A qualitative study of the influence of loss aversion and increased social pressure in physician pay-for-performance

Abstract

Background: Payers are increasingly using behavioral economic principles to design reimbursement programs, despite little evidence regarding the efficacy of these principles on influencing physicians' clinical behaviors. We conducted a qualitative follow-up study to a randomized controlled trial testing how loss aversion and increased social pressure in physician pay-for-performance (P4P) influenced the motivations and behaviors of participating physicians. Methods: We interviewed 22of the 33 physicians who participated in a previously reported randomized controlled trial using a standardized script. The trial included all physicians in an integrated physician network who treated patients with one of five chronic conditions (diabetes, asthma, COPD, CAD, or CHF). Seven participants were in the " active control" arm whose incentive was the same as the year prior; nine were in the loss aversion arm and had the option to withdraw up to 50% of their expected bonus early (averaging $ 19,292); and twelve were in the increased social pressure arm and had 50% of their bonus based on group performance, rather than 30%. Transcripts from these interviews were coded by two trained observers and analyzed using nodal analysis with NVivo 11. Results: The most frequently discussed topics were the financial salience of the interventions (3.6 references per interview), changes in personal motivations (2.3 references per interview), and changes in group dynamics (2.0 references per interview). Participants in the increased social pressure arm demonstrated significantly increased group cohesiveness and individual empowerment that were associated with changes in clinical behavior, including offering additional testing and arranging for closer follow-up. Participants in the loss aversion arm demonstrated little change in their clinical behavior. The ineffectiveness of this arm was attributed to a lack of participation in the voluntary incentive-only two of the eleven physicians withdrew any money-rather than a lack of salience of the principle of loss aversion to physicians necessarily. Barriers to success included clerical burdens of P4P and using outcomes measures rather than process measures for performance bonuses. Conclusions: Incorporation of increased social pressure into physician P4P incentive programs can lead to significant changes in physicians' personal motivations, group dynamics, and clinical behaviors. Lack of participation in pre-funded bonuses may indicate limitations to feasibility of using voluntary loss aversion in physician P4P programs.