Identifying Predictive Risk Factors for Future Cognitive Impairment Among Chinese Older Adults: Longitudinal...

Area of Science:

• Gerontology
• Public Health
• Epidemiology

Background:

• Cognitive impairment poses a significant societal burden in China, driving the need for effective clinical prediction models.
• Existing models' accuracy across diverse socioeconomic groups and subpopulations remains unclear.

Purpose of the Study:

• Identify key health information domains for predicting cognitive impairment in Chinese older adults.
• Examine disparities in predictive accuracy across different demographic and socioeconomic subsets.

Main Methods:

• Utilized data from the Chinese Longitudinal Healthy Longevity Survey (CLHLS).
• Quantified predictive ability of various factors (demographics, ADLs, cognitive tests, etc.) and existing models using Area Under the Curve (AUC) via cross-validation.
• Analyzed prediction accuracy across general, male, female, rural, urban, educated, and uneducated older adults.

Main Results:

• Demographics (AUC 0.78), cognitive tests (AUC 0.72), and instrumental activities of daily living (ADLs) (AUC 0.71) were strongest predictors in the general population.
• Predictive models showed significantly higher accuracy for females and those with no formal education compared to males and educated individuals.
• Existing models and key risk factors demonstrated lower predictive power for male, urban-dwelling, and educated older adults.

Conclusions:

• Demographics, cognitive tests, and instrumental ADLs are crucial for predicting cognitive impairment in Chinese older adults.
• Significant disparities exist in prediction accuracy across gender, education, and potentially urban/rural divides.
• Further research and model refinement are essential to ensure equitable risk assessment for all socioeconomic groups in China.