This Community is Archived.

This community is no longer active as of December 2018. Thanks to those who posted here and made this information available to others visiting the site.

Martin Roland, BM, BCh, MD on the Effects of Financial Incentives on Provider Behavior

Today we are delighted to welcome Martin Roland, BM, BCh, MD as a Keynote Speaker for this Breakthrough Opportunities event on Designing Provider Incentives with the Commonwealth Fund.

Dr. Roland directs the Cambridge Centre for Health Services Research (CCHSR), a joint research centre between the University of Cambridge and RAND Europe.

In this video, Dr. Roland discusses the positive and unintended negative effects of financial incentives on provider behavior, such as those seen with pay-for-performance models.

We look forward to hearing your thoughts on Dr. Roland’s video below, and encourage you to submit ideas sparked by the topics covered:

(Note for those viewing this update via email, please click through to view Dr. Roland’s video and read the video transcript on GHDonline: )

Due to the audio quality of Dr. Roland's video, we have included a transcript below.

Video Transcript

I’m Martin Roland, I’m professor of Health Research at the University of Cambridge in the U.K. and for 35 years I was a practicing family practitioner.

How have you seen provider incentives affect quality of care?

Provider incentives can have an impact, but they don’t have a very great impact, and they certainly don’t have the impact that payers often think that they’re going to. I think what the evidence and the literature do show is that if you combine quality improvement methods of different sorts that might include pay-for-performance [(P4P)] and provider incentives, then you can get substantial changes. But expecting P4P on its own to produce a big effect doesn’t usually turn out to be successful.

What aspects are key to creating effective incentive programs?

It’s important to think about the possible effect of financial incentives on disparities; there has always been concern that these types of schemes might widen disparities in care. For example, patients from more affluent areas might be easier to treat, easier to reach targets of one sort or another, than patients living in more socioeconomically deprived areas.

There are a number of ways that have been described in the literature of trying to avoid that effect, either by giving specific incentives to people working in socioeconomically deprived areas, or arranging the payment system in some way to avoid those disparities.

This is a real example which faced me some years ago in an earlier P4P scheme in England, where there was an incentive for a primary care practice to achieve 50% of all the women who were eligible for cervical smears having had a pap smear, and there was another incentive for reaching an 80% target. Now this is in an area with a high ethnic minority population and where it was quite difficult to get women to come in for their pap smears. We could relatively easily reach the 50% target, in fact we could get 70%, but 80% was extremely difficult. Above 70% there was very little incentive to do more because there was no additional resource coming into the practice unless we got to 80%.

Many schemes might have a threshold below which they wouldn’t pay anything, and above which they might say well that’s the maximum you might reasonably achieve, but would have some sort of incentive for improvement between those bounds.

It’s absolutely critical that clinicians are involved in the development of those programs. What you need to achieve is an alignment between the professional incentives physicians have when they’re seeing their patients, with managerial incentives of the organization that they’re working for or around, with the financial incentives. You’re trying to get the professional incentives, managerial incentives, and the financial incentives all going in the same direction. If you get them going in opposing directions, that’s when you can really run into problems, where, for example, the financial incentives are getting physicians to do something that they don’t really believe that they should be doing.

What are the challenges associated with designing provider incentives?

One of the key questions I think is the issue of whether you’re incentivizing individual physicians, or what you do when the physicians and other clinicians are working within groups. In the U.K., when we had a hospital-based P4P scheme which was addressing issues both of cardiac surgery but also specialist medicine in hospitals more generally, the incentive there was not to the physicians who were by-and-large salaried, but the incentive was to the team so that they could engage additional staff. So, for example, they could engage a nurse who would focus on quality improvement activities, and help on those things within the team. Not denying the importance of non-financial incentives, there is a range of ways in which you might construct financial incentives to benefit team-members and the team as a whole in different ways.

Incentive schemes can have unintended consequences and people often forget that those are likely to happen. It’s important to anticipate what the unintended consequences might be, and try to reduce those to a minimum. Part of that will depend on the size of the incentives, and this is something that is quite difficult for payers to get right. Talking still about financial incentives, if they’re too small they are not going to have an effect. I think it’s pretty clear that very small incentives don’t influence physicians’ behavior. On the other end, if they’re too large, that will increase the chance that you will get unintended effects, maybe physicians doing things for patients that aren’t necessarily in their best interest, or straightforward gaming, or the risk that physicians might actually cheat. So it’s getting a balance there so that you get an enhancement of professional incentives with as few adverse consequences as possible.


Matteo M Galizzi Advisor Replied at 4:01 AM, 10 Dec 2015

Prof Roland raises several key issues. First, despite what is commonly argued in policy debates, there is little rigorous evidence on the overall effectiveness of financial incentives schemes on providers' performance.
Second, there is clearly large untapped potential for further experimentation in this area. For instance, given the inherent heterogeneity of socio-demographic, health, and ethnic conditions in countries such us the UK and the US, financial incentives should be designed in a way that they account for risk-adjusted performance. Also, at an organizational level, there is relatively little done in the direction of what Prof Volpp and Prof Loewenstein are doing in the area of individual financial incentives in health: that is, 'supercharging' financial incentives with behavioural economics insights, and for instance combining financial incentives with loss aversion, overweighting of small probabilities and similar. The reason is that organizations are usually assumed to be less likely to fall prey of 'behavioural' biases that affect individuals. Rigorous evidence on this point, however, is scarce, and there is also some evidence from the behavioural science that organizations that aggregate views from many different interacting stakeholders may amplify rather than reduce the behavioural biases that are present at an individual level.
Third, behavioural science suggest that, at an individual level, financial incentives work best when the financial incentives are conditional on objective which are clearly specified, simple, clearly measurable, preferably one-shot, and are under full control of the individual who is therefore fully accountable for the change in behaviour. There is thus clearly a trade-off in the context of healthcare organizations, where the objective of the performance should typically depend on a holistic outcome which encompasses the complex aggregation of a broad range of efforts depending on a multiplicity of healthcare actors in the organization.

This Community is Archived.

This community is no longer active as of December 2018. Thanks to those who posted here and made this information available to others visiting the site.