Interrater reliability of nutritional risk screening using NRS-2002 in hospitalized patients with cancer

PURPOSE: Malnutrition is common among hospitalized cancer patients and has been linked to adverse clinical outcomes. Accurate nutritional risk screening is essential for rapid nutritional intervention. Although the NRS-2002 tool is widely used in clinical practice, its interrater reliability in oncologic settings has not been adequately examined. This study aimed to evaluate the interrater reliability of nutritional risk screening among hospitalized patients with cancer using the Nutritional Risk Screening 2002 (NRS-2002) tool. DESIGN: Retrospective cross-sectional study. METHODS: Thirty-five trained clinical nurses and one registered dietitian independently performed NRS-2002 assessments within 24 h of admission among 434 hospitalized cancer patients. Cohen's kappa statistics were used to evaluate interrater reliability for each NRS-2002 component, including nutritional status, disease severity, and risk classification derived from the total score. RESULTS: Nurses identified fewer patients at nutritional risk than the dietitian (73 vs. 211). In addition, the overall agreement between the registered dietitian and clinical nurses using the NRS-2002 was fair (kappa = 0.343, 95% CI: 0.275 to 0.411). Interrater reliability was fair for nutritional status (kappa = 0.241) but poor for disease severity (kappa = 0.198), whereas perfect agreement was observed for the age component (kappa = 1.000). Among cancer types, patients with breast cancer (n = 11) showed complete agreement (kappa = 1.000), though this finding should be interpreted cautiously because of the small sample size, followed by those with gastrointestinal tumors (kappa = 0.489, 95% CI: 0.289 to 0.689) and esophageal cancer (kappa = 0.408, 95% CI: 0.212 to 0.604). Nurses with junior professional titles (kappa = 0.347) showed slightly higher consistency with the dietitian than those with intermediate titles (kappa = 0.339). CONCLUSION: Overall, this study demonstrated fair interrater reliability of the NRS-2002 for assessing nutritional risk among hospitalized patients with cancer. To improve the reliability of nutritional risk screening, we recommend standardizing the operational definitions of disease severity scores, implementing structured and periodic training programs for clinical nurses, and establishing regular audit and supervision mechanisms. These measures are essential for improving screening consistency and, in turn, patient assessment and subsequent care.