Outcome models in clinical studies: implications for designing and evaluating trials in clinical nutrition.

The selection of appropriate outcome variables in clinical nutrition is particularly challenging, since nutrition is an adjunct therapy in most cases. Therefore, its effect may be confounded with the primary therapy, and classic biomedical outcomes may not reflect the effect of the nutritional intervention. This paper scrutinizes different alternatives to the biomedical perspective.Five different outcome models are proposed and analyzed for their suitability in clinical nutrition studies: biomedical, patient-centered/-reported, health economic, decision-making, and integration of classical and patient-reported endpoints. Most published studies in the field of clinical nutrition make use of biomedical endpoints, but the growing importance of patient-centered/-reported and health economic outcomes is recognized. We recommend avoiding to focus solely on biomedical endpoints in clinical nutrition studies. The availability and value of a broader set of outcome-models should be acknowledged.Patient-centered/-reported, health economic or combined endpoints are particularly useful to assess the effect of nutritional therapies, especially when applied in conjunction with a primary therapy. The proposed outcome models can also contribute to refine clinical nutrition guidelines in assessing the clinical relevance of the study results.