Episodes of care are defined as the bundle of medical treatments used to treat an illness over a specified time period. Because all treatments are bundled together, these episodes have been thought to provide a superior unit of analysis in pay-for-performance (P4P) systems. In fact, Oxford Health Plan pioneered episode payment in the 1990s. [ Later, the health plan abandoned episode payment due to data and computer system difficulties ]
How has episode-based pay-for-performance worked in California? A paper by Robinson, Williams and Yanagihara (2009) looks at Integrated Healthcare Association ( IHA ) initiative. This association of health plans, hospital systems, and medical groups manages California’s P4P program. The P4P program gives large physician groups a P4P score based on all of its commercial HMO patients. Using the Thomson Reuters Medical Episode Grouper (MEG), $264 million was spent on the California P4P between 2003 and 2007. Although this is the largest P4P program in the nation, it amounts to less than 2% of California physician’s income.
Using medical groups rather than individual physicians has a number of advantages. First, a larger number of patients are attributed at the medical group level compared to the individual provider. Also, having multiple physicians within a single group treating a patient does not complicate the analysis. Despite the use of medical groups rather than physicians, Robinson and co-authors found a number of problems with episode-based P4P:
Small Sample Size: The IHA technical committee stated that a physician organization must have a minimum of 30 episodes to be evaluated. However, most physician organizations did not have enough episodes to be scored. The main problem is that most enrollees are healthy during the year and thus do not generate as many episodes. However, this may be less of a problem if P4P was to be applied to Medicare beneficiaries.
Data Completeness: “Prior to P4P, there was little incentive for the medical group to fully code what is done to the patient (procedures) and why it is done (diagnoses) on encounter forms.” If P4P were applied to the Medicare population, procedure codes would likely be coded more accurately (because physicians are mostly paid by the procedure), but the diagnoses fields would likely suffer from similar problems.
Today in 2009, IHA has mostly abandoned episode-based P4P. The metrics to be used going forward include: the percentage of prescriptions filled which are generic, the percentage of ambulatory surgery procedures that take place in freestanding centers compared to hospitals, ER visits per 1000 enrollees, non-maternity hospital visits per 1000 enrollees.
The authors note that episodes of care still may be appropriate for high-volume, high cost procedures in orthopedics and interventional cardiology, but California’s enthusiasms for episode-based P4P seems to be waning.