Validity of Mini Nutritional Assessment tool among an elderly population in Yeka sub-city, Addis Ababa, Ethiopia

Abstract

Background: The widely used nutritional assessment tool used for the elderly includes the Mini Nutritional Assessment (MNA) and Mini Nutritional Assessment–Short Form (MNA-SF) tool. These tools have not been evaluated for their validity and reliability among the elderly population of developing countries such as Ethiopia. This study aimed to evaluate the reliability and validity of the full and short form of the Mini Nutritional Assessment tool among the elderly in Ethiopia. Method: We evaluated the reliability and validity of the tools using a community-based cross-sectional study among 506 elderly individuals. Accuracy, sensitivity, specificity and cut-off point were evaluated to determine the validity of both the full MNA and MNA-SF tool. Reliability was assessed using Cronbach’s α coefficient. The criterion-related validity of the MNA tool was evaluated by computing the correlation between the total MNA score and anthropometric measurements. The Youden index was used to determine best cut-off points of the full MNA and MNA-SF. Result: The mean MNA score was 19.9 ± 4.5. Cronbach’s α value of the full MNA tool was 0.7. The overall accuracy of the full MNA was 91% (95% CI, 87.5%–94.9%). The sensitivity and specificity of the full MNA tool using an established cut-off point was 87.9% and 89.6% respectively. Youden index analysis showed that the best cut-off point to detect the malnourished and those at risk of malnutrition using the full MNA was 16 (sensitivity 90.4% and specificity 86.8%). The reliability of the MNA-SF as measured by Cronbach’s α was 0.5. The overall accuracy of the MNA-SF was found to be 93% (95% CI, 0.90–0.96). By using the Youden index the best cut-off point for MNA-SF to detect malnutrition was 7.5 (sensitivity 85.7% and specificity 89%). Conclusion: The full MNA tool was a valid and reliable tool to identify elderly individuals who are malnourished, at risk of malnutrition and well-nourished with modulation of cut-off points. However, the short MNA tool was valid and but not reliable in this study.