Population-based dementia prediction model using Korean public health examination data: A cohort study

Area of Science:

• Gerontology
• Public Health
• Epidemiology

Background:

• Dementia poses a significant global health challenge, necessitating early identification and prevention strategies.
• Existing dementia prediction models often lack specificity for middle-aged and Asian populations.
• There is a critical need for practical and precise dementia risk assessment tools tailored to diverse demographics.

Purpose of the Study:

• To develop and validate a dementia prediction model using a large-scale Korean health examination dataset.
• To identify key risk factors for dementia incidence in middle-aged adults.
• To assess the model's performance in predicting 10-year dementia incidence.

Main Methods:

• Utilized national Korean health examination data from 331,126 adults (40-69 years) from 2002-2003.
• Employed Cox proportional hazards models to construct prediction models for the total group and sex-specific subgroups.
• Validated model performance using receiver operating characteristics curves, C-statistics, and calibration plots.

Main Results:

• Achieved a C-statistic of 0.81 for the total group, indicating good discriminative accuracy.
• Identified significant dementia risk factors including age, female sex, underweight, hypertension, psychiatric/neurological disorders, cardiovascular disease history, diabetes, smoking, and lack of exercise.
• Risk factor significance varied slightly in sex-specific analyses, with low body weight and hypertension not significant in females.

Conclusions:

• National public health examination data can be effectively leveraged for dementia prediction in middle-aged Asian populations.
• The developed model facilitates early dementia identification and supports targeted public health interventions.
• This approach offers a practical tool for risk stratification and preventative care in clinical and public health settings.