The specific aim is to develop criteria for the appropriate use of parenteral nutritional support in the health care sector. These recommendations will be based on a meta-analysis of the evidence that parenteral nutritional support does more good than harm in 12 clinical conditions. The meta-analysis will be as rigorous in its attempts to protect the review from bias and error as we expect investigators to be in generating primary data in clinical trials. A working protocol will consist of: literature search, development of inclusion/exclusion criteria, listing of papers selected and rejected (and reasons), assessment of study quality by individuals who are blinded to authors and study results, duplicate data abstraction, listing of study characteristics, tests for homogeneity, statistical pooling, consideration of potential type I and type II errors, and consideration of possible publication bias. We will also examine evidence on safety, costs and cost-effectiveness. The product of this project will be put through the adjudication process of the American College of Physicians' Clinical Efficacy Assessment Project (CEAP) for external review of possible dissemination. This study will have considerable impact on prospective payment for patients who are treated with this technology.
