Throughout the modern history of medicine, health care has been paid for in a “fee-for-service” framework – physicians performed services and physicians were paid for those services accordingly. A common criticism is that fee-for-service incentivizes physicians towards overuse without improving outcomes. The “pay-for-performance” model addresses this issue by tying physician reimbursements to measures of quality. A recent study evaluates whether pay-for-performance models actually improve outcomes.
The research team began by performing a literature search of articles that included the keywords “value-based purchasing,” “pay for performance,” and “accountable care organizations.” They identified 80 articles published between 2010 and 2015, studying 44 incentive schemes in 10 countries. Using a vote count method, they analyzed articles to quantify the percentage of outcomes that improved under the schemes.
On average, 56% of outcome measures were positively impacted per scheme, both in the US and abroad. However, they found that studies with weaker-designs, which didn’t account for baseline differences or trends in samples (i.e. case-control and before-and-after studies), had 25 percentage point higher outcome improvements compared to more strongly-designed studies.
While the results are inconclusive, the study identified some notable ideas. Three schemes specified that incentive payments would not go to physicians but instead to facilities. Rewards could then be used for investments such as advancing clinical care capacities and hiring support staff. These schemes saw higher percentages of positive changes in outcomes compared to those that did not by 24 percentage points. While significant, only 3 schemes fell into this “specific use” category. Furthermore, schemes in which performance improvements were rewarded actually showed a lower percentage of outcome improvements compared to those that did not by an average of 20 percentage points.
The researchers note that increasing the size of incentive payments have little power to motivate changes in physicians’ behavior. While tentative, the study may indicate that directing rewards back to the facilities for investments in care has greater incentive power than income supplementation alone. Going forward, the future of constructing alternative payment schemes may lie in identifying fundamentally what incentives – monetary or otherwise – are capable of motivating physician behavior.
This Policy Prescriptions® review is written by Daniel Liaou as part of our collaboration with the Health Policy Journal Club at Baylor College of Medicine. Mr. Liaou is a second year medical student.
This article reviews the literature on the use of financial incentives to improve the provision of value-based health care. Eighty studies of 44 schemes from 10 countries were reviewed. The proportion of positive and statistically significant outcomes was close to .5. Stronger study designs were associated with a lower proportion of positive effects. There were no differences between studies conducted in the United States compared with other countries; between schemes that targeted hospitals or primary care; or between schemes combining pay for performance with rewards for reducing costs, relative to pay for performance schemes alone. Paying for performance improvement is less likely to be effective. Allowing payments to be used for specific purposes, such as quality improvement, had a higher likelihood of a positive effect, compared with using funding for physician income. Finally, the size of incentive payments relative to revenue was not associated with the proportion of positive outcomes. PMID: 27815451