Although mental incapacity is becoming increasingly important in clinical practice, little information is available on its frequency in medical inpatients. We aimed to estimate the prevalence of mental incapacity in acutely admitted medical inpatients; to determine the frequency that medical teams recognised patients who did not have mental capacity; and to identify factors associated with mental incapacity. Over an 18-month period, we recruited 302 consecutive acute medical inpatients. Participants were assessed with the MacArthur competence tool for treatment and by clinical interview. Cognitive impairment was measured by the mini-mental state examination. 72 (24%) patients were severely cognitively impaired, unconscious, or unable to express a choice and were automatically assigned to the incapacity group. 71 (24%) refused to participate or could not speak English. Thus, 159 patients were interviewed. Of these, 31% (95% CI 24-38) were judged not to have mental capacity. For the total sample (n=302), we estimated that at least 40% did not have mental capacity. Clinical teams rarely identified patients who did not have mental capacity: of 50 patients interviewed, 12 (24%) were rated as lacking capacity. Factors associated with mental incapacity were increasing age and cognitive impairment. Mental incapacity is common in acutely ill medical inpatients, and clinicians tend not to recognise it. Screening methods for cognitive impairment could be useful in detecting those with doubtful capacity to consent.