Barriers to clinical audit

In this issue of the Journal, Ofri and Spillane1 lament the difficulties in establishing a clinical quality registry, with much of the frustration directed at the research governance framework, particularly site governance. In this editorial, we compare this with our experience in establishing a national clinical quality registry across Australia and New Zealand and point out some of the problems with the current system of governance for clinical quality registries. While some progress has been made towards making research approval easier, particularly by having a (nearly) nationally accepted ethics approval process for public hospitals, there are still fundamental problems with the way pragmatic research and quality improvement are governed. Without major structural change, the difficulties associated with unnecessary steps will persist. The current research governance framework was constructed to protect patients being treated under experimental conditions. The problem is that the same framework applies to comparative effectiveness research, where health care interventions already in use are compared. In this type of research, the risk to participants is theoretically no greater than the risk of the usual care. Further removed from experimental research is the use of clinical audit for quality improvement and the use of clinical quality registries. These activities are observational in nature and use systematically collected data to simply monitor treatment for a specific condition or intervention with the intent of identifying opportunities to improve care and outcomes. They do not control the treatment or intervention given, yet these registries are also subject to the same ethics and governance requirements used for new and potentially harmful therapies. As described by Ofri and Spillane, these hurdles can be so onerous as to make the audit impractical. We have experienced similar hurdles in the establishment of the ANZ Hip Fracture Registry, which began in 2012 and is now a full binational registry. Although many along the way agreed that the research governance requirements seemed unnecessary or onerous for this audit of usual care, it was the only system available to achieve the aims of the Registry. As a result, registries need to employ people to specifically deal with these requirements, as the national research framework requires an ethics review and then separate governance reviews for all sites involved. This makes registries covering many sites unmanageable without dedicated staff. The inefficiency of the process is highlighted in the observation that when identical applications are submitted to multiple research governance organizations, the purported deficiencies in the application that require attention are almost universally different from each organization, or, as Ofri and Spillane found, relate to formatting of documents rather than content. Research governance is currently driven by risk mitigation rather than societal benefit. When there is doubt regarding risk, approval is not given and another round of the application process is triggered. Quality improvement uses data to change local processes to achieve specific standards of clinical care2 and the risks of not having these processes in place is not considered in the review process. National or binational registries, like the ANZ Hip Fracture Registry are essentially large-scale quality improvement projects designed to bring about immediate improvements in the way health care is delivered to patients at the time of care. Ofri and Spillane call for centralized governance oversight, but this is difficult because local sites vary in their ability to participate, so some local input at each site is inevitable. However, we argue that the societal benefit achieved by systematically auditing usual care makes providing treatment without audit unethical. It is an expectation of practicing clinicians to undertake audit and quality improvement activity with a view to constantly improving care. We must have a system that accommodates this default position. We contend that to continue intervening without audit (the path of least resistance) is a disservice to patients which in the case of breast cancer has the potential to limit life expectancy. Much is made of consumer involvement in research, but this is not the case for research governance. Most people would agree that audit of clinical practice should be mandatory and a 2021 systematic review indicated that health care consumers were generally supportive of using health data for research, particularly if the data is de-identified or anonymised.3 A governance framework for audit is needed that enables and facilitates quality improvement activity. Failure to do so is a missed opportunity that ultimately impacts the lives of people.