Many recently hospitalized patients lack knowledge about important aspects of their medications. We evaluated whether a structured discharge interview could improve medication knowledge.

Patients discharged with at least one discharge medication were recruited from two general internal medicine services (one experimental and one control) of a teaching hospital. During a 3-month baseline period, usual care at discharge was provided in both services. During the ensuing 3-month period, observation was continued in the control service; residents in the experimental service implemented the intervention, which consisted of a structured patient-centered discharge interview during which a computer-generated individualized treatment card was discussed with and provided to patients. One week after discharge, patients' knowledge about their medications was assessed by telephone.

We enrolled a total of 809 patients. After adjustment for patients' characteristics and for the effect of time, the intervention significantly increased the percentage of medications for which patients correctly knew the purposes (adjusted difference = 6%; 95% confidence interval [CI]: 3% to 8%; P <0.001), possible side effects (adjusted difference = 19%; 95% CI: 9% to 29%; P <0.001), and precautions to observe (adjusted difference = 9%; 95% CI: 2% to 19%; P <0.001). However, the number of medications that patients discontinued after discharge was not modified. Patients with a better knowledge of side effects of their active treatment were less likely to discontinue their medications, but there were no associations with other types of knowledge.

A structured patient-centered discharge interview, performed by residents using a standardized treatment card, significantly increased patients' knowledge about their medications. Its effects on compliance require further study.