Predictors of oral health in Australian children

Abstract

Introduction: Dental caries is a preventable non-communicable disease (NCD) and the causal biological mechanisms have been established in the literature. Despite this, the Global Burden of Disease Study 2017 estimated 500 million children globally had dental caries in their primary dentition. As with many other NCD’s dental caries is associated with behavioural and lifestyle factors. Effective prevention of NCD’s understands the social and environment factors which influence individual’s experience of disease. Newton and Bower (2005) discussed the absence of theoretical frameworks guiding empirical research were preventing more complex models from being quantified explaining the broader social determinants. Subsequently, Fisher-Owens et al. (2007) published a conceptual model of children oral health, which used a multi-level ecological perspective to conceptualise the various influences. The relationships described in conceptual model have been quantified in multiple countries including Brazil, New Zealand, the Unites States and more. In Australia, the studies investigating influences on Australian children’s oral health generally employed methodologies that were qualitative, cross-sectional or focused on singular factors failing to account for confounding factors. This thesis aimed to describe the patterns of oral disease in a nationally representative cohort of Australian children and estimate the effect of community, family, and child-level influences on oral health. Methods: The Longitudinal Study of Australia Children (LSAC) is a government-led investigation into the environment in which Australian children grow up and its impact on development and well-being. Data collection commenced in 2004 and adopted a biennially cross-sequential dual cohort design. At baseline, the LSAC recruited two geographically and demographically representative cohorts of children aged zero and four years of age. Children’s parent or guardian (herein referred to as carers) reported at each study wave their child’s experience of dental caries and injury since the previous wave (two years). Between LSAC waves six and seven (2015 to 2016), a nested cross-sectional study the Child Health CheckPoint undertook a variety of health assessments in LSAC’s birth cohort. Intra-oral photographs were taken of the children and their parents as part of the CheckPoint study. The photographs were assessed using previously validated indices for the health of their teeth, oral hygiene and orthodontic status. Analysis of available unadjusted dental variables included a descriptive analysis of frequencies over time and Generalised estimating equations (GEE) were used to assess age, period and cohort effects. GEE were used to estimate the regression coefficients of the community and family-level factors meeting the Fisher-Owens conceptual model domains across the study period (waves one to seven; or 2004 to 2016) with carer-reported caries. Concordance between biological parents and children’s oral health in the Checkpoint study was assessed by Cohen’s kappa, Spearman’s and Pearson’s correlation coefficients. Structural equation modelling (SEM) were used to model child-level factors for carer-reported caries experienced at ages four and eight. Results: A total of 10,090 children were recruited at baseline in the LSAC with n= 5,107 in the birth and n= 4,983 in the kindergarten cohorts. The majority of children were born in Australia, non-Indigenous and lived in a major urban area. Increasing age was a significant predictor of dental caries, whereas no differences were observed between time periods and cohorts. Caries was highest at age eight for the B (n=1,234, 30.5%) and K (n=1,355, 31.5%) cohorts. On a community-level, children living in Queensland (OR 1.48, 95% CI 1.35-1.62) and low socio-economic status (SES) areas (OR 1.32, 95% CI 1.20-1.44) had an increased odds of dental caries over time. On a family-level, predictors of dental caries over time were younger mothers, lower parental education, increasing poor parental global health scores, parents currently smoking, English as a main language at home and Indigenous parents. In the Checkpoint study, 13.0% of children (95% CI 11.3 - 14.9) had poor oral hygiene, compared with 2.1% (95% CI 1.4 - 3.0) of parents. Gingivitis measured by the modified gingival index in photographs showed a high parent-child correlation coefficient (CC 0.49, 95% CI 0.44 to 0.53). On a child-level, past dietary behaviours had a direct effect on dental caries at 4/5 years (β= .28, 95% CI [.22, .32]) and 8/9 years (β= .10, 95% CI [.05, .14]). The greatest predictor of carer-reported dental caries at age four was diet, whereas a history of carer-reported dental caries was the greatest effects observed in eight-year-old model (β= .26, 95% CI [.20, .32]). In the four and eight year old models, Indigenous status and childhood development significantly predicted on carer-reported caries indirectly mediated through diet. Conclusions: This thesis identified a number of influences of dental caries in children. It is evident dental caries is a chronic disease, as the greatest predictor of caries in mid-childhood was a past experience of caries. The high concordance for gingival health between child-parent pairs supports the genetic and behaviours links established in previous studies. A focus on prevention in early childhood urgently needs to be prioritised by all relevant stakeholders to prevent caries. Using empirical statistical models to quantify the influence of community was challenging. Drawing on this thesis, broader public health and implementation science literature, a new conceptual model has been proposed which further develops on the Fisher-Owens concepts to provide a holistic view of the complexities of children’s oral health. By mapping context and stakeholders in this new model future research can be guided to promote stakeholder collaboration and implementation of findings for the betterment of children’s oral health.