OBJECTIVES To examine the incidence of Medicare Part D enrollees potentially affected by drug-drug interactions (DDI) and associated hospitalization, ER visits and death. METHODS A retrospective cohort study of 2006 Medicare Part D enrollees from six states (Alabama, California, Florida, Mississippi, New York, and Ohio) was conducted. Enrollees with Part D claims for either the object or precipitant (precipitant alters the effects/pharmacokinetics of object) drug identified to result in a drug-drug interaction were included. Enrollees were identified as having a DDI if the object and precipitant drug were taken during overlapping dates, with the first day of overlap being the DDI event date. Medicare Part A records for 2006 were evaluated to examine hospitalizations and ER visits within 30 days of an object drug precipitation of the potential DDI. The Medicare beneficiary file was used to examine death. RESULTS Data were available for 2,707,769 enrollees. Overall, 15.3% of enrollees were taking an object and precipitant drug, thus being at risk for a potential DDI. After controlling for age, gender, ethnicity, and comorbidities, the odds ratio for hospitalization and for ER visits when comparing enrollees with a DDI to enrollees taking an object drug without DDI potential were: 1.57 and 1.57 for warfarin DDI's; 1.28 and 1.48 for digoxin DDI's; 1.37 and 1.04 for beta-blocker DDI's; 1.27 and 1.09 for ACE/ARB inhibitor DDI's; 2.20 and 1.90 for methotrexate DDI's; 1.37 and 1.23 for clonidine DDI's. Significant outcomes were not found for pimozide, MAOI/SSRI, and thiopurine DDI's. There was no increase in risk of mortality. CONCLUSIONS Potential DDIs are still a significant problem among the elderly. Intervention strategies to reduce the incidence of potential DDI's should be a priority in CMS quality assurance activities.