Screening for Diabetes Risk Using Integrated Dental and Medical Electronic Health Record Data.

Area of Science:

• Oral health and systemic disease screening
• Public health and preventative medicine
• Health informatics and predictive modeling

Background:

• Undiagnosed diabetes and prediabetes pose significant public health challenges.
• Dental settings offer opportunities for early dysglycemia identification.
• Previous studies focused on urban, Hispanic populations.

Purpose of the Study:

• To recalibrate a dysglycemia identification approach for a White, rural cohort.
• To evaluate the value of an integrated dental-medical electronic health record (iEHR) for diabetes/prediabetes detection.
• To assess predictive model performance using combined dental and medical data.

Main Methods:

• Analysis of iEHR data from 4,560 dental patients (≥21 years) in rural Wisconsin.
• Inclusion criteria: no prior diabetes diagnosis, periodontal exam, and recent glycemic assessment.
• Multivariate logistic regression to develop and assess predictive models for glycemic status.

Main Results:

• An iEHR-integrated model demonstrated the best predictive performance (AUC=0.71).
• Key predictors included demographics, dental health (missing teeth, periodontal pockets), and medical factors (obesity, hypertension, hyperlipidemia, smoking).
• The model achieved 70% sensitivity and 62% specificity.

Conclusions:

• Dental care teams can identify patients at risk for undiagnosed diabetes or prediabetes.
• Integrating dental findings with medical EHR data significantly enhances prediction accuracy.
• Early identification through combined data sources can improve glycemic outcomes and prevent complications.