Strategies for evaluatingpublic sector managed care using administrative data

Public sector managed behavioral care is changing faster than researchers can evaluateits effects. This chapter presents a format for managed behavioral health evaluations which is less expensive and easier to implement than methods which require prospective data collection. A framework is presented for evaluating managed care interventions based on two criteria: (1) access to care, defined as first contact with a behavioral health clinician; and (2) changes in utilization patterns. These issues can be addressed using three pre-existing electronic administrative data sources: Medic-aid paid claims and enrollment rosters; state substance abuse datasets; and state or county mental health treatment datasets. Various linkage strategies are presented, as well as statistical methods that handle multiple repeated time elements, such as policy periods, changing insurance status, and episodes of care. Potential controls and comparisons are also discussed. Finally, the human subjects implications of these designs will be addressed.