A blinded randomised controlled trial of prescribed oral sip-feed supplements compared with routine hospital practice was undertaken in acute admissions to a geriatric medicine department. Patients were eligible for inclusion if they were admitted from home, were not obese (BMI>75th percentile), had no swallowing difficulties and were not deemed to be in the terminal stage of illness. On admission they were stratified by nutritional status (BMI<5th, >5th to <25th, >25th to <75th percentile) and randomised. The intervention group received 120 ml oral sip-feed supplement prescribed three times per d in the medicine prescription chart (22·5 g protein, 2260 kJ (540 kcal) energy/d) distributed at medication rounds for the duration of hospital stay. The control group received routine hospital care. Outcomes were patient compliance with supplement, total energy intake and nursing staff views of the method. Patients were randomised to receive supplements (n 186 of total n 381). Half had full compliance and three-quarters at least moderate compliance. Total energy intake was significantly increased, on average, in the intervention group (P=0·001). The proportion of patients meeting estimated minimum energy requirements was significantly increased (P=0·023), but was still <50 % for the sample of patients in the intervention group. The present study suggests this method is acceptable to patients and staff and improves total energy intake. However, the amount prescribed did not ensure minimum energy requirements were met in all cases.