Using A Reclassification Method To Evaluate The Improvement Of Mini Nutritional Assessment-Taiwan Version In Predicting The Mortality Of Older Taiwanese

• Main Authors: Chen,YaHui, 陳雅惠
• Other Authors: Wang, JiunYi
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/01304741951976334876

碩士 === 亞洲大學 === 健康產業管理學系健康管理組 === 100 === Background: Malnutrition is a common problem in geriatric health. The Mini Nutritional Assessment (MNA) has been widely applied in gerontological care. To normalize the MNA for Taiwan population, the MNA-Taiwan version 1 (MNA-T1) and version 2 (MNA-T2) have been developed. The validity of the two versions requires further investigation. Objectives: The study aimed to determine the improvement of the two versions in predicting the mortality risk of older Taiwanese. Methods: We analyzed the 1999 and 2003 data of the “Survey of Health and Living Status of the Elderly in Taiwan (SHLSET)”. A sample of 4,261 subjects with 53 years old or older was recruited. The Net Reclassification Improvement (NRI) and the Cox regression with Akaike's information criterion (AIC) were used to evaluate the improvement of the MNA-T1 and MNA-T2 in predicting the 4-years follow-up mortality risk. Results: The NRI were 0.004 (p = 0.726) and −0.010 (p = 0.454) if the MNA in the predictive model was replaced by the MNA-T1 and MNA-T2, respectively. The NRI was −0.012 (p = 0.052) if the MNA-T1 in the model was replaced by the MNA-T2. The model fitting index AIC were 8,987.3, 8,204.4, 8,218.1 and 8,214.2 for the Cox regression models with CV (control variables) only, CV + MNA, CV + MNA-T1 and CV + MNA-T2, respectively. It implied that the fitting of the model gets much better if any of the three MNA versions was included in the model. However, there were only small differences among the three versions. Conclusions: The result shows that the mortality-predictive ability of the MNA-T1 and the MNA-T2 did not significantly superior to the original MNA. Beacuse there is no “gold standard” for identifying subjects with malnutrition, we just take mortality as an external reference in this study. More studies may be required to confirm the validity of the modified versions.