Many believe that underprovision of mental health services will be one strategy managed care plans will adopt to discourage loser enrollees because of the correlation of mental health costs with costs of other services, but this contention has never been subject to empirical scrutiny. Are mental health services any more likely to be under-provided in managed care than treatments for other conditions? The proposed research would address this question by identifying the conditions whose treatments are most affected by selection incentives in managed care. We are particularly interested in mental health, but we consider the incentives to provide care for mental health in relation to the incentives for other services. This proposal introduces a new approach, based on economic theory, to derive a measure of plan strictness in rationing of services. We show how this measure can be used as a computationally straightforward index of incentives to over and under-provide services for various conditions. We intend to empirically implement this concept in this research, and based on the results, make concrete recommendations for policies of risk adjustment and """"""""carve outs"""""""" for dealing with selection-related incentives in markers for managed care plans. Service distortions generated by selection-related incentives affect the plan choices of enrollees. Carve out programs may attenuate such incentives by removing the mental health care from the competitive strategy of health plans. We test the empirical importance of this effect. Adverse selection is a enduring threat to the efficiency of provision of mental health care in insurance markets. Understanding how adverse selection works in managed care and what can be done about it should be a very high priority for research.
