A random national sample of 501 physicians and 298 nurse practitioners was presented a case vignette describing a patient with epigastric pain and endoscopy showing diffuse gastritis. Respondents were encouraged to request further information and then were asked for recommendations. History available if requested included substantial use of aspirin, coffee, cigarettes, and alcohol, and severe psychosocial stress. More than one third of the physicians chose to initiate therapy without seeking a relevant history. Nearly half of all physicians indicated that a prescription would be the single most effective therapy; 65% recommended a histamine antagonist. By contrast, only 19% of nurse practitioners opted to treat without taking further history; the nurse sample asked an average of 2.6 questions vs 1.6 for physicians; only 20% of the nurses recommended a prescription medication. These findings raise concerns about the adequacy of basic history taking in this setting and the underuse of nonpharmacologic approaches in favor of excessive reliance on prescription drugs, even when not indicated by clinical circumstances.