How useful is competition in improving hospital performance?

Over the last few years there have been attempts in England to investigate scientifically the claim that competition between health care providers leads to better outcomes. The arguments for and against were eloquently presented in a pair of papers in the BMJ in 2011 by my colleague Nicholas Mays (challenging the notion) and Simon Stevens, then President of Global Health at UnitedHealth Group, claiming there was good evidence to support the benefits of competition.

My personal view at the time was that, while the authors of three English studies that had been published had conducted the most rigorous analyses possible (Reform, Competition and Patient Outcomes in the National Health Service, What do we know about competition and quality in health care markets? and Does hospital competition save lives? Evidence from the English NHS patient choice reforms), the evidence was inevitably limited by researchers' reliance on routine administrative data (Hospital Episode Statistics, or HES).This meant they had to make some heroic but dubious assumptions (eg acute myocardial infarctions are easily clinically identifiable; adjusted mortality rates are purged of case mix differences).

As a result, I felt that their conclusions over-estimated the benefits of competition. At best it looked as though competition might have a marginal benefit, but not enough to warrant claims such as 'Death by market power'.

However, it was also clear that methodological shortcomings shouldn’t close discussion on this topic because if competition had benefits, it was vital that the NHS exploit this means of improving quality. To that end my somewhat forlorn hope at the time was that someone might find more appropriate data that was not subject to the limitations of HES.

I was therefore delighted, as co-editor of the Journal of Health Services Research & Policy, when we received a paper doing just that. Recently published, the paper by Yan Feng and colleagues makes use of the national patient-reported outcome measure (PROM) data.

Using PROMs enabled the researchers to assess the association of market concentration (a proxy for competition) and the outcome of hip replacement for 337 hospitals in 2010/11. And what they found was that competition had no significant influence on outcome.

Clearly this finding cannot necessarily be generalised to other surgical operations, let alone medical or mental health conditions. But given that in England the principal area of competition in health care is elective surgery, it seems unlikely that a stronger association would be found in other clinical areas.

My other reason for delight in being able to publish this important paper (revealing my bias, being the co-editor) is that it is another example of the usefulness and power of PROMs data for investigating key aspects of health services. In recent years these data have been used in a number of ways:

• to reveal the persisting slight (but statistically significant) social inequity in both the use of surgery and its outcome
• to show that patients distinguish between their outcomes (heath gain) and their experiences, thus scotching some clinicians' claims that PROMs just provide an indication of patients' experience
• to show that patients in areas with high surgical rates are in just as great a need for surgery (according to their self-reported heath status and quality of life) as they are in areas with low rates, thus throwing into question whether commissioners are justified in reducing investment in elective surgery in high rate areas.

While Yan Feng and colleagues' paper is not definitive proof of the lack of impact of competition on outcomes in England, it certainly should give policy-makers cause for thought before relying unduly on this fashionable chimera. There may be aspects of health care in which competition can indeed improve productivity (in the true sense, not just cost-cutting). But at present we still await rigorous evidence to demonstrate that.

Nick is Professor of Health Services Research in the Department of Health Services Research & Policy at London School of Hygiene & Tropical Medicine