DISENROLLMENT OF MEDICARE CANCER-PATIENTS FROM HEALTH MAINTENANCE ORGANIZATIONS

OBJECTIVES. There is concern that financial incentives in health maint enance organizations (HMOs) might result in pressures to induce sicker members to disenroll. The authors compared disenrollment rates of Med icare HMO enrollees with cancer with disenrollment rates for cancer-fr ee enrollees, using Medicare enrollment files linked to population-bas ed tumor registry data from the Surveillance, Epidemiology, and End Re sults (SEER) Program. METHODS. The authors identified all aged Medicar e beneficiaries who enrolled in an HMO located in a SEER reporting are a during 1985 to 1989. Time to disenrollment was analyzed using a prop ortional hazards model. RESULTS. Overall, cancer patients were no more likely to disenroll than others. However, persons diagnosed with canc er after enrollment were less likely to disenroll. than other persons in Independent practice Association (IPA) model HMOs (relative risk [R R] = 0.79). Persons diagnosed with cancer after enrollment in group- a nd staff-model HMOs were about equally likely to disenroll as other pe rsons (RR = 0.91). Persons diagnosed with cancer less than 18 months b efore enrollment were at high risk for disenrollment from both IPA and group-/staff-model HMOs (RR = 1.47 and 1.35). There was substantial v ariation among HMOs in overall disenrollment rates and in RRs for dise nrollment by cancer patients. CONCLUSIONS. The low disenrollment rates of patients diagnosed after enrollment do not support the contention that features intrinsic to managed care make HMOs unattractive to the seriously ill. Monitoring of selective disenrollment could be used as a screen for possible access and quality problems.