Figure. Regression-Adjusted Rates of Fluoride Varnish Applications During Well-Child Medical Visits, (N = 328 661).

The figure presents unadjusted and regression-adjusted probabilities. Unadjusted probabilities are derived from logistic regression models predicting fluoride varnish applications during well-child medical visits. Models are estimated separately for each of the independent variables of interest (ie, age, state, and year). Regression-adjusted results are from a model controlling for child sex, age, insurance type, visit year, and state. In all models, standard errors are clustered at the county level, and the delta method was used to calculate standard errors for predicted probabilities. We calculated probabilities of fluoride varnish application during a well-child medical visit for each variable category of interest using model estimates to compute the mean of predicted probabilities for the entire sample after setting the covariate of interest (eg, Connecticut) while keeping all other covariates at their observed values.

^a Indicates difference in predicted probability is statistically significantly different from the reference group (child age = 5 years, state = Connecticut, year = 2016) at the 5% level.