Are selection bias assumptions always appropriate?
William J Miller, PhD, Patricia Ketsche, PhD, Susan A. McLaren, MPH, John Newman, PhD, and Andrew T. Sumner, ScD. Institute of Health Administration, Georgia State University, University Plaza, Atlanta, GA 30303, 404/651-4133, wmiller@gsu.edu Background: Selection bias, the assumption that healthier individuals are more likely to choose managed care plans because they do not have the longstanding ties to providers that sicker individuals have, is so widely accepted that actuaries routinely reduce capitation rates paid to managed care organizations (MCOs) participating in voluntary managed care programs. This assumption is based on the MCO model that typically offers a limited number of providers to a given population of patients. Methods: Utilization data from two populations of Georgia Medicaid recipients was analyzed. One which chose to join the primary care case management program and one which did not choose to join. Results: Utilization rates for the two populations varied dramati
