Improving diagnosis in outpatient care is a patient safety priority. This simulation study evaluated the process of diagnosis in the primary care setting. Investigators contrasted physicians' diagnostic accuracy conducting a primary care visit in their usual manner versus using a clinical decision support tool. Each visit employed a standardized patient (an actor reporting symptoms consistent with a given diagnosis) and the visits with and without decision support were matched for complexity. The tool improved diagnostic accuracy significantly: 68% of visits using decision support reached the correct diagnosis versus 59% of usual care visits. The duration of visits and number of subspecialty consultations did not change with or without decision support. Physician participants rated the usability of the decision support tool favorably overall. These data suggest that decision support can be feasibly integrated into primary care to improve diagnostic accuracy.