Classifying Delayed Dental Care Using Machine Learning: A National Health Interview Survey Analysis

Area of Science:

• Public Health
• Health Services Research
• Health Equity

Background:

• Delayed dental care exacerbates oral health issues and reflects broader healthcare access disparities.
• Understanding the prevalence and drivers of delayed dental care is crucial for addressing oral health inequities.

Purpose of the Study:

• To estimate the prevalence of delayed dental care among US adults over 12 months.
• To identify disparities in delayed care across demographic and socioeconomic groups.
• To evaluate the utility of machine learning (ML) for identifying at-risk individuals.

Main Methods:

• Cross-sectional analysis of the 2023 National Health Interview Survey (NHIS) sample adult file (N=54,927).
• Descriptive statistics stratified by age, sex, education, insurance, and race/ethnicity.
• Development and evaluation of ML classifiers (LightGBM) using accuracy, precision, recall, F1-score, AUC, and SHAP for interpretability.

Main Results:

• 14.5% of US adults reported delaying dental care in the past year.
• Higher delay rates observed in adults aged 35-64, Black/African American individuals, multiracial individuals, and those with lower educational attainment.
• Lack of dental insurance significantly increased the likelihood of delayed care. LightGBM achieved 84.87% accuracy, with education, income-to-poverty ratio, and insurance status as key predictors.

Conclusions:

• Delayed dental care is prevalent and disproportionately affects specific demographic and socioeconomic groups.
• Interpretable ML models can aid in risk stratification for targeted interventions.
• Policies expanding dental coverage and reducing financial barriers are recommended to mitigate inequities.