Physician knowledge, financial incentives and treatment decisions for depression

BACKGROUND: Two important policy levers to affect health care delivery are financing and informational interventions. Unfortunately, these two approaches have not been considered simultaneously and little is known about how their effects compare. AIMS OF THE STUDY: This paper estimates the relative role of financial incentives (prepaid versus fee for service) and provider information (perceived knowledge of antidepressant medications and skill in counseling for depression) on quality of care for less and more severely depressed patients and their health and cost outcomes. METHODS: We develop a theoretical model of provider behavior and estimate a reduced form using a multinomial probit model with heteroskedastic covariances. The likely effects of changing provider knowledge about depression treatment in primary care are then simulated and contrasted with the effects of a shift toward prepaid managed care as opposed to fee-for-service care. The empirical model is estimated using data from the Medical Outcomes Study. RESULTS: We conclude that financing and information have different effects and that their combination can achieve the conflicting goals of improved health outcomes and reduced direct treatment goals. Moreover, including family income as one important dimension of social cost suggests that the combination of informational interventions and a shift to prepaid care may dominate either one intervention in isolation from a social cost perspective. Specifically regarding information, we found that increasing provider knowledge could have the highly desirable effect of greater targeting of treatments to sicker patients while not raising overall treatment rates much - a treatment pattern that many hoped managed care could achieve, but for which there has been little evidence. CONCLUSIONS: Our analysis illustrates the value of considering these widely different policy goals simultaneously. We learned that variation in physician knowledge generally had stronger associations with clinically relevant practice patterns for depression than did a complete change in financing strategy. The moderate change in perceived knowledge we simulated (not near the extremes of observed values of perceived knowledge) was associated with enough improvement in appropriateness of care to more than offset the reduction in appropriateness with a complete shift from fee-for-service to prepaid managed care. IMPLICATIONS FOR HEALTH POLICY: The paper demonstrates the importance of considering different interventions simultaneously. Combining informational and financial interventions simultaneously can achieve better quality of care and reduce health care costs, something neither intervention can in isolation.