OBJECTIVES: To examine determinants of drug therapy problems (DTPs) among home health patients receiving telephonic medication therapy management (MTM). METHODS: Retrospective data were used from patients (n=418) who participated in a randomized, controlled trial of telephonic MTM in a home health population. Data were extracted from two databases: 1) home health records from in-home nursing assessments (OASIS-C), and 2) MTM records from pharmacists’ telephonic assessments. The primary dependent variable was the number of DTPs identified by the pharmacist during the initial MTM encounter. Secondary dependent variables were the presence of subgroups of DTPs, including safety, effectiveness, indication, and adherence. Univariate and multivariate linear and logistic regression models assessed the relationship between DTPs and the independent variables: age, sex, race, functional status (cognitive, vision, hearing impairment and limited verbal understanding), hospitalization risk score, number of medications, Beers medications, narrow therapeutic index medications, self-reported adherence, and history of impaired medication management. RESULTS: Hospitalization risk score, Beers medications, hearing impairment, vision impairment, number of medications, and history of impaired medication management were significant in univariate analysis (p<0.2) and entered into a multivariate linear regression model for the primary outcome. The overall regression model was significant (p = 0.009, R-squared = 0.029), and the number of Beers medications was found to be an independent predictor of number of DTPs (p = 0.008). Univariate variables significantly associated with indication-related problems included race, vision impairment, and living alone, but only race remained significant (p = 0.013) in a regression model (p=0.007, R-squared = 0.055). For adherence-related problems, race, Beers medications, and ethnicity were significant by univariate analysis, but only race (p=0.002) and Beers medications (p=0.048) remained significant in the regression model (p = 0.002, R-squared = 0.074). CONCLUSIONS: Through identification of predictors of DTPs, these results may inform allocation of MTM resources among this population.