A nurse researcher asks whether a structured discharge-teaching program reduces medication errors compared with usual teaching. Which study design would most directly support an estimate of the program's effect?