Background: The period preceding death is often characterized by increased utilization of medical resources. There is an ongoing debate on the reasons for and the appropriateness of increased utilization at the end of life.
Objectives: To study end-of-life practices in the PACE program and to assess the contribution of individual characteristics versus program site to the variation in use of services.
Research design: Retrospective analyses of utilization and health status data for the last 3 years before death. Multivariate regression techniques were used to estimate models predicting utilization of services and the percent of variation explained by individual characteristics and program sites in relation to time from death.
Subjects: The study comprised 2160 persons enrolled in 10 PACE sites who died before 2000.
Measures: Monthly utilization by service type (eg, hospital, nursing home, primary care physicians), socioeconomic, and health status data.
Results: Utilization of health services increases as early as 7 months before death, with the largest increase in the last month. The increase is dominated by hospital use. During the last month before death variation across program sites explains twice as much of the variation in service utilization as does variation in individual characteristics.
Conclusions: The variation in end-of-life practices across PACE program sites, which are not attributed to differences in individual characteristics, raises two important questions: what are the causes for these variations; and are these variations desirable? Further research is required to answer both questions.