Clinical registries: the urgent need to address ethical hurdles

Australia is committed to improving the measurement of its delivery of health services in order to improve quality and safety of care and increase transparency for consumers.1 The Australian Commission on Safety and Quality in Health Care, established as an independent statutory authority in June 2011, has responsibility for leading this work. On its program is the drafting of national arrangements for clinical registries to monitor quality of care.2 Clinical registries synthesise patient treatment and outcome data, and report back to clinicians and contributing hospitals to allow outcomes and patterns of practice to be benchmarked (after risk adjustment) against other Australian and international units.

Currently, nine registries operate in Australia at the national level.3 In developing a national registry program, we can learn from the experience of Sweden, where registries account for more than 25% of all health expenditure. Sweden has invested heavily in the development of information technology infrastructure to enable efficient and effective data-collection capacity. Registries are solidly supported by clinicians who routinely use and report registry findings back to clinical units.4,5 A study has identified that by investing $70 million annually, Swedish registries will return a cost savings over 10 years of over $7 billion.6 Not surprisingly, expert groups in the United States have also been calling for the expansion of their clinical registry programs.5

Registries are currently and incorrectly understood to be databases primarily established for research purposes, rather than as entities which monitor the appropriateness and effectiveness of health care. In order to establish registries, researchers are generally required to submit applications for ethics approval to each hospital ethics committee from which patient data will be sought. Some ethics committees have approved an opt-out approach to consent for patient participation in data collection, while others have required registries to actively seek consent from patients (opt-in) for inclusion of their data in clinical quality registries.

Registries required by ethics committees to adopt active opt-in consent processes rarely achieve an acceptable recruitment rate, and response bias is common. A New South Wales pilot trauma registry recently received ethics committee approval to recruit patients using an opt-in consent approach. It achieved a recruitment rate of 53%.7 In contrast, the National Joint Replacement Registry and the Victorian State Trauma Registry rely on opt-out consent and have achieved recruitment rates of more than 99%.8,9 A Canadian registry using opt-in consent achieved a recruitment rate of 39%, despite having dedicated research nurses at contributing sites recruiting patients.10 That registry found that those who opted in were significantly more likely than those who refused consent to be alert at admission, to be alive at discharge, and to speak English or French.

In the case of registries designed to monitor and improve quality of care, opt-out consent minimises recruitment bias and ensures that all groups are well represented in outcome assessment. Registries depend heavily on the acceptability of their results to clinicians. However, this support would not be sustained if the outcome results could be manipulated or subjected to bias. The high recruitment rate offered by opt-out consent provides a strong defence against these concerns. Opt-in consent discriminates against sicker, disadvantaged and minority groups. This then means that feedback provided to government, hospitals, clinicians and the public might be misleading, and policies developed using results from unrepresentative data sources may result in misallocation of resources or misplaced and unnecessary interventions. Registries that are unable to obtain data from the majority of the eligible population are capable of providing little, if any, return on investment, and have significant potential to do harm. This makes such projects unethical.11 If registry data are to be used to monitor quality of care, it may be appropriate and even desirable for data to be collected without the requirement for consent.

In considering applications to establish clinical registries, ethics committees must be satisfied that consent is given voluntarily, based on sufficient information and adequate understanding of the risks to them, and that patients have the capacity to consult others before agreeing to participate.11 Those opposing opt-out consent cite concern that patients might not receive and understand information about the registry or have the capacity to opt out. To address these concerns, registries must ensure that every effort is made to ensure that information reaches the patient in a language they understand and that patients or their next of kin can opt out. Privacy concerns exist regardless of the consent model used at recruitment, but they can be minimised by ensuring that data are housed securely, a data-access policy is implemented, access to data is restricted and the governing body is aware of its responsibility for safeguarding patient information.

If registries are to be relied upon as pivotal tools for monitoring quality of care across our health services, current thinking about registries and how they should be established and operate needs to be revisited. An opt-in approach to recruitment, where the most vulnerable are also the least likely to be included, is neither ethical nor acceptable. Nor is it sensible to expect registries, whose primary purpose is to promote quality of care (that might otherwise be thought of as a responsibility of government and/or health services), to apply to numerous hospital ethics committees for approval and receive responses with varied and sometimes inconsistent conditions.

It is undoubtedly true that an important secondary function of registries is the provision of a data spine for high-quality research. Given the cumbersome processes registries are required to go through to obtain ethics committee approval from each hospital before conducting research, the level of research activity that they currently achieve is remarkable. A more sensible approach would see the establishment of national, registry-specific ethics committees responsible for overseeing registry research governance and activity. One of the most important features of registries is that they have the support of clinical groups; it is important to ensure that changes in structure do not jeopardise this relationship. Without clinical buy-in, registry data will cease to be respected and used by clinicians to drive practice improvement. We therefore suggest that clinicians and other registry stakeholders, including patient representatives, be involved as members of such ethics committees as well as of governance bodies for the registries.

In December 2011, a National Health Information Agreement was signed by federal and state health authorities, the Australian Institute for Health and Welfare, the Australian Bureau of Statistics and the Australian Commission on Safety and Quality in Health Care, committing all parties to working together to collect and facilitate the use of health information in the national interest.12

We propose that a formally constituted and incorporated umbrella registry be party to this agreement. The umbrella body should represent registries focused on high-risk areas of medicine,13 which will be accredited in much the same way as pathology providers and hospitals, and will be approved to collect, synthesise and report clinical information. Registries collecting identifiable patient information should be required to demonstrate that they have a secure system for data capture and housing. To be accredited, registries will need to show that data are reviewed regularly, appropriate systems exist to address aberrant findings and output is being used to drive practice improvement. Operating principles and technical standards exist to help guide this work.14 To ensure that registries do not become repositories of data remote to research, they are best housed within an environment that fosters research activity.

Australia is at a crossroads in the development of national registries to monitor quality of care. We can continue to invest in developing registries using an inappropriate approach that equates registries with research projects, requiring them to apply for approval to each hospital ethics committee, and creating a patchwork approach across Australia — or we can make important changes to improve the efficiency and effectiveness of data collection to drive quality of care.

Inclusion of data in high-quality registries should be part of a health service, and funding should be contingent on this occurring. Making changes to the way in which registries operate will see the Australian population benefiting from rigorous rather than ad-hoc monitoring of quality of care.