Factors, including socioeconomics, education level, and insurance status, appeared to play a major role in delayed dental care among U.S. adults, according to a study recently published in the International Dental Journal.
Furthermore, machine learning models may help identify and classify risk patterns across populations, the authors wrote.
“Our findings confirm the central role of socioeconomic status, education and insurance coverage as primary determinants of delayed dental care,” wrote the authors, led by Dr. Giang Vu, PhD, of the University of Central Florida School of Global Health Management and Informatics (Int Dent J, February 6, 2026, Vol. 76: 2, 109407).
There were three aims of the study: one, estimating the prevalence of delayed dental care among U.S. adults; two, detailing disparities across socioeconomic and demographic groups, and; three, evaluating the value of machine learning in identifying high-risk individuals. The researchers conducted a cross-sectional analysis of data from the 2023 U.S. National Health Interview Survey. The sample included 54,927 adults ages 18 and older, they wrote.
To measure delayed dental care, respondents were classified as either having delayed care in the past 12 months or not. Survey weights were applied to generate nationally representative estimates; descriptive statistics were stratified by key demographic factors, and supervised machine learning classifiers were developed to predict delayed dental care.
Delays were more common among adults ages 35 to 50 (18.3%) and 51 to 64 (17.2%) and less frequent among those 65 and older (11%), with 14.5% of U.S. adults reporting they delayed dental care in the previous year. Racial and ethnic disparities were apparent, with a higher prevalence among Black/African American adults (19.1%) and individuals reporting multiple races (21.1%) compared with white adults (14.2%), they wrote.
Additionally, adults without dental insurance were significantly more likely to delay care, and a clear educational gradient showed higher delays among those with lower levels of education. Among the machine learning models evaluated, LightGBM performed the best (accuracy = 84.87%), with education, income-to-poverty ratio, and insurance status emerging as the strongest predictors.
However, the study had limitations. The analysis was based on self-reported survey data, which may be affected by recall errors and social desirability bias, the authors added.
“Policies that expand dental coverage, reduce financial barriers and target outreach to high-risk populations may mitigate inequities,” Vu and colleagues concluded.
