In this issue of Quality and Safety in Health Care, there is a paper that should stimulate considerable debate (see page 99);1 indeed, we have published three commentaries alongside the paper2^–4 to initiate this and a further response from the authors (see page 90).5 Despite major initiatives to improve patient safety, there is a perception that attempts to improve safety have made slow progress. Hence, Coiera and Braithwaite argue for the implementation of market based control mechanisms as an incentive to promote patient safety. Their proposal is modelled on the “cap and trade” approach to creating a market in emissions trading, a key component of the Kyoto protocol that allows organisations that are successful in reducing carbon emissions to sell credits to organisations that have been less successful.

The parallel between emissions trading to improve the environment and patient safety event trading to improve healthcare safety is fascinating; each of our commentators is intrigued by the proposal. However, each of them believes that the model, while intriguing, is unlikely to be implemented or effective. Coiera and Braithwaite have responded to these commentaries with a further robust argument!

Chin and Wilkes2 are concerned that a market of this type will lead to further widening of inequalities through a variety of mechanisms including lobbying power, resource constraints, ability to pass or shift costs and gaming. If market mechanisms were to enhance inequalities, this would be unacceptable (see page 88). While Chin and Wilkes believe that measurement of patient safety and holding healthcare organisations accountable in a fair manner, consistent with Coiera and Braithwaite’s proposals, are important, neither is unique to market solutions.

Donaldson3 expresses a fundamental concern—he argues that healthcare is not a public good in the same way as the environment (see page 87). He also points out that the introduction of quasimarkets in healthcare has been largely unsuccessful in addressing issues of quality and safety. Instead he calls for more explicit and better developed methods to determine priorities for investment in constrained healthcare systems. He also, quite rightly, raises the question as to whether the emissions trading model has yet shown itself to be effective—indeed, Coiera and Braithwaite themselves accept that it is too early to evaluate that.

Meltzer4 points out that incentives for patients and payers to avoid errors through competitive market forces already exist, in contrast to carbon emissions prior to trading, thus making the argument less compelling (see page 86). He highlights the challenge of measurement and of how an appropriate level of adverse events might be set. He also believes that such a system is likely to increase healthcare costs. Meltzer flags up one element of the proposal that he describes as compelling; that is the underlying assumption that some level of harm is appropriate or acceptable because reducing harm is costly.

In addition to the concerns expressed in these commentaries, I believe there are several other issues that need to be considered before pursuing an MBC approach. First, this approach is very top down; it appears to ignore the importance of engagement of healthcare staff in improving safety. The mantra of “first do no harm” is embedded within the culture of most healthcare professionals, and when patient safety incidents occur, they are rarely due to negligence or intended actions but largely reflect the inevitabilities of human error and the inadequacies of systems. A top-down model such as that proposed here is likely to provoke resistance among professional groups.

It is also likely to provoke resistance among patients and the public. What level of acceptability would this engender within the public domain, particularly given Meltzer’s comments that an underlying implication is that there is a level of acceptable harm? One of the challenges to patient safety has been the fact that the value placed upon harm produced by healthcare is often quite different to the value placed upon injury or ill health arising de novo. Fundamentally, patients access healthcare with the expectation that it will make them better and find the concept that it might make them worse very difficult to understand.

Another issue of relevance is the complexity of healthcare. The issue of carbon emission is arguably much more straightforward in both its measurement and its aetiology than harm caused by healthcare. This complexity in healthcare may explain why some of the methods of quality and safety improvement that have been effective in industrial settings are more difficult to apply in healthcare. Market-based control is likely to be similar in this respect.

A key problem, also flagged up by our commentators, is that of measurement. We know that incident reporting significantly under-reports for a variety of reasons. Equally, there is evidence to suggest that those organisations that report more incidents have a better and more effective safety culture.6 Any market-based mechanism that penalised higher rates of incidents would have the potential effect of switching off the tap of reporting, upon which much safety improvement depends. The approach would be replete with perverse incentives. Furthermore, the use of measures of safety or quality from routine information systems, such as the AHRQ indicators suggested by Coiera and Braithwaite, would need to take account of the fact that routine data quality and completeness are hugely variable, not only across different healthcare systems but also within countries, again emphasising the considerable challenge of measurement of safety within this proposal.

In conclusion, this is a fascinating and intriguing proposal that merits wide debate. But is this proposal something that healthcare systems should take seriously and consider testing?

Could this approach dramatically improve safety within healthcare systems, or should we once again heed HL Mencken’s view that “For every complex problem there is an answer that is clear, simple, and wrong”?