Abstract

Background: Nutritional care for hospital in-patients is potentially important but challenging.

Objective: To investigate the association between nutritional status and clinical outcomes.

Methods: Eight prevalence surveys were performed at Haukeland University Hospital, Norway, during 2008-2009. In total 3279 patients were classified as being at nutritional risk or not according to the Nutritional Risk Screening (NRS 2002) tool. The initial four questions of NRS 2002 assess dietary intake, weight loss, body mass index (BMI) and illness severity.

Results: The overall prevalence of nutritional risk was 29%. Adjusted mean days for hospitalisation was 8.3 days for patients at nutritional risk and 5.0 days for patients not at risk (p < 0.001). In adjusted models, patients at nutritional risk had increased one-year mortality (OR 4.07, 95% CI 2.90-5.70), morbidity (OR 1.59, 95% CI 1.18-2.13), and were 1.24 (95% CI 1.16-1.32) times more likely to have had a new admission during the three previous years and the one subsequent year, compared to patients not at risk. A 'positive' response to the initial four questions was associated with increased risk of morbidity and mortality. Patients with a reduced dietary intake during the last weeks had OR 1.72 (95% CI 1.03-2.85) for one-year mortality. Patients with a positive answer on all the initial four questions had ten times increased risk for mortality the following year, OR 13.0 (95% CI 4.52-37.6).

Conclusion: The four initial questions of the NRS 2002 robustly identify nutritional risk and were strong predictors of hospitalisation, morbidity and most importantly mortality among hospitalised patients. Thus, these simpler and short questions are robust indicators for subsequent poor outcomes.