Investigators: Edward Norton, Justin B. Dimick, Mousumi Banerjee, Scott E. Regenbogen, Donald Likosky
Funding (subcontract): National Institute on Aging, 2012-2016 (2 P01 AG 019783 11)
Aiming to reduce variation and reduce expenditures around hospitalizations, the Center for Medicare and Medicaid Services is moving steadily toward episode-based bundled payments. The hope is that bundled payments will give hospital and physicians more ?skin in the game? and thus stronger incentives to coordinate care and reduce the use of unnecessary services. Whether such reforms will be successful in reducing hospital episode costs remains uncertain however. Mechanisms underlying variation in hospital efficiency?and the extent to which they are likely to be leveraged by CMS? financial incentives?have not been well established. For example, practice patterns around hospitalization episodes in Medicare patients may be influenced by market forces originating in the under-65 population, including price differentials between public and private payers for specific services and related substitution effects. In addition, variation in hospital episode payments may be driven by aspects of hospital quality not immediately influenced by bundled payments.
Extending our previous P01-funded research, this project will explore determinants of variation in hospital efficiency, incorporating both claims data from the under-65 population and clinically rich outcomes and safety culture information from 50 hospitals in Michigan. We will first assess spill-over and substitution effects from the private market on episode efficiency in Medicare patients. We will then examine relationships between episode payments and various domains of hospital quality, including technical quality (e.g., risk-adjusted morbidity and mortality and processes of care) and hospital safety culture (e.g., teamwork and communication). Finally, we will evaluate the direct effects of the new CMMI bundled payment initiative on costs and quality, as well as its spillover effects on related but untargeted conditions and on the under-65 population.