An American physician called Ernest Codman advocated vigorously the publication of surgical mortality rates in Boston in the early twentieth century. His reward was the withdrawal of practice rights at Massachusetts General in 1914. Undeterred, he established his own hospital, published his mortality data and was bankrupted as his colleagues’ outcomes remained buried! Codman died a pauper.
Over a hundred years later it is proposed to publish mortality rates by surgical team in the English NHS. This proposal is the product of Sir Bruce Keogh leading colleagues in cardio-thoracic surgery to publish mortality rates by centre. These data can now be consulted on the Health Care Commission website. They are a product of review of local practitioners and data collection in each centre.
Sir Bruce is now Medical Director in the Department of Health. His remit is not only to develop outcome measures but also to improve the quantity, quality and timeliness of information about medical practice
Risk-rating outcome data can be problematic. In addition to the problems of small numbers and the fact that increasingly care is delivered by multi-disciplinary surgical and medical teams, where attribution of error may be difficult, there are inevitable incentive effects.
Publication of cardiac surgery mortality rates in Pennsylvania and New York led to surgeons declining to treat complex cases in order to improve their scores.
The production of distributions of mortality, crude and adjusted-for patient risk requires a “six sigma” approach to its investigation. Not all practitioners can be investigated in great detail and so usually management, clinical and non-clinical, will focus their attention on the tails or extremes of the distribution i.e. three standard deviations each side of the mean or six sigma.
Attention to these extremes identifies good and poor practice which can be investigated by peers. However, such data are threatening and may affect admitting behaviour. Publication of cardiac surgery mortality rates in Pennsylvania and New York led to surgeons declining to treat complex cases in order to improve their scores. This improved average performance but the excluded, high-risk patients had poorer outcomes1.
Such perverse effects will have to be carefully managed in the NHS. Also, it has to be emphasised that these data are only a part of a system of measurement of clinical performance that is emerging from Whitehall. In addition to comparative mortality data, the NHS is investing in hospital episode statistics (HES), patient level information on costs (PLICs) and patient reported outcome measures (PROMs).
Commissioners and providers are to be “armed” with comparative activity level for consultants, comparative PLICs and PROMs that assess pre- and post-treatment physical and psychological functioning.
Such armaments will be useless unless managed by the medical profession with vigour. Doctors will have the capacity to identify outliers in terms of activity, cost and outcomes, both mortality and PROMs. This transparency is a vital ingredient into professional transparency.
Confucius argued that “without trust we cannot stand”. If the profession is to prosper and deliver observably excellent care to patients, they will have to manage better and be seen to use these radical investments in comparative performance to improve accountability.
Reference:
- Dranove, D, Kessler, D et al, Health care report cards may fail patients, National Bureau for Economic Research, NBER working paper 8697, 2002
Author's competing interests: none declared
(Picture: Wellcome Images)