Objectives:
to compare the results of hierarchical and conventional multivariate models in the investigation of determining factors for tooth decay in children.
Methods:
an analysis was conducted of an investigation of children served by the Family Health Strategy in Recife, Pernambuco, carried out 2006. The dependent variable was the prevalence of tooth decay (ceo-d≥1); The independent variables were ordered in blocks: socio-economic, life/family context, dental health care. Variables statistically associated with ceo-d ≥1 in univariate analyses with a level of significance ofp≤0.20 were included in the multivariate models. In the hierarchical mode, the order of inclusion of variables was guided by a multidimensional conceptual model. Associated variables with a level of significance of p ≤ 0.10 were retained in the final model. In the conventional logistic model, the significant variables from the univariate analyses were simultaneously included and adjusted. Variables associated with ceo-d with p ≤ 0.10 were retained.
Results:
in the conventional model the factors associated with ceo-d≥1 were high household density, low level of education of caregiver; studying at public school, seeