2008. Written correspondence relating to these activities was reviewed and analysed (Box 1). An overview of the process of ethics review and consultation for the NIEHS is provided in Box 2. A summary of the NIEHS ethics review experience is described in Box 3 and issues raised during the ethics review are categorised in Box 4.

Before the NIEHS could proceed, it was necessary to work through an elaborate process of ethics review and community consultation. This involved correspondence with 73 entities (Box 2). Only eight of the entities were Human Research Ethics Committees (HRECs) with formal approval authority, but those HRECs operated as gateways to 31 separate community organisations whose sign-off was essentially a pre-condition to ethics approval. The 39 entities with approval authority raised a total of 80 discrete ethics-related issues (Box 4).

As the NIEHS is a relatively large and fully funded project, it was possible to channel staff effort and project resources to manage the ethics review and consultation requirements. We did not formally track that level of effort, and quantifying it retrospectively is unscientific, but we estimate that over 22 months the effort absorbed 50% of one full-time staff member’s time and 15% of each of the three chief investigators’ time. Researchers pursuing smaller multisite studies, and those without full funding support, are likely to struggle to negotiate an ethics review process on this scale.

Centralised ethics review is a widely discussed policy reform for improving the efficiency of multisite research. The UK has moved aggressively in this direction.25 Specially constituted committees within National Health Service trusts are empowered to confer approvals that operate throughout the UK. Although this reform does not appear to have eliminated the workload associated with gaining approval from local ethics committees, early reports suggest that it has streamlined overall approval processes markedly.26-28

Similar reforms are pending in Australia. Under direction from the Australian Health Ministers’ Advisory Council, the National Health and Medical Research Council (NHMRC) established the Harmonisation of Multi-centre Ethical Review (HoMER) project during early 2007. This project involves development of a nationally harmonised system in which a single ethics and scientific review by an approved committee will cover multiple institutions, within and across states and territories.18 New South Wales and Queensland already operate intrastate versions of such a system, and Victoria is currently developing one.

It is unclear how centralised ethics review will operate in the context of community-based studies, like the NIEHS, in which there are many local stakeholders. Centralisation will not change the desire of those stakeholders to have a say in how research is done in their communities, nor should it. Community organisations bring an important perspective — which an expert group focused on scientific issues and the requirements set forth in the National statement on ethical conduct in human research29 will not have — and the ethics review process is deficient without this perspective.30 In addition, effective community engagement enables effective implementation of research findings.

On the other hand, many layers of ethics approval and consultation almost certainly deter or derail large-scale investigation and surveillance in areas of national importance, such as Indigenous health.31 If our experience with the NIEHS is a reliable guide, the administrative and logistical burdens of such ethics review are substantial. From a population health standpoint, this is an unfortunate outcome. Efforts to “close the gap” in the health of Aboriginal and Torres Strait Islander peoples depend on rigorous clinical and epidemiological research, some of it at the national level. Large-scale studies, arguably more than their small-scale counterparts, have the potential to produce the paradigm shifts in knowledge, public opinion and political will that are needed to make major strides towards improving Indigenous health.

How can this tension be addressed? Namely, how should regulators manage the apparently unavoidable trade-off between the desirability of ethics review that takes into account a variety of perspectives (including local community views and values) and the practical burdens and barriers created by subjecting multisite studies to multiple ethics reviews?

Our experience sheds light on the tension but provides no simple answers. Nonetheless, it throws up several insights and ideas that may help to inform the HoMER project team as it tackles the challenge of streamlining ethics review in the context of multisite Indigenous health research.

First, the HoMER project is a welcome initiative. Although centralisation of the formal ethics and scientific review processes will be an incomplete solution for many types of multisite studies, it should help. If implemented effectively, a harmonised system will reduce barriers for all types of multisite research, chiefly by consolidating the HREC component of the process. We also believe that it has the potential to strengthen rather than weaken ethics oversight of multisite research.

Second, linking community consultation to the HREC process, by making the former a pre-condition of the latter, is problematic. Both are important, but they are qualitatively different steps. The National statement on ethical conduct in human research specifies that research methods be “respectful and acknowledge the cultural distinctiveness of discrete . . . communities or groups” and calls for “evidence of support for the research project from relevant Aboriginal and Torres Strait Islander communities.”29 On the ground, the engagement process involves trust and relationship building, and occurs at several levels. For example, basic courtesy dictates the need for introductions and reasonable notice before researchers “arrive”. Logistical necessities, particularly when local staff are involved or the research intersects with existing services, demand another type of dialogue. Ensuring that community leaders are comfortable with and support the study requires a different form of engagement again — done well, it involves discussions before, during and after the field component of the research.

Drawing these various layers into the HREC process through a requirement of sign-off from community organisations runs the risk of making this part of the review haphazard, reductionist and artificial. In our experience, the basis for determining which community organisation’s approval should operate as a precondition to HREC approval was unclear and variable. HREC expectations of the community’s response — essentially a red or green light for the study — will often be out of step with the tenor of the dialogue between researchers and community members. Another way of articulating the problem is that HRECs tend to handle community approval as a kind of population-level analogue of the contemporary bioethics concept of a patient’s informed consent, but the analogy is not well suited to the realities of community engagement.

We believe that a more appropriate model would be for researchers to include an outline of an engagement strategy in their ethics application. The strategy would detail the project’s consultation activities, including the plan for approaching communities, how local Indigenous people will be engaged, and what governance arrangements will guide the process. The ethics committee would review the strategy and, through the standard mechanisms for ongoing monitoring of research projects, researchers would then report annually on the strategy’s roll-out. Ethics committees should be notified promptly of any problems encountered or departures from what was approved, akin to notification of adverse events.

Third, any overseer of engagement processes must recognise the need for flexibility. In multisite studies, the processes will often need to vary according to the scope and nature of the research and the types of communities involved. In some discrete Aboriginal communities, the appropriate organisation to engage in the research process may be obvious. But when the Aboriginal community is a relatively small minority dispersed within a larger population group, the appropriate structure and process may be less clear; in such cases, researchers will need time to develop their strategy, and latitude regarding how it is executed on the ground.

Fourth, in parallel with the centralisation of formal ethics review for multisite studies, options for centralising community consultation processes should be explored. This suggestion will strike some readers as odd: how can local matters be handled by central bodies? Consider a committee constituted of local leaders and stakeholders from a mix of communities who convene regularly to review multisite study proposals, focusing on issues relevant to communities and special populations. The committee could be a standing one, but with several rotating seats to ensure a degree of representation from certain communities and groups directly connected with the project under review.

Critics may reject this last idea as destined to be insufficiently sensitive to local concerns. However, the goal must be a reasonable balance between the accommodation of local concerns and a level of efficiency that does not stymie valuable research. Tipped too far towards efficiency, the research may be ill-informed and lack adequate ethics oversight. Tipped too far towards local consultation, social benefits that flow from multisite studies in important fields of research, such as Indigenous health, will fall out of reach.

2 Entities involved in ethics review and consultation for the National Indigenous Eye Health Survey (NIEHS)*

* Shaded boxes with a solid outline indicate entities from which ethics approval was required; unshaded boxes with a solid outline indicate community organisations from which consent was not a precondition to approval by the relevant ethics committee; and boxes with a dashed outline indicate other state and national entities that were consulted, typically at the suggestion of an ethics committee or community organisation, whose consent was not a precondition to ethics approval. † Although this was a hospital-based ethics committee, it operated as a committee with state-wide approval authority for purposes of the NIEHS.