Financial incentives for physicians to improve health care

KEY POINTS

• Accumulating research shows that the effects of financial incentives to improve the delivery of health care are mixed; where effects are observed, they tend to be modest and short lived.

• Barriers to providing better care may lie outside the incentivized physician’s sphere of control.

• Small tweaks to payments for physicians are unlikely to fix systemic problems and can cause harm if they mean that other opportunities for change are not addressed.

• We need to design payment systems that do not get in the way of better care, and accept that payments alone are not likely to deliver change where it is most needed.

Two linked research papers examine the use of financial incentives to improve access to follow-up care with community-based physicians after patients are discharged from hospital.^1,2 In one, a $25 premium was introduced for primary care physicians who saw patients within two weeks of hospital discharge.¹ In the other, premium payments were introduced for psychiatrists who provided outpatient care to patients in the month after hospital discharge or in the six months after a suicide attempt.² That both studies observed no change in follow-up visits after introduction of the incentive payments is not surprising. A growing list of systematic reviews suggests that any changes in care induced by financial incentives tend to be modest and short lived,^3–6 especially when incentives are introduced without complementary changes to the organization of health care delivery.⁷

Incentive programs include true pay-for-performance initiatives in which payments are tied to defined performance measures, notably the United Kingdom’s Quality and Outcomes Framework, introduced in 2004 to improve quality in primary care.⁵ There are also many examples of incentive schemes in which premiums are applied to desired processes of care with no new efforts to measure quality or outcomes, as is the case for incentives that target follow-up care.^1,2 A common logic underlies both: you get what you pay for. If you want higher performance, quality or value, or simply more physician follow-up after hospital discharge, design an incentive that rewards it.

A recent review of pay-for-performance programs found evidence of short-term improvements in some care processes, such as guideline-recommended screening and prescribing, but no impact on longer-term patient outcomes.³ A review of the Quality and Outcomes Framework literature found that there were short-term improvements in patient outcomes, but also that the improvements were not sustained.⁵ An earlier Cochrane review of the effect of financial incentives on care provided by primary care physicians similarly found positive but modest effects for some physician behaviours, such as screening, referral and recording patient information, but noted the low quality of studies and potential for selection bias.⁶ A systematic review of systematic reviews observed that incentive programs appear more likely to show desired effects for process rather than outcome measures, if desired behaviours are specific and easy to measure and are in areas where there is clear room for improvement.⁴

Both recent examples of incentives for postdischarge physician follow-up^1,2 are specific, in that premiums reward particular processes of care (physician visits) and the visits are easily measurable in administrative data. Before the start of these programs, one-third of patients had no physician follow-up within two weeks of discharge¹ and the probability of a psychiatric visit within 30 days was less than 0.3,² so there seemed to be ample room for improvement in follow-up. Physician follow-up seems like a reasonable target for incentives, and yet the findings for both studies are resoundingly null.

In interpreting the lack of effect, the authors of both papers suggest charitably that the design of such incentives must be more carefully examined, with consideration for principles of behavioural economics.^1,2 Certainly there is much to learn from decades of research in this field about how the design of incentives (size, choice of target, timing, framing of the incentive as a loss or a gain, and delivery of the payment itself) can improve their effectiveness.^7,8 However, efforts to improve the design of incentives for individual physicians will go only so far to solve a problem that is rooted in the behaviour of individual physicians.

Before spending too much more time and creativity developing tastier carrots (or pointier sticks), it may be beneficial to consider the degree to which postdischarge follow-up is within the control of individual physicians. The authors of the linked studies also acknowledged that barriers remain outside physicians’ control. ^1,2 Poor communication between inpatient and outpatient care can mean that the responsibility for sharing information about hospital admission and booking follow-up appointments falls to patients, which, in conjunction with varying patient mobility, health literacy, finances and social supports, may limit follow-up.¹ Where interventions have been found to reduce psychiatric readmissions, they include components such as patient education before and after discharge, structured needs assessment, formal transition managers and communication between inpatient and outpatient providers.⁹ To the extent that factors that determine follow-up are beyond the control of community-based physicians, financial incentives targeting physicians will not improve follow-up.

Of course, incentivizing behaviour change isn’t the only rationale for premium payments. Additional payments may appropriately reflect the time required to support patients as they transition back to the community after an admission and may reward providers who are already providing needed follow-up care. However, financial incentives are not the only way to provide appropriate compensation, and the false promise that small tweaks to payments for physicians will fix systemic problems can do real harm if it means we fail to pursue other opportunities for change.

It may be that incentives targeting groups or organizations rather than individuals are more powerful in changing patient care processes or increasing coordination.⁷ In other jurisdictions, where incentives to outpatient physicians have targeted follow-up care, they were aligned with incentives to prevent readmission at the hospital level, and with processes to facilitate communication between inpatient and outpatient providers.¹

Incentives may offer neat and tidy solutions to neat and tidy problems, as suggested by evidence showing that discrete incentives work in the short term.^3–6 However, the most intractable problems in Canadian health care — which certainly include managing transitions from hospital to community — are messy. They involve siloed care, poor transfer of information and ambiguous accountability structures. To achieve higher performance, quality or value, we need to take a systems approach to the design of payment systems, and accept that changes to payments alone are not likely to deliver change where it is most needed.

See related articles at www.cmaj.ca/lookup/doi/10.1503/cmaj.160816 and www.cmaj.ca/lookup/doi/10.1503/cmaj.170092