ASSESSING QUALITY AND EQUITY IN HEALTH CARE
Department of Health Care Policy, Harvard Medical School
Department of Biostatistics, Harvard T.H. Chan School of Public Health
The last two decades have been characterized by an increasing focus on healthcare provider performance measures, most often utilizing multiple binary response outcomes. In this problem, data arise from multiple clusters where (a) outcomes within clusters are more similar than outcomes between clusters; (b) within-cluster covariates vary across clusters; (c) clusters are observed repeatedly over time; and (d) multiple binary measures are observed for each unit within the cluster. In this talk, we describe methods to determine whether quality of care in schizophrenia care varies by race/ethnicity and over time; and (b) whether these patterns differ across counties within states using Medicaid claims data from California, Florida, New York, and North Carolina during 2002–2008. Random effects approaches for handling within-county correlation and item response theory models for handling multiple binary outcomes per beneficiary are used to determine if where you live matters.
Thanks: Marcela Horvitz-Lennon (Rand Corporation), Rita Volya (Harvard Medical School), Rachel Garfield (Kaiser Family Foundation), Julie Donohue and Judith Lave (University of Pittsburgh Graduate School of Public Health). This research was supported by R01MH087488 from the National Institute of Mental Health.