The purpose of this study was to investigate the impact of Medicare Part D on prescription utilization, health services utilization, and health care expenditures in the general Medicare population – as well as Medicare sub-populations, including non-Hispanic blacks (NHBs), Hispanics, near poor individuals, and persons with higher disease burden. A retrospective analysis of Medicare beneficiaries (N=32,228) was conducted using the Medical Expenditure Panel Survey 2002 to 2009 data. Multivariable quantile regression was used to estimate the following outcomes, adjusting for socio-demographic characteristics: 1) number of prescription fills; 2) out-of-pocket (OOP) drug expenditures; 3) total drug expenditures; 4) OOP health care expenditures; 5) total health care expenditures; 6) number of hospitalizations; and 7) number of emergency department (ED) visits between the pre-Part D (2002-2005) and post-Part D (2006-2009) periods. All expenditures were inflation-adjusted to 2009 dollars. The average age of the study sample was 71.0 (SD=14.5). In the general Medicare population, Part D was associated with decreases in OOP drug expenditures (-25.7% to -33.6%; p<0.0001) and OOP health care expenditures (-22.1% to -24.3%; p<0.0001) as well as increases in the number of prescription fills (5.8% to 8.4%; p<0.0001) and total drug expenditures (75th percentile: 5.5%; 90th percentile: 10.2%; p<0.0001). Part D was not associated with changes in total health care expenditures in the general Medicare population. Changes in hospitalizations and ED visits were tested at the 90th percentile, and the results were not statistically significant. In sub-group analyses based on race/ethnicity, non-Hispanic whites (NHWs) experienced more significant reductions in OOP drug and/or health care expenditures when compared with NHBs and Hispanics. Near poor beneficiaries experienced larger reductions in OOP drug expenditures than beneficiaries with middle- to high-income, while Medicare beneficiaries with three or more conditions experienced more substantial reductions in OOP drug and OOP health expenditures after Part D was introduced, compared with those with fewer than three conditions. Part D resulted in increases in medication utilization and reductions in OOP drug and OOP health care expenditures among Medicare beneficiaries. Part D was not associated with differences in total health care spending. The effects of Part D were more pronounced in Medicare subgroups, including NHWs, near poor individuals, and patients with higher chronic disease burden. / text