Can clinical governance deliver quality improvement in Australian general practice and primary care? A systematic review of the evidence

Improving the quality of primary health care is a complex undertaking. Over the past decade, clinical governance has been promoted as a systematic, integrated approach to assuring safe, good quality health care.1 Its intent is to move beyond the organisational “magic bullet” of single strategies (eg, professional education, audit, or risk management) to a systematic, multifaceted approach to quality improvement using a range of locally implemented strategies.

In Australia, most work on clinical governance has been in hospitals2 In the past 5 years, however, there has been an emerging interest in clinical governance in primary health care. Failures in health care delivery occur frequently in primary health care, just as they do in hospitals. In a study using anonymous reporting of adverse outcomes in Australian general practice, errors were estimated to occur in one in 1000 consultations;3 of 433 errors, nearly 70% were attributed to failures in the processes of health care, and 30% to the failure of individual general practitioners’ knowledge and skills.4 Partly in response to this, the Royal Australian College of General Practitioners (RACGP) recently introduced clinical governance into its accreditation standards.5 The Victorian Healthcare Association has developed a range of resources and guidelines for clinical governance in community health.6 Improving quality, safety, accountability and performance is one of the four priorities in the draft national primary care strategy.7 The final report of the National Health and Hospitals Reform Commission (NHHRC) advocated strategies and infrastructure for quality improvement, and these are integral elements of the NHHRC’s recommendations on continuous learning, monitoring and interprofessional and intersectoral collaboration.8

Given the policy development in this area, there is a need to clarify models of clinical governance and the evidence relating them to quality improvement. Here, we review the literature on different models of clinical governance, exploring their relevance to the Australian primary care sector, and their potential impacts on quality and safety.

The key question we examined was: What models of clinical governance deliver quality of care? We undertook a systematic review and realist synthesis of the literature.9 Realist synthesis is particularly relevant for complex social interventions, as it aims to develop explanatory analyses of why and how these interventions may work in particular settings and contexts. It is therefore relevant for policymakers who are keen to explore whether or not a model may work in a given context.10 Realist synthesis involves developing theories and models of how mechanisms work, using careful analysis of the literature and interviews with stakeholders.

We developed an operational definition of clinical governance by reviewing health policy on clinical governance in Australia, the United Kingdom and New Zealand (Box 1). We also reviewed the seminal papers on clinical governance referenced in those policy documents,11-16 and in the position documents on systematised quality approaches in health care in the United States.17 Our definition of quality used the nine dimensions from the 2001 National Health Performance Framework18 (Box 2).

Eighty-seven studies were assessed as being of high quality (Box 3) and, of these, 19 explored the outcomes of clinical governance. These 19 articles are the focus of this review. Seven were randomised controlled studies, 11 were longitudinal observational studies, and one was a well theorised and constructed case study. Six studies were from England,20-25 two from the US,26,27 four from Australia,28-31 two from New Zealand,32,33 and one each from Spain,34 the Philippines,35 Belgium,36 and Germany.37 Only one study compared clinical governance models in two countries, Holland and the US.38

Most studies addressed capability.20-23,25-32,35-37 There were no studies that addressed continuity or appropriateness of care, and few that addressed responsiveness20,24,25,27,34,35 and accessibility.20,26,30,33,35 Only four studies addressed safety.24,25,31,37 Although many studies explored incentive systems, very few described a clinical governance process, but rather posited a kind of black box between the incentive and the outcome. As a result, only four studies25,30-32 explored the relationship between clinical governance and efficiency.

The models that emerged from the literature and interviews are summarised in Box 4. The strategies most used were audit, performance against indicators, and peer-led reflection on evidence or performance. Assessment of clinical competence was described relatively infrequently in the studies as a governance strategy; there are more articles describing the “enticement” of practitioners into clinical governance than “hard governance” measures like assessment of professional competence.39

Box 5 outlines the evidence for each model’s contribution to the quality domains. Interventions at different levels can improve capability of care, but this improvement appears to be related to the type of care and how easy it is to systematise care processes. Improvements in capability are more easily achieved in prescribing practice than in chronic disease management, where the evidence for improvement is varied and context-dependent. There are fewer data on effectiveness (ie, whether the intervention results in a measurable health improvement) than on capability (ie, whether the intervention results in an improvement in skills or practice), but in two uncontrolled studies, there was an improvement in the outcomes of chronic disease (angina,20 asthma20 and diabetes29) without an improvement in capability. Both these studies suggest that trends and/or unmeasured contextual effects lead to changes in outcomes irrespective of clinical governance measures.

Most evidence supports governance models which use targeted, peer-led feedback on the clinician’s own practice. There is less evidence supporting interventions across entire health systems, such as national level performance indicators, unless this is accompanied by appropriate levels of support, including infrastructure support. Top-down models to drive quality improvement (external accreditation, or economic incentives for quality measures for doctors) resulted in no improvements in a range of quality domains, and in one country with limited resources (Philippines),35 led to a decline in accessibility. Sustainability may also deteriorate under a top-down model of governance without support and infrastructure, a risk also noted at interview by workers within the Aboriginal health sector.

