Predictive model of 4-year cognitive decline risk in middle-aged adults in rural area of Xi'an
- VernacularTitle:西安地区农村人群4年认知功能下降风险预测
- Author:
Ling GAO
1
;
Yucheng PANG
1
;
Suhang SHANG
1
;
Liangjun DANG
1
;
Shan WEI
1
;
Jin WANG
1
;
Qiumin QU
1
;
Kang HUO
1
Author Information
- Publication Type:Journal Article
- Keywords: cognitive decline; predictive model; risk assessment; cohort study; rural population
- From: Journal of Xi'an Jiaotong University(Medical Sciences) 2025;46(5):811-817
- CountryChina
- Language:Chinese
- Abstract: Objective To develop a risk predictive model of cognitive decline in a prospective cohort study in rural area of Xi'an and compare the predictive performance with that of the classical CAIDE model.Methods The cohort was established between October 2014 and March 2015 in two selected villages in rural Xi'an.Mini-Mental State Examination(MMSE)was applied to assess global cognition at baseline and 4-year follow-up,and cognitive decline was defined as a drop of ≥4 points in MMSE after 4-year follow-up.Participants were randomly split into training set and validation set in a ratio of 7∶3.The Logistic regression analysis was used to develop the predictive model,and the area under the receiver operating characteristic(ROC)curve was applied to assess the performance of the risk model.Results Occurrence of cognitive decline after 4-year follow-up was 4.15％.Future cognitive decline was significantly predicted by age,low education and stroke(AUC in training set=0.73,95％CI:0.63-0.79;AUC in valid data=0.77,95％CI:0.67-0.87),while the classical CAIDE model did not predict the risk of cognitive decline well(AUC=0.68,95％CI:0.61-0.75).The results differed after stratification by APOE genotype,and showed a better predictive value of both our model(AUC=0.87,95％CI:0.78-0.96)and CAIDE model(AUC=0.89,95％CI:0.81-0.98)in APOE ε4 carriers.Conclusion The predictive model was developed based on age,educational level and stroke,and it predicted relatively well 4-year cognitive decline as compared with traditional CAIDE model,especially in APOE ε4 carriers.However,the model should be validated after longer follow-up and further improved to increase its predictive value.
