To evaluate the association of institutional residence and chronic disability with end-of-life Medicare utilization in 1989 and 1999.
Retrospective study of nationally representative survey data linked to Medicare claims.
Medicare beneficiaries who completed the National Long-Term Care Survey in 1989 or 1999 and died within 1 year.
Medicare utilization [hospital, hospice, and skilled nursing facility (SNF)] in the last year of life and at death among the nondisabled, chronically disabled, and institutionalized.
Medicare utilization over the last year of life was similar in 1989 and 1999 for the proportion hospitalized in all groups (eg, 76% vs. 73% among the nondisabled); there were significant increases in SNF use (eg, 9% vs. 20% among the nondisabled). Comparisons at the time of death were somewhat different, with significant increases in hospice use in all groups (eg, 4% vs. 22% in the nondisabled); there was a significant decrease in the proportion dying in the hospital only for the nondisabled group (53% vs. 40%). Utilization was similar for the nondisabled and chronically disabled, but was lower in the institutionalized (eg, proportion hospitalized in the last year of life was 51% in the institutionalized, 73% in the nondisabled, and 77% in the disabled group in 1999).
Despite dramatic growth in hospice use, hospital and SNF utilization generally did not decline. Institutional residence, which is not usually included in analyses of Medicare claims, was strongly associated with utilization. This may affect the usefulness of claims-based efficiency indicators.
From the *Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health; and Departments of †Medicine and ‡Oncology, Johns Hopkins School of Medicine, Baltimore, Maryland.
Supported by an unrestricted educational grant from Pfizer and grant K07CA96783 (to S.M.D.) from the National Cancer Institute.
Reprints: Sydney M. Dy, MD, MSc, Room 609, 624 N Broadway, Baltimore, MD 21205. E-mail: email@example.com.