Effect of Financial Bonus Size, Loss Aversion, and Increased Social Pressure on Physician Pay-for-Performance

Key Points Question Does increasing bonus size or adding the behavioral economic principles of social pressure or loss aversion improve pay-for-performance effectiveness among physicians? Findings In this randomized clinical trial of 54 physicians and cohort study including 66 physicians and 8188 patients, increased bonus size was associated with improved quality relative to a comparison group, although adding increased social pressure and opportunities for loss aversion did not improve quality. Meaning Increasing pay-for-performance bonus sizes may be associated with improved effectiveness, whereas adding the behavioral economic principles of social pressure and loss aversion may not be.


SUPPLEMENTAL PRO FORMA for ENHANCED GROUP INCENTIVE*
The bar graph below shows the additional incentive dollars you can receive through group performance versus prior years.
-Blue Bar: In 2014, you earned $3,000 of your CI incentive from the PHO pool based on the Trinity PHO score of 79%.
-Red Bar: In the current 2016 year, with the new program design and if your group performs the same as 2014, you would earn $4,590 of your CI incentive based on your group performance.
-Green Bar: In the current 2016 year, with the new program design and if the group performance increases to 90%, you would earn $5,095 of your CI incentive based on your group performance. That means, in 2016 if your group performs at 90%, you could earn $2,095 more than you did in 2014 based on your group performance. "Group" refers to the performance of the physicians in Arm 3 Enhanced Group Incentive only.
The current Group (Arm 3) performance shows the following metrics that are hurting the Group

eMethods 1. Propensity Matching Methods and Graphs for the Area of Common Support
Propensity matching was performed in a two-step approach because not all physicians had historic trend data. In the first step, we used a logistic model with a dependent variable of participation in the Trinity PHO and independent variables of physician demographics, 2015 (pre-) composite quality score (on measures included in the study), and the trend from 2014-2015. This resulted in a match for 28 of the 33 physicians. The remaining 5 physicians were matched using a similar model without the 2014-2015 trend because these physicians did not have adequate historical data. In total, all 33 physicians in the RCT who received larger bonus sizes were matched to a physician in the no larger bonus size group in a 1:1 match using a 2 digit match.
The area of common support is shown below using kernel density. To test the trend in performance we ran the following linear regression clustering at the physician level and weighting by number of measures (when indicated): Where year is a continuous variable and trinity indicates whether the physician is in the Larger Bonus Size (LBS) group. 2 Physicians are included only if they are included in the main analysis.
This analysis demonstrated no significant differences in the trend in performance (Year x Trinity interaction term) in the years prior to the intervention.

eFigure 8. Sensitivity Analysis of Cohort Study Without Imputation (Using Complete Case Data)
The estimate is the effect of the association between larger bonus size and higher achievement of evidence-based quality measures. The error bars indicate 95% confidence intervals.

. Sensitivity Analysis of Cohort Study Without Physician Fixed Effects
The estimate is the effect of the association between larger bonus size and higher achievement of evidence-based quality measures. The error bars indicate 95% confidence intervals.

Stable Set of Physicians, Weighted
Year -0.007