Construction of a nomogram prediction model for Alzheimer's disease among the elderly in community
10.19485/j.cnki.issn2096-5087.2025.09.003
- Author:
ZHANG Tao
;
LIN Junfen
;
GU Xue
;
XU Le
;
LI Fudong
;
WU Chen
- Publication Type:Journal Article
- Keywords:
Alzheimer's disease;
the elderly;
nomogram
- From:
Journal of Preventive Medicine
2025;37(9):875-880
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish a nomogram prediction model for Alzheimer's disease (AD) among the elderly in community, so as to provide the evidence for early screening and prevention of AD.
Methods:Based on the Zhejiang Healthy Aging Cohort Study, the elderly aged 60-90 years who completed the baseline survey were selected as the study subjects. Follow-up surveys were conducted from 2015 to 2016 and from 2019 to 2021. Sociodemographic characteristics, lifestyle factors, medical history, and waist circumference were collected through questionnaire surveys and physical examinations. Cognitive function was assessed using the Mini-Mental State Examination (MMSE), and a diagnosis of AD was made based on the Alzheimer's Disease Assessment Scale-Cognitive Subscale and medical history. The participants were randomly divided into training and validation sets at 8∶2 ratio. LASSO regression was used to screen for predictive factors. Multivariable logistic regression model was used to analyze predictive factors and construct a nomogram. The model was analyzed and evaluated using the receiver operating characteristic (ROC) curve and decision curve analysis (DCA).
Results:A total of 6 988 elderly were included at baseline, with a mean age of (68.19±6.63) years. There were 3 438 males (49.20%), and 3 550 females (50.80%). The median follow-up duration was 4.90 (interquartile range, 3.80) years, with 817 new cases of AD were identified, yielding an incidence of 11.69%. LASSO regression and multivariable logistic regression showed that age (OR=1.017, 95%CI: 1.005-1.030), gender (female, OR=1.820, 95%CI: 1.533-2.165), educational level (primary school, OR=0.813, 95%CI: 0.673-0.980), physical exercise (not active, OR=1.572, 95%CI: 1.260-1.980), dining companions (spouse and children, OR=0.771, 95%CI: 0.598-0.995), baseline MMSE score (OR=0.843, 95%CI: 0.821-0.866), and waist circumference (OR=0.981, 95%CI: 0.973-0.989) were risk predictors for AD among the elderly in community. The prediction model demonstrated an area under the ROC curve of 0.740 (95%CI: 0.698-0.783) in the validation set, with a sensitivity of 0.731 and a specificity of 0.667. DCA indicated that when the probability threshold was 0.060 to 0.325, the clinical net benefit was relatively high.
Conclusion:The AD risk prediction model constructed in this study has good discrimination and clinical practicability, can be used for early screening of AD among the elderly in the community.
- Full text:2025111308270338438社区老年人阿尔茨海默病列线图预测模型构建.pdf
