There is a large evidence base for nutritional intervention in acutely ill and post-operative hospitalised patients, but the evidence base for nursing home (NH) residents is small. The prevalence of poor nutrition in NHs is high and baseline nutrition appears to be an important determinant of response to nutritional intervention. Residents with mininutritional assessment (MNA) scores above 23.5 tend to show less response than those with lower scores. This relates in part to failure to increase intake in the better nourished as well as to actual response to increased intake. At the low end of the MNA spectrum, the increasing prevalence of multiple pathologies tends to result in a reduced response, but randomised controlled studies in this group is probably not ethical. Most studies have tended to investigate the intermediate group with MNA scores of 17–23.5 or equivalent using other scales. Interventions have usually resulted in increased intake of calories and micronutrients. Other end points have variously shown responses including weight, immunological measures, infection rates, decubitus ulcers, falls and fracture rates. Many studies have been too small to demonstrate benefit and some are likely to have suffered from type l errors – showing benefit by chance. Poorly quantifiable variables likely to be of importance include the local environment and catering as well as pathophysiological variability.