A Causative Approach to Demographic and Socioeconomic Factors Affecting Parental Ratings of Child Oral Health.

Abstract

Many studies have investigated associations between demographic, socioeconomic status (SES), behavioral, and clinical factors and parental ratings of child oral health. Caries experience, pain, missing teeth, malocclusions, and conditions and treatments likely to negatively affect the child or family in the future have been consistently associated with poorer parental ratings. In contrast, effect sizes for associations between demographic and SES indicators (race/ethnicity, country of birth, family structure, household income, employment status, and parental education levels) and parental ratings vary greatly. The primary objectives of this study were to estimate effect sizes for associations between demographic and SES variables and parental ratings of child oral health and then to consider possible causal implications. This article uses a nationally representative data set from 24,664 Australian children aged 5 to 14 y, regression analyses guided by a directed acyclic graph causal model, and sensitivity analyses to investigate effects of demographic and SES factors on parental ratings of oral health. One in 8 children had oral health rated as fair or poor by a parent. Indigenous children, older boys, young children with a migrant parent, children from single-parent families, low-income households and families where no parent worked full-time, and children whose parents had lower education levels were much more likely to receive a fair or poor parental oral health rating in crude and adjusted models. This cross-sectional study helps to clarify inconsistent findings from previous research and shows many demographic and SES variables to be strong determinants of parental ratings of child oral health, consistent with the effects of these variables on other health outcomes. Sensitivity analyses and consideration of the potential for chance and bias to have affected these findings suggest that many of these associations may be causal. Based on regression analyses driven by a directed acyclic graph causal model, this research shows a strong impact of demographic and socioeconomic determinants on parental ratings of child oral health, consistent with associations between these variables and other oral and general health outcomes. Many of these associations may be causal. We demonstrate the value of causal models and causal thinking when analyzing complex multilevel observational data.