Online Program

Return to main conference page

Thursday, January 11
Thu, Jan 11, 2:00 PM - 3:45 PM
Crystal Ballroom F
Use of CMS Data and Tools to Help Researchers

Bayesian Framework for Health Care Innovation Award (HCIA) Evaluation of Complex/High-Risk Patient Targeting Awardees (304140)

*ADRIJO CHAKRABORTY, NORC at the University of Chicago 
Erin Ewald, NORC at the University of Chicago 
Sai Loganathan, NORC at the University of Chicago 
Edward Mulrow, NORC at the University of Chicago 
Shriram Parashuram, NORC at the University of Chicago 

Keywords: Health Care Innovation award, Bayesian methods, Evaluation

Bayesian methods generate results as a probability of program effectiveness given the observed data. This statistical approach has been gaining acceptance in both the scientific and public policy communities and is becoming a more attractive option, in part due to the availability of software that makes it easier for non-programmers to perform Bayesian analysis, as well as the intuitive interpretation of outcomes. The method is attractive to policymakers because it generates simple probability statements that are well suited for policy-related risk assessments. We explore Bayesian methods for studying the impact of care improvement efforts on Medicare spending for participants in five of the Centers for Medicare & Medicaid Services Health Care Innovation Award programs. We analyze and compare the results based on both classical econometric and Bayesian methods and discuss the usefulness of Bayesian analysis in this context. This approach may help policymakers make informed decisions using more intuitive and interpretable results.