Although malnutrition is associated with poor clinical outcome, it cannot be inferred that better nutrition will improve clinical outcome. Efficacy of a proposed regimen is best established by prospective, randomised, controlled trials. Cost effectiveness is only an issue if efficacy exists. Patients with long term temporary, or permanent, inadequate bowel syndrome are candidates for parenteral nutrition. Most of the prospective, randomised, controlled trials testing the value of nutritional support in other diseases, however, have failed to show that this treatment has a beneficial clinical effect. Areas where these trials have shown a possible clinical benefit include the perioperative care of patients with upper gastrointestinal cancer, elemental diet treatment of Crohn's disease, and branched chain amino acid infusions in hepatic encephalopathy. Even in these instances, it is not clear that such treatment will prove to be cost effective (compared with other currently available treatments).