Abstract
BACKGROUND: Despite the high burden of malnutrition in the country, there is a lack of a simple and valid tool to screen elders in Ethiopia. The Mini Nutritional Assessment (MNA) tool has been validated for comprehensive geriatric assessment to identify malnutrition in other countries. However, there is a lack of evidence on the potential validity and reliability of the tool for institutionalized elders in Ethiopia. This study was aimed at determining the validity and predictive performance of MNA tool for malnutrition among Ethiopian institutionalized elderly. METHODS: A facility-based survey was conducted on randomly selected 164 elders in geriatric centers to evaluate the validity, reliability, and predictive performance of full MNA against hemoglobin (Hgb) and ideal body weight (IBW) measured under standard procedures. The data was presented in ROC graphs, and reliability was evaluated with Cronbach alpha. The receiver-operating characteristic curve (ROC) analysis was used to assess the predictive performance of the tool. The area under the curve (AUC) with its 95% CI was reported. The Youden index, at maximum sensitivity and specificity, was used to obtain optimal cutoff points. RESULTS: The internal consistency of the tool was good (α = 0.80). The full MNA score can better predict Hgb (AUC = 0.845; 0.783-0.899) and percentage of IBW (AUC = 0.90; 0.842-0.941) at specified cutoff points. A full MNA can predict malnutrition or risk of malnutrition based on percentage IBW at a sensitivity and specificity of 97.3% and 72.2%, respectively. CONCLUSIONS: The full MNA has the potential to be a reliable and valid nutritional assessment tool for institutional elders.