OBJECTIVES:To estimate prevalence of potentially inappropriate medication prescribing (PIP) among elderly patients in two outpatient practices and to determine the factors associated with PIP.METHODS: We conducted a retrospective chart review in a general family medicine practice (FM) and a specialty geriatric practice (SG) for the six month time period of 1/1/2004-6/30/2004. Inclusion criteria were age ? 65 years old, being an established patient of the practice, and having at least one visit during the study period. A sample of fifty charts was randomly identified from each practice. An occurrence of PIP was defined as having chart documentation of a prescription for a medication on the 2003 revised Beers list of drugs to be avoided in the elderly. RESULTS: In all, 21% of patients had at least one occurrence of PIP. Of these, 24% of subjects had 2 or more occurrences. Differences in PIP prevalences in FM and GS (18% and 24% of subjects, respectively) were not statistically significant. Overall, the most frequent medications of concern were oral estrogens and fluoxetine. Multivariate analysis showed that higher number of medications (? 9 medications vs. <5 medications, OR, 0.11; 95% CI, 0.02-0.69) was significantly associated with lower odds of PIP occurrence. Age, race, gender, number of clinic visits, and number of diagnoses were not significantly associated with PIP. CONCLUSIONS: Our retrospective chart review indicated that PIP in elderly patients was common in both ambulatory settings, supporting results of prior pharmacy database analyses. The fact that no significant differences were found in PIP occurrence in the two practices and that general demographic patient variables were not significantly associated with occurrence of PIP suggest that inappropriate prescribing may be endemic to physician prescribing patterns. More research is needed to study physician prescribing patterns in elderly populations and intervention strategies designed to reduce PIP.