Confusion in 91 hospitalized patients over 60 years of age who had sustained hip fracture and undergone subsequent surgical repair was examined in seven hospitals in five states. The relationship of nursing activities, patient, surgical, and environmental variables to the extent of confusion apparent in the sample was explored. Data were gathered by record review, patient testing, observation, and interview. Patient testing and interview were carried out on the first, third, and fifth postoperative days; the patient's caregiver on those days was also interviewed. The major dependent variables used in multiple regression analysis were memory (immediate and recent), observed behavior, and patient's self-report of mental clarity. Results showed that the most consistent predictors of postoperative confusion were the presence of confusion on admission to the hospital or confusion during the preoperative period. Greater postoperative mobility and freedom from urinary problems resulted in less confusion. Being male and of increased age were risk factors. The presence of timepieces and patients' use of television were associated with higher levels of mental clarity. Patients in private rooms tended to do better on memory testing than patients in rooms shared with others. Nursing personnel's major action in response to patient confusion was that of conscious attention to orientation, explanation, or reassurance. Use of restraints was relatively frequent.