A method for adjusting capitation payments to managed care plans using multivariate patterns of health and...

A multivariate procedure for identifying case-mix dimensions from discrete health variables is presented. Since the dimensions are generated only from health use data and not service use data, they can be used for adjust capitation rates to provide incentives to treat persons not currently well integrated in standard health care system (e.g., very ill persons, the uninsured) or to promote specific health outcomes. The procedure is illustrated with data from Social/Health Maintenance Organizations (S/HMO) since they provide both acute and long-term care (LTC) services. Thus, case-mix measures to adjust S/HMO reimbursements have to represent both medical conditions and the degree, and type, of functional impairment. From 31 health and functioning items, six case-mix dimensions, and scores for individuals on each, were calculated. The multivariate distribution of scores in S/HMO enrollees, and in Medicare eligible, comparison samples, were examined in each site to see how their health differed. S/HMO enrollees were healthier and less frail than persons remaining in the Medicare FFS system. Such differences are important in adjusting capitation rates to provide incentive to accept clients with complex health problems.