A commitment-based intervention was evaluated for improvement of adherence to a 10-day antibiotic regimen. Experimental Ss made verbal and written commitments for adherence and completed tasks designed to increase their investment in a medication regimen. Control Ss performed similarly structured tasks unrelated to the medical regimen. Adherence, measured by unannounced pill counts, was significantly higher for experimental subjects than for control Ss. Self-reported adherence was significantly correlated with posttest self-efficacy but not with pretest self-efficacy. Adherence to a medical regimen may be improved by strategies conceptually based on the investment model of commitment, which provides a useful framework for further study of adherence.