Bacchus Marsh baby deaths: Australia should learn from the UK and publish clinician performance data

The United Kingdom has been where the Victorian government is now. After a review confirmed 11 newborn and stillborn deaths at Victoria’s Bacchus Marsh hospital were potentially avoidable, the state is set to overhaul its health services.

It also took the deaths or serious injury of babies and children to change things in England. The General Medical Council’s 1998 inquiry into the deaths of 29 babies undergoing heart surgery in the 1980s and early 1990s at the Bristol Royal Infirmary concluded that there had been serious professional misconduct by three doctors. In response, the UK government decided to publish hospital death statistics.

The Victorian government’s review was chaired by Stephen Duckett, director of the health program at Grattan Institute. Its report into the safety of care in Victorian hospitals, released last week, concluded that the potentially avoidable deaths were the result of a series of catastrophic clinical and governance failures.

Clinical outcomes were not being monitored properly and there was an inadequate response when things went wrong. The report states that there is no way of telling whether “all hospitals are consistently providing high-quality, safe and continuously improving care”.

That’s a damning deficiency, and the authors set out 179 recommendations about how to reform the system for the better. Among these is a recommendation for government to legislate to establish a Victorian Health Performance Authority, independent from the department to:

provide the public with hospital safety and quality performance data on a quarterly basis that covers all safety and quality indicators against which hospitals are monitored, for both public and private hospitals.

But the report stops short of recommending that data about individual clinicians be made available to patients, parents and the general public. At the report’s launch, one of the authors argued that the Victorian system is simply not mature enough to allow publication of such data, particularly because it is difficult to make like-for-like comparisons between hospitals and doctors.

This was exactly the argument made a long time ago, in 1866, by the United Kingdom’s Royal College of Surgeons, to put a stop to the annual publication of death statistics from 1861 to 1865 for English hospitals.

The excuse no longer holds: like-for-like comparisons are now perfectly possible, as recent experience in England testifies. And comparisons can be made not just between hospitals but between individual doctors.

The Society for Cardiothoracic Surgery in Great Britain and Ireland led this initiative, collecting data about individual surgeons and devising its own approach to adjusting risk between surgeons. From 2004, the Society started publishing annual data on the internet about the performance of cardiac surgeons.

The Society also evaluated what happened next. Doctors weren’t foregoing more risky operations to protect their outcome data, because they were confident the risk-adjustment was correct. And survival rates following surgery improved across the board because doctors compared and questioned their performance relative to their peers. They’ve published the data annually for hospitals and individuals ever since.

Following this experience, publication of performance data has since been rolled out across the NHS, starting in 2013 with the publication of activity and death rates for hospital doctors in ten specialities. The NHS Choices website now provides information for every surgeon about the number and type of operations they have conducted and the 30-day, risk-adjusted death rates of their patients.

Performance is summarised as “OK” if these rates are within the safe range. But it’s also possible to see each doctor’s actual performance compared to the national average in their area of specialisation.

Information about the performance of maternity services includes the annual number of births, type of delivery, the proportion of mothers who are breastfeeding within 48 hours, and the results of the most recent inspection of the service by the Care Quality Commission, an independent health regulator.

The Royal College has also changed the stance it took 150 years ago, and now supports publication of performance data in Britain; the current release covering more than 5,000 surgeons for more than 28 procedures. As the Royal College of Surgeons now says, these data are designed to help patients understand “more about the nature of a surgeon’s work and their recovery after an operation”.

Publication of the data is also believed to be instrumental in driving performance improvements, because “surgeons are more likely to reflect on their practice and be inspired to improve while providing patients with accurate information on their surgeon’s outcomes”.

It’s taken a long time, and far too many cases of unavoidable deaths or harm to patients, but the medical profession in England is now fully committed to the publication of data about their own performance.

This latest report into the avoidable deaths at Bacchus Marsh suggests that such openness and transparency remain a distant prospect in Australia. This raises the question of whose interests are being protected by non-disclosure. It’s certainly not those of patients, parents and the Australian public.

Andrew Street has received project funding from the United Kingdom's National Institute of Health Research, the English Department of Health's Policy Research Programme, and the European Union. The views expressed are his own.