The Effects of Physician-Level Pay-for-Performance in an Outpatient Setting
Posted: 29 Jun 2007
We exploit variation in the switch from salary to pay-for-performance (P4P) compensation for physicians at a large multi-site community-based health center in the United States to estimate the effect of a change in physician payment on physician productivity and patient health outcomes.
In 2004, the health center switched the majority of its physicians from a salary to a P4P schedule, which consisted of a substantially reduced base salary compared to its pre-P4P level and a productivity-based component. Depending on specialty, physicians were paid between $12 and $66 for each outpatient encounter and an additional $5 for each of a number of different counseling and screening services administered.
We study the effect of P4P on the number of patient encounters and the number of patients seen per physician. We also study whether P4P led physicians to modify their patient recruitment and retention efforts in order to attract and keep more care-intensive patients and to what extent any increase in patient encounters was achieved by reducing the amount of time spent per encounter rather than total hours worked by the physician. We also study whether the switch to P4P led physician to overprovide services (supplier-induced demand). We study whether there was a faster increase in the reported provision of quality procedures (counseling and screening services) relative to procedures not separately reimbursed.
We also study the extent to which patient health outcomes were measurably affected by P4P-induced changes in physician behaviors. Whether P4P will raise or reduce a patient's downstream utilization and thus cost depends on whether the quality procedure involves screening or counseling. As screening increases the likelihood of diagnosing chronic conditions such as HIV, depression, or diabetes, this type of quality procedure will tend to raise a patient's number and service intensity of future visits. Conversely, to the extent that counseling helps prevent sexually transmitted disease, unintended pregnancy, and periodontal disease, this type of quality procedure will tend to reduce future utilization. Similarly, early detection of some diseases, especially cancer, may reduce the total cost of treatment.
Finally, we estimate the effect of the P4P initiative on physicians' incomes and on the center's net revenues. The reduction of the base salary as well as the specialty-specific payment amounts for encounters and quality procedures were chosen, so that physicians who did not modify their productivity would earn the same as before the switch to P4P. Thus, we can use the switch to a piece rate to estimate an implied labor supply elasticity, as the marginal remuneration for an additional patient encounter or quality procedure increased from zero to single- or double-digit dollar amounts. The magnitude of the labor supply elasticity, in turn, determines whether P4P increased physicians' annual compensation. To the extent that P4P led to more prevention and early detection of chronic decrease, the increase in physician wages might be offset by a decrease in patients' treatment cost, leaving the net effect on the center's net revenues ambiguous.
Keywords: pay-for-performance, primary care, provider reimbursement
JEL Classification: J33
Suggested Citation: Suggested Citation