Of the 11 software packages most commonly used in Australian general practice, only four enable the user to participate in regional or aggregated data quality activities. No system had inbuilt data quality checks (eg, “How often are angiotensin-modulating drugs prescribed without renal function being checked?”). Most have some inbuilt searches for items that relate to funding initiatives or chronic disease management. The ability to do other searches was quite variable and often required significant computer and database knowledge. Only one package allowed patient access through a web portal, to which both the practice and the patient must have subscribed. Achieving some degree of standardisation may be an essential step in the implementation of clinical governance.40

An emerging theme in the interviews was the lack of clarity around the definition and purpose of clinical governance. Concerns were expressed that the concept might denote governance of or by clinicians, rather than strategic approaches to improving clinical care. There was resistance among professional groups to the idea of externally set performance indicators, which had the capacity to grow beyond indicators which would improve practice. The Aboriginal sector in Australia has pioneered the development of locally relevant performance indicators, and also reported on the organisational costs of reporting to multiple disconnected performance indicators. Many interviewees commented on the difficulties of using information systems to best improve practice in Australia, with a great deal of energy being expended in extracting data from systems that did not intuitively allow this. This was in contrast to other settings (eg, Kaiser Permanente, and Pegasus Health’s data on prescribing practice), where regionalised aggregated data were available and could be compared with individual practitioners’ data.

The evidence in favour of clinical governance in general practice and primary care is fragmented. The largest body of evidence relates to the areas that are most easily measured, such as the quality of prescribing practices. There is less evidence for quality improvement in chronic disease management and complex areas like health care for older people and mental health care. However, the development of outcome indicators in chronic and complex care is challenging, and the lack of evidence of the impact of clinical governance may reflect measurement difficulties rather than a genuine lack of impact. More research is needed into interventions to improve safety, sustainability, efficiency and responsiveness in primary care.

The most acceptable mechanisms to drive clinical governance are those that recognise professional leadership and are perceived as being locally relevant and allowing reflection on one’s own professional practice. There is a reasonable evidence base to support the role of well resourced peer support networks to facilitate clinical governance. Clinical governance strategies that rely on guidelines alone have not proven effective among GPs. Performance indicators, clinical standards and guidelines — advocated in the Australian Government’s response to the NHHRC report,41 and in the draft National Primary Health Care Strategy7 — have a role in guided reflection and quality improvement. However, exclusive or excessive use of these top-down mechanisms can result in opportunity costs for the service,42 disruption of teamwork if the reporting task is delegated,43 and constructing clinical activities around indicators rather than need.44 An increase in the measurement of common chronic conditions that were the subject of incentives, over others that were not, was noted after the introduction of the Quality and Outcomes Framework in the UK.45

The Australian Government’s response to the NHHRC report proposed combining GPs and state-funded community health systems under the one regional structure, Primary Health Care Organisations (or Medicare Locals).41 This would provide a platform for regional networks to support clinical governance,46 and would presumably build on structures such as the Divisions of General Practice network or Victoria’s Primary Care Partnerships. The UK experience suggests that networks of practices interact to seek corroboration of the value of new information and evidence within local networks, before these practices establish patterns of clinical governance.47 Trust may be undermined if regional-level involvement in clinical governance is seen to be only the collection of data for reporting purposes.

A central difficulty in introducing clinical governance is a lack of consensus among primary care workers about its meaning. Clinical governance remains a poorly understood term, often equated with bureaucratic control, or medical dominance.48 There is a need for regional or national organisations to display leadership in explaining and demonstrating what clinical governance actually involves. Examples of leaders in this area are the RACGP in their standards work and the work of the Victorian Healthcare Association.

The Aboriginal community-controlled sector is in the vanguard of clinical governance in Australia. The development of clinical indicators in some services (eg, Kimberley Aboriginal Medical Services49) preceded mainstream work in this area. Across the sector, there has been a concerted and long-term development of capacity and personnel to drive clinical governance activities. The opportunity cost of resource-poor services directing personnel to data collection for reporting purposes rather than for data translation to service improvement is also well understood within the sector. Input from this sector should be sought from others in Australia to inform the implementation of clinical governance across all primary health care.

The lack of good information on practice in Australia is a critical constraint for clinical governance activities. In some organisations in the US and New Zealand, regional networks have access to detailed service-use data. In Australia, this level of feedback is only provided through the National Prescribing Service. In all other areas, specific clinical software is interrogated to generate routine data on individual practices, but the software itself is frequently insufficiently usable by service staff. As a minimum, service providers should have access to their own data on prescribing, tests ordered, referral patterns, and the demographic patterns and illnesses of their patients. Ready extraction of these data with a unified coding system should be built into the standards developed for clinical software.

For clinical governance to become second nature in Australian primary care, health care practitioners need to be actively engaged as partners in quality improvement. The evidence is strongest for improvements that are driven by health professionals at the practice level, with support from regional networks. Such activity at regional and service levels needs support through structural changes initiated at the national level. This should include funding of time for clinical governance, supported by information systems which provide ready access for practitioners to their own clinical data.

3 Data searching and inclusion flowchart