Abstract
BACKGROUND/OBJECTIVES: The Mini Nutritional Assessment Short-Form (MNA-SF) is the recommended screening tool for older persons. Data on interrater reliability in clinical routine are rare. Thus, we wanted to quantify the interrater reliability of the MNA-SF in hospital. SUBJECTS/METHODS: This observational cross-sectional study was undertaken retrospectively. The study population comprised 105 participants. Risk of malnutrition was measured twice with the routine MNA-SF performed by nurses (within 24 h after admission) and a dedicated dietician (one to three days after the first MNA-SF). The MNA-SF score was analyzed for interrater reliability between nurse and dietician. RESULTS: Participants' mean age was 82.4 (±7.1) years and 71 (68%) were women. The mean total MNA-SF score was 7.4 (±2.4) assessed by dietician and 7.8 (±2.3) assessed by nurse. The intra-class correlation coefficient between the total MNA-SF scores was 0.74 (0.61; 0.82), indicating moderate reliability. For the MNA-SF nutritional status, Cohens Kappa was 0.37 (p < 0.05) showing a fair agreement. CONCLUSION: Multiple misclassifications were observed between malnutrition and risk of malnutrition. Because mean scores were near the border between malnutrition and risk of malnutrition, we recommend to consider the total MNA-SF score in addition to the three risk groups to assess nutritional risk in geriatric hospital patients.