Validation of the Framingham Diabetes Risk Model Using Community-Based KoGES Data

Area of Science:

• Endocrinology
• Preventive Medicine
• Epidemiology

Background:

• The Framingham Diabetes Risk Model (FDRM) has limitations in predicting diabetes risk according to current clinical standards.
• Evaluating the FDRM's validity in the Korean population is crucial due to potential ethnic and clinical variations.

Purpose of the Study:

• Assess the original FDRM's performance in a Korean cohort.
• Develop and validate a modified FDRM incorporating updated predictors and clinical features.
• Enhance diabetes risk prediction accuracy for the Korean population.

Main Methods:

• Utilized a community-based Korean cohort (n=5,409) for an 8-year follow-up study.
• Modified the FDRM by redefining predictors like hypertension and incorporating additional clinical factors.
• Evaluated model performance using Area Under the Curve (AUC) for internal and external validation.

Main Results:

• The modified FDRM, including age, obesity, hypertension, lipid profiles, and glycemic markers, achieved an AUC of 0.82.
• An expanded model with additional clinical features showed a slightly higher AUC (0.83) but lacked sustained performance in validation.
• Both modified models demonstrated strong external validity with AUCs exceeding 0.82 in independent datasets.

Conclusions:

• The original FDRM provides acceptable diabetes prediction in Koreans, but significant improvement is achieved with redefined predictors.
• The modified FDRM demonstrates enhanced accuracy and validity for diabetes risk assessment in the Korean population.
• Further evaluation of the modified model's long-term validity and clinical utility is warranted.