Dynamic aspects of prescription drug use in an elderly population

Abstract

Objective: This study explores longitudinal patterns in outpatient prescription drug use in an elderly population.

Data sources/study setting: Enrollment records and prescription drug claims were obtained for a sample of elderly Pennsylvanians (N = 27,301) who had enrolled in the Pharmaceutical Assistance Contract for the Elderly (PACE) program at any time between July 1984 and June 1987. Study Design. The study tracks monthly prescription fill rates for sampled PACE beneficiaries from their initial enrollment month through disenrollment, death, or the end of the study (whichever occurred first). We specify two-part multivariate models to assess the effect of calendar time, length of time in the PACE program, and progression to disenrollment or death both on the probability of any prescription use and on the level of use among those who filled at least one prescription claim per month. Control variables include age, gender, race, income, residence, and marital status.

Data collection/extraction methods: Data were extracted from administrative files maintained by the PACE program, checked for errors, and then formatted as person-month records.

Principal findings/conclusions: We find a strong positive relationship between drug use and the length of time persons are PACE-enrolled. Persons whose death occurs within a year have much higher prescription utilization rates than do persons whose death is at least a year away, and the differential increases as death nears. Persons who fail to renew PACE coverage use significantly fewer prescription drugs in the year prior to disenrollment. Holding age and other factors constant, we find that average levels of prescription use actually declined over the study period.