Prescription Drug Spending for Medicare+Choice Beneficiaries in the Last Year of Life
Published in: Journal of Palliative Medicine, v. 9, no. 4, Aug. 2006, p. 884-893
Posted on RAND.org on December 31, 2005
BACKGROUND: In 2006, Medicare implemented its prescription benefit plan. Therefore, insights into medication costs at the end of life may help guide clinicians to navigate Medicare Part D coverage for chronically ill individuals. OBJECTIVES: The authors examined drug spending by disease and demographics for Medicare+Choice (M+C) beneficiaries in the last year of life (LYOL). RESEARCH DESIGN: Retrospective review of M+C decedents' drug claims and enrollment data collected between January 1998 and December 2000, supplemented by the Medicare denominator file and 1990 Census data. SUBJECTS: Four thousand six hundred two beneficiaries in a large national managed care organization. MEASURES: The authors analyzed the relationship between prescription drug expenditures and sociodemographic descriptors, insurance characteristics, and cause of death. RESULTS: The mean annual number of prescriptions filled was 36.9; the managed care organization (MCO) paid $539 and beneficiaries paid $627. Higher expenditures were significantly correlated with female gender, higher number of comorbidities, and whether beneficiaries obtained the insurance as an employer-based retiree benefit. Minority beneficiaries had 26% fewer prescriptions. Increasing levels of annual median household income corresponded with a 20% increase in the number of prescriptions and a 25% increase in mean out-of-pocket expenses, between those with a median household income of less than $20,000 and those with $40,000 or greater. In the LYOL, chronic obstructive pulmonary disease and diabetes had the highest average number of prescriptions and total expenditures. Individuals dying from strokes or other unclassifiable conditions had the lowest average number of prescriptions and average total expenditures. CONCLUSION: Medication expenditures in the LYOL were highly dependent upon selected sociodemographic, insurance characteristics, and disease states.