Online Program

Statistical and Trial Design Considerations for Kentucky's New Medicaid Program (307866)

*Elizabeth F Bair, University of Pennsylvania 
Erica Dixon, University of Pennsylvania 
Will Ferrell, University of Pennsylvania 
Kristin A Linn, University of Pennsylvania 
Kristen Underhill, Columbia University 
Atheendar S Venkataramani, University of Pennsylvania 
Kevin G Volpp, University of Pennsylvania 

Keywords: stepped-wedge, trial design, Medicaid

In January 2018 the Center for Medicare and Medicaid Services approved a Section 1115 waiver for the Commonwealth of Kentucky’s Medicaid program called Kentucky HEALTH. Kentucky HEALTH bundles together a collection of provisions, including a Community Engagement requirement, monthly premiums, and Medicaid lockouts. During the planning process, the state agreed to incorporate experimental design into the implementation of Kentucky HEALTH. We compare three possible trial designs: a two-arm trial randomizing individuals; a stepped-wedge trial randomizing individuals with cross-sectional measurements; and a stepped-wedge cluster randomized trial following cohorts of individuals. Important study design, statistical, and operational considerations for each trial are discussed. We also investigate power and sample size considerations leveraging survey data collected from Kentucky HEALTH eligible Medicaid beneficiaries in 2018 to inform realistic parameter settings.