Objective. To determine the effect of a diagnosis of Alzheimer's disease or related dementias (ADRD), and the timing of first ADRD diagnosis, on Medicare expenditures at end of life.
Data Sources. Monthly Medicare payment data for the 5 years before death linked to the National Long-Term Care Survey (NLTCS) for decedents between 1996 and 2000 (N=4,899).
Data Extraction Methods. Medicare payment data for the 5 years before death were used to compare 5-year and 6-month intervals of expenditures (total and six subcategories of services) for persons with and without a diagnosis of ADRD during the last 5 years of life, controlling for age, gender, race, education, comorbidities, and nursing home status. Covariate matching was used.
Principal Findings. On average, ADRD diagnosis was not significantly associated with excess Medicare payments over the last 5 years of life. Regarding the timing of ADRD diagnosis, there were no significant 5-year total expenditure differences for persons diagnosed with dementia more than 1 year before death. Payment differences by 6-month intervals were highly sensitive to timing of ADRD diagnosis, with the highest differences occurring around the time of diagnosis. There were reduced, non-significant, or negative total payment differences after the initial diagnosis for those diagnosed at least 1 year before death. Only those diagnosed with ADRD in the last year of life had significantly higher Medicare payments during the last 12 months of life, primarily for acute care services.
Conclusions. ADRD has a smaller impact on total Medicare expenditures than previously reported in controlled studies. The significant differences occur primarily around the time of diagnosis. Although rates of dementia are increasing per se, our results suggest that long-term (1+ year) ADRD diagnoses do not contribute to greater total Medicare costs at the end of life.