Katharine Browne raises doubts about the motivational model upon which pay for performance (P4P) schemes are built. Physicians are not selfish knaves.
Pay for performance (P4P) schemes are alternatives or supplements to the traditional ways of paying physicians. These schemes are adopted as a means of boosting quality and efficiency in healthcare delivery. Improving efficiency is an urgent concern with the expiry of Canada’s health-care accord.
P4P schemes attach financial incentives to care and services in order to motivate the achievement of certain quality measures. These measures include both specified targeted outcomes (such as the achievement of lower cholesterol levels in cardiac patients), as well as the performance of specific activities (such as advice to increase exercise or to quit smoking).
P4P schemes are well established in the United States and the United Kingdom, and have started to make their way to Canada. The implementation of P4P formed part of Ontario’s Primary Health Care Reform – an undertaking to improve access to, and quality of, care at the first point of contact between the patient and the healthcare provider.
The success of P4P is hard to measure, but it hasn’t been overwhelming. For example, it appears that there have been minimal increases in efficiency and quality of care.
P4P can also produce unintended consequences, such as physicians gaming the system in an attempt to reap financial rewards without actually meeting the specified outcomes. Consequently, much discussion surrounding P4P schemes has focused on which activities should be incentivized, the appropriate level of incentive needed to generate the desired result, how to rank and value outcomes, and how to monitor and assess performance. In short, significant energy has been devoted to determining how to incentivize, and comparatively less energy has been devoted to the question of whether to incentivize. This is a serious mistake.
The debate surrounding P4P often overlooks the appropriateness of the model of human motivation upon which such schemes are built. All financial incentive schemes are founded on the selfish actor model, according to which individuals are presumed to be selfishly motivated to maximize the satisfaction of their own preferences. On this assumption, accompanying policies follow a carrot and stick methodology. The idea is that if individuals are driven by self-interest, we can reward and punish them to get them to do what we want them to do.
But the growing consensus is that the selfish actor model is deeply flawed. First, experimental evidence in behavioural economics reveals higher levels of other-regarding behaviour than the selfish actor model predicts. For example, when two economists performed the classic Prisoner’s Dilemma experiment on actual prisoners, those prisoners cooperated far more than the model predicted.
Second, theoretical advances in evolutionary biology have challenged the selfish actor model. It has been replaced with a picture of humans as a cooperative species with psychologically altruistic elements that allow us to live together successfully in groups.
Third, the incentive structures that accompany the selfish actor model often lead to undesirable results by crowding out other motivations. For example, a well-known daycare study revealed that fining parents who arrived late to pick up their children actually increased the percentage of tardy parents.
Finally, there is evidence that management techniques in line with something other than the selfish actor model can generate more desirable outcomes. Toyota’s success in the automotive industry has been attributed to its innovation in management. It has shifted away from the assembly line model and its inherent presumption of self-interest, to an emphasis on other values such as teamwork. Notwithstanding recent trouble over safety standards, it is hard to deny the success that Toyota has enjoyed.
This suggests a need to revise the selfish actor model, along with the policies that are shaped by it. Julian Le Grand’s work on policy responsiveness to different motivational structures exemplifies this approach. Applied to health care, Le Grand says that if we assume physicians are selfish ‘knaves,’ then health care policies should rely on punishment and incentives. If they are altruistic ‘knights,’ then policies should aim to leave physicians to do their jobs. If they are non-independently motivated ‘pawns,’ then policies should explicitly direct physicians to do the right thing.
From all this we can draw three main conclusions. First, the selfish actor model and P4P schemes are challengeable. Second, the flaws of the selfish actor model provide a diagnosis for some of the concerns raised about P4P schemes. Finally, a rejection of the selfish actor model could transform the health care industry in the way that Toyota transformed the automotive industry.
Katharine Browne is a Postdoctoral Research Fellow at the Centre for the Study of Mind in Nature (CSMN) at the University of Oslo, Norway. She will be joining the Novel Tech Ethics research team in the fall. @BrowneKatharine