Validation of mini nutritional assessment long-form tool for screening malnutrition among elderly population, Meki town, Ethiopia: a cross-section validation study

Abstract

Background: Among various malnutrition screening tools, the mini nutritional assessment (MNA) long-form is a simple and valid tool used to assess nutritional status in the elderly population. It is easily administered within 10 to 15 minutes and does not require laboratory investigation. Although it has been validated and used worldwide, not yet validated in Ethiopia. Objective: To validate the MNA long-form for screening malnutrition among the elderly population aged 60 years and above living in the community. Method: Community based cross-sectional validation study was conducted among 176 elders from March 30 to April 30, 2020, using a simple random sampling technique in Meki town, East Ethiopia. All apparently healthy elderly people aged 60 years and above who have been living in Meki town for 6 months and above were included. Whereas amputated, bedridden, edematous, those with visible of deformity, and known liver and/or kidney disease were excluded. Data were collected using pretested Afan Oromo and Amharic version MNA long-form questionnaires by trained 2 BSc nurses. A fasted venous blood sample was collected by trained 2 BSc laboratory technicians for serum albumin concentration analysis using the Bromocresol green (BCG) method. Lastly, collected data were edited, entered into Epidata version 3.1 then exported to SPSS version 25 for Analysis. Mean. Standard deviation (SD), frequency, and percent (%) were calculated for the variable of interest. P-value< 0.05 was used to declare significance. Criterion-related validity and its subtype concurrent validity were assessed. Sensitivity, specificity, internal consistency, total diagnostic accuracy, positive predictive value (PPV), and negative predictive value (NPV) were calculated. The receiver-operating characteristic curve (ROC-curve) was done to calculate the area under the curve (AUC) and a new optimal cut off value with its optimal sensitivity and specificity. II Result: Out of 176 subjects, 44.3% were males. The mean (SD) age of 67.6 (5.79) years. The prevalence of malnutrition was 18.2% using MNA long-form (<17 points) and 13.1% using Serum albumin concentration level (<3g/dl). MNA long-form had strong internal consistency with Cronbach’s alpha 0.605 and a significant correlation was found with all items (Spearman's rho (rs)>0.242; P<0.05). MNA long-form showed that criterion-related validity was significant with serum albumin concentration (Spearman's rho (rs) =0.746; p <0.05). Also, concurrent validity was significant with the selfperception of nutritional status (Spearman's rho (rs) =0.514; p<0.05). Additionally, it showed moderate agreement with serum albumin concentration (Weighted kappa (Kw) =0.556; p<0.05). According to original cut off value, MNA long-form had a sensitivity of 93.5%, a specificity of 44.6%, PPV of 65.4%, and NPV of 86.0%. However, using a new cut off value <20.5, to identify markers of malnutrition with serum albumin concentration used as a golden standard, found a sensitivity of 97.6%, and a specificity of 82.8%. The AUC (95% CI) showed an excellent overall accuracy of 92.7% (88.5, 96.9). Conclusion and recommendation: MNA long-form tool, can be used as a valid screening tool to assess nutritional status among the elderly population in Ethiopia with modification of the original established cut off value. A future study using a combination of various biomarkers as a golden standard to validate the MNA long-form is needed