?
Heterogeneous effect of incentive contracts in healthcare
Besstremyannaya G., Golovan S.
The paper analyzes heterogeneous hospital behavior in response to an incentive scheme with linear increase of the reward function for measured quality. The main contribution of the paper is to reassess a stylized fact in the literature, which asserts that pay-forperformance systems lead to greater improvements from providers with lower baseline quality. We suggest that this “fact” is the result of measurement error, and we develop a structural model and an estimation approach to recover the effects of the linear incentive scheme on quality. Compared to threshold-based incentive schemes that pay a flat bonus for meeting a quality target, where incentives operate mostly for providers below and near to the target, linear incentives may have large effects on providers with any initial level of quality. The predictions of our theoretical model, which are verified empirically using the longitudinal data for US Medicare hospitals involved in value-based purchasing, show a direct association between prior quality and quality improvement owing to the incentive scheme. We argue that the differences between true quality and its measurable proxy may cause regression towards the mean, which needs to be excluded in empirical estimations of the impact of the incentive scheme. Our theoretical model suggests that the effect of the inaccuracy in quality measurement can also be diminished by increasing the share of hospital funds at risk. An extension of the model considers fixed cost of hospital’s investment into quality improvement. The hospitals split into two groups, and the effect of incentive on quality increase in the group that chooses to bear the fixed cost is larger than in the group that refuses the burden of the fixed cost.