AIM: To describe the burden of malnutrition in an at-risk population and validate a screening tool.
METHODS: A 10-year cross-sectional study was conducted in a remote hospital. Dietitians screened participants using the Malnutrition Screening Tool or the Adult Nutrition Tool. Participants scoring ≥2 utilising the Malnutrition Screening Tool or ≥4 utilising the Adult Nutrition Tool, or already under the care of a dietitian, were assessed for malnutrition using a Subjective Global Assessment. The validity of the Adult Nutrition Tool was compared to the Malnutrition Screening Tool in participants undergoing a Subjective Global Assessment.
RESULTS: Of 980 study participants, 70.1% identified as Aboriginal Australian and 42.6% were at risk of malnutrition. Three-hundred and seventy-four (58.8%) participants were malnourished. The Adult Nutrition Tool showed superior area under the curve (0.74, 95% confidence interval [0.67, 0.81], p < 0.001) than the Malnutrition Screening Tool (0.64, 95% confidence interval [0.56-0.73], p 0.001). An Adult Nutrition Tool score ≥2 demonstrated superior sensitivity (98.2%) and specificity (16.7%) compared to a Malnutrition Screening Tool score ≥2 (sensitivity 97.2% and specificity 14.3%). An Adult Nutrition Tool score of ≥3 demonstrated reduced sensitivity (92.0%) but superior specificity (20.2%) than screening scores ≥2 for both tools.
CONCLUSION: Malnutrition risk is high in a remote Australian hospital. This study confirms the Adult Nutrition Tool and the Malnutrition Screening Tool are valid tools to predict malnutrition in a unique remote hospital, recommending an 'at-risk' threshold of ≥2 for the Adult Nutrition Tool to aid in the early detection of malnutrition.