Measuring the relative performance of providers of a health service

A methodology is developed and applied to compare the performance of publicly fundedagencies providing treatment for alcohol abuse in Maine. The methodology estimates aWiener process that determines the duration of completed treatments, while allowing foragency differences in the effectiveness of treatment, standards for completion oftreatment, patient attrition, and the characteristics of patient populations. Notably, theWiener process model separately identifies agency fixed effects that describe differencesin the effectiveness of treatment ('treatment effects'), and effects that describe differencesin the unobservable characteristics of patients ('population effects'). The estimated modelenables hypothetical comparisons of how different agencies would treat the samepopulations. The policy experiment of transferring the treatment practices of more costeffective agencies suggests that Maine could have significantly reduced treatment costswithout compromising health outcomes by identifying and transferring best practices.