Using latent class analysis of survey data to explore parent behaviors and attitudes regarding children's COVID vaccinations.

We conducted exploratory data analyses of the National Immunization Survey-Child COVID Module (NIS-CCM). NIS-CCM is a random digit dialing survey of parents and guardians of 6-month- through 17-year-old children. We conducted latent class analysis to identify different population segments of children according to combinations of parental/guardian intent, attitudes, and behaviors regarding their children's COVID vaccinations. We analyzed data collected between October 2021 and April 2023. The latent class model variables with the greatest variation in conditional probabilities across latent classes included vaccination status, intent to vaccinate, and confidence in the safety of COVID vaccine. The latent classes included one with vaccinated children despite low confidence in vaccine safety and one with unvaccinated children despite high confidence in vaccine safety and high vaccination intent. The latent class that increased the most in size over the period studied was the class most hesitant to vaccination. We further analyzed demographic and environmental characteristics associated with membership in different latent classes using multinomial logistic regression, with the goal of informing strategies to increase vaccination coverage with different population segments. Socioeconomic status and race/ethnicity were associated with latent class membership. Specific social and environmental factors, such as having many family and friends' children vaccinated for COVID and receiving a provider's recommendation for vaccination were associated with more favorable-to-vaccination class membership. While the exploratory analyses do not establish causal relationships, the research shows the value of latent class analysis with high-quality survey data to inform further research on strategies to increase vaccination coverage.