AIMS: In the elderly, mortality prediction models are important for clinical decision-making and planning of services required. METHODS: A total of 247 Kahrizak Charity Foundation (KCF) residents aged ≥ 65 years were followed up for approximately 39 months. At the baseline, the questionnaires of Barthel Index (BI), Mini-Mental State Examination, Geriatric Depression Scale, Mini Nutritional Assessment and Norton Index was given. Medical history was recorded and anthropometric values were also measured at the baseline. Fasting blood samples were collected at baseline. Mortality and its causes were recorded during the study. RESULTS: A total of 30% (74) of participants died during the study. The variables that had a significant association with mortality in the Cox regression hazard model were entered into the principal components analysis (PCA). The older people's mortality index (OPMI) was developed by four variables extracted from PCA, including BI, age, hemoglobin and mid-arm circumference. Cut-points of these components were calculated using ROC curve analysis. Based on neural network analysis, there was no significant difference in relative importance of OPMI components. OPMI was correlated to mortality (r = -.351, P = 0.000) and survival (r = -0.355, P = 0.000). OPMI score was defined as the number of adverse predictors present. An increasing hazard ratio for mortality was observed from scores 1 to 4 (2.10, 4.56, 7.12 and 13.85, respectively). CONCLUSION: Our suggested model could predict mortality in KCF residents well. The new developed model could be a practical, easy and non-expensive index for mortality prediction in elderly care facilities.
AIMS: In the elderly, mortality prediction models are important for clinical decision-making and planning of services required. METHODS: A total of 247 Kahrizak Charity Foundation (KCF) residents aged ≥ 65 years were followed up for approximately 39 months. At the baseline, the questionnaires of Barthel Index (BI), Mini-Mental State Examination, Geriatric Depression Scale, Mini Nutritional Assessment and Norton Index was given. Medical history was recorded and anthropometric values were also measured at the baseline. Fasting blood samples were collected at baseline. Mortality and its causes were recorded during the study. RESULTS: A total of 30% (74) of participants died during the study. The variables that had a significant association with mortality in the Cox regression hazard model were entered into the principal components analysis (PCA). The older people's mortality index (OPMI) was developed by four variables extracted from PCA, including BI, age, hemoglobin and mid-arm circumference. Cut-points of these components were calculated using ROC curve analysis. Based on neural network analysis, there was no significant difference in relative importance of OPMI components. OPMI was correlated to mortality (r = -.351, P = 0.000) and survival (r = -0.355, P = 0.000). OPMI score was defined as the number of adverse predictors present. An increasing hazard ratio for mortality was observed from scores 1 to 4 (2.10, 4.56, 7.12 and 13.85, respectively). CONCLUSION: Our suggested model could predict mortality in KCF residents well. The new developed model could be a practical, easy and non-expensive index for mortality prediction in elderly care facilities.
Authors: Y K Chang; L F Huang; S J Shin; K D Lin; K Chong; F S Yen; H Y Chang; S Y Chuang; T J Hsieh; C A Hsiung; C C Hsu Journal: Sci Rep Date: 2017-10-04 Impact factor: 4.379