Pathways to enhancing the quality of stroke care through national data monitoring systems for hospitals

In Australia, over 50 000 new strokes occur each year and 420 000 people live with the effects of stroke.1 All are at high risk of recurrent vascular events. Evidence exists for effective treatments to prevent stroke and subsequent disability,2 but these treatments are not universally applied, creating unwarranted clinical variation and gaps in care.3

Monitoring quality is a major focus for funders, providers and consumers of health care. In Australia, as in other countries, there is a range of systems in place to collect data, but this creates inconsistency, duplication and data excesses.4,5 Since most people who suffer acute stroke are admitted to hospital,6 an integrated system of monitoring care in this setting has strong potential to provide an appraisal of the quality of care.

In addition to government-collected hospital data, two complementary national programs currently exist for assessing adherence to clinical guidelines for stroke care: the National Stroke Foundation (NSF) audit program3 and the Australian Stroke Clinical Registry (AuSCR).7 The NSF audit program, which alternates annually between acute and rehabilitation hospitals, consists of an organisational survey of the resources available to provide stroke care, and a detailed, retrospective clinical audit of at least 40 consecutive eligible patient health records.3,8 On completion of the clinical audit, the NSF provides participating hospitals with a report comparing their data with the national results. In Queensland, the NSF has also established the StrokeLink program in partnership with Queensland Health. StrokeLink is designed to enhance the audit feedback process with additional outreach visits to identify local barriers and develop action plans for practice change.9 Conversely, the AuSCR addresses the limitations of routine (administrative) hospital data that omit key quality-of-care indicators (eg, access to an acute stroke unit). Use of the AuSCR ensures that all eligible patients admitted to participating hospitals have a standard minimum dataset (MDS) collected and an assessment of the health of survivors at 3–6 months after the index event.7 Currently, 44 hospitals (including three private hospitals) in New South Wales, Queensland, Tasmania, Victoria and Western Australia have registered over 12 000 episodes of care in the AuSCR since 2009. The AuSCR provides live, web-based reports to hospital clinicians to enable real-time tracking of adherence to clinical indicators, patient characteristics and outcomes.

The NSF audit and the AuSCR form important infrastructure for monitoring the quality of hospital stroke care and add value to existing government data. They have independent systems, and share consistent definitions for common variables, but use separate secure web-based data entry systems. This creates duplication of effort through data entered manually. The NSF audit collects de-identified data, whereas the AuSCR includes personal identifiers to allow the follow-up of patients in the community and data linkage with other datasets. Harmonising national data collection processes and establishing reliable linkage of data from different repositories has the potential to improve efficiency given constrained resources for these activities.

The desire for a national system for monitoring the quality of stroke care is endorsed by the Australian Stroke Coalition, an alliance of clinical networks and professional organisations working in the field of stroke established in 2008,10 and the Australian Council on Healthcare Standards Clinical Indicator Program, which endorsed a stroke MDS as part of the National Performance Indicator Set (NPIS) variables.11 The NPIS data collection is voluntary and includes the four AuSCR clinical indicators as well as a subset of four additional NSF audit variables.

In September 2011, a workshop was held to achieve consensus on recommendations and outline the next steps for implementing a national approach to data collection and quality improvement for stroke care in Australia. The workshop was convened by the AuSCR consortium, which consists of The George Institute for Global Health, the National Stroke Research Institute (now part of the Florey Institute of Neuroscience and Mental Health), the NSF and the Stroke Society of Australasia (SSA). It was held in Adelaide on the day before the annual scientific meeting of the SSA. The workshop was attended by experts in stroke and health systems researchers (14 clinicians, four government representatives, three non-government organisation representatives, two consumers and 12 university academics and researchers). The Box outlines the key outcomes of the workshop.

Since the workshop, several activities have been organised to progress the vision for a national monitoring and quality improvement system for stroke. This has included further uptake of the AuSCR in Queensland and funding from the National Health and Medical Research Council (NHMRC) for a Partnerships for Better Health grant known as Stroke123 over 4 years from 2012 to test several recommendations from this workshop (Box). The primary objective of Stroke123 is to show whether comprehensive data, coupled with an evidence-based clinical practice improvement program, is more effective than the status quo. This will be achieved by obtaining a better understanding of the data linkage potential for stroke data across Australia using the AuSCR data to link with government and non-government datasets (eg, NSF audit data), and by undertaking a non-randomised, multicentre, historically controlled, prospective cohort study in Queensland. A second national workshop was convened by the Florey Institute and NSF in Melbourne in April 2013 to provide an update on progress, and to discuss issues specifically related to stroke data linkage projects.12 The recommendations of the initial workshop were reinforced as ongoing objectives to be achieved.

In summary, much can be gained from bringing together diverse groups with a vested interest in a single, clinical population (patients with stroke) to discuss important issues related to the routine monitoring of the quality of care in hospitals. The issues reported in this article are not unique to one clinical population, and the strategies developed are broadly relevant to other conditions. We hope this experience will encourage further discussion and progress in improving clinical care monitoring in the health sector. Moreover, any integrated system for monitoring the quality of care should incorporate an active process of continuous quality improvement aimed at effecting change in clinician behaviour when an area of underperformance is found.