Episode-Based Payment for the Medicare Outpatient Therapy Benefit

Abstract Objective To conduct an analysis of Medicare outpatient therapy episodes of care and associated payment implications. Design Retrospective observational design using Medicare claims data. To descriptively analyze the composition of outpatient therapy episodes, both variable- and fixed-length episodes are explored. The variable-length episode definition organizes services into episodes based on the time pattern of therapy service utilization, using 60-day clean periods. Fixed-length episodes are also examined, beginning with the first therapy utilization in calendar year 2010 and extending 30, 60, and 90 days. Setting The study is focused on community-dwelling users of outpatient therapy. Participants The sample includes all Medicare patients who used outpatient therapy beginning at any point in 2010. Interventions Not applicable. Main Outcome Measures Mean episode payments and episode lengths in calendar days. Results Variable-length outpatient therapy episodes have a mean payment of $881. On average, outpatient therapy episodes last 43 calendar days. Mean therapy durations for the 30-, 60-, and 90-day fixed-length episodes are 20, 31, and 38 calendar days, respectively. The 30-, 60-, and 90-day fixed-length initial episodes account for 40%, 55%, and 63%, respectively, of total Medicare payments. Simulations of episode-based payment illustrate the difficulty of avoiding a large number of substantial underpayments, because of the right-skewed distribution of total actual payments. Conclusions A strength of episode payment is that it reduces cost and potentially wasteful variation within episodes. Given the substantial variation in therapy episode expenditures, absent improvements in available data and in predictive information, a pure lump sum episode payment would result in substantial revenue changes for many episodes. Additional data are needed to better explain the wide variation in episode expenditures.