A geriatric medication algorithm designed to reduce inappropriate prescribing was tested in a resident outpatient clinic. The medications of patients over 65 years old taking more than three medications (n = 41) were compared pre- and post-algorithm using the paired t-test. Pre-algorithm, the average number of drugs was 5.8 per patient (SD 1.62). Fifteen medications (6.4%) were discontinued, seven were substituted for a less toxic medication, and five were added. Post-algorithm, the average number of drugs was 5.6 (SD 1.69), mean difference 0.3 (SD 0.67), p < 0.025. Drugs discontinued were more likely to be high risk compared with drugs used at baseline; drugs added were less likely to be high risk. In this pilot study, the authors conclude that the algorithm helps resident physicians reduce inappropriate prescribing.