We studied the natural course of cognitive state in 256 consecutive hip fracture patients who were admitted from an independent living situation. We employed a treatment programme that focused on preventing postoperative cognitive impairment. Cognitive function was assessed with the SPSMQ screening test. The incidence of postoperative cognitive impairment among those lucid on admission was 13%, which generally was reversed before discharge. Thirty-seven percent were cognitively impaired on admission; of those, 51% reached normal test scores while in hospital. Those who recovered within the first week had as good a prognosis during the first year as those who remained lucid throughout the hospital stay. Cognitive impairment was associated with an increased complication rate, e.g. a three-fold increase of early fracture displacement and a four-fold increase of wound infection. This increased risk was present even in patients with mild/moderate cognitive impairment and could not entirely be explained by age. Our results suggest that it is possible to decrease postoperative cognitive impairment by routine monitoring of cognitive status, a high level of continuity and a reorientation programme. The routine assessment of the cognitive function is recommended in geriatric patients who are admitted for surgery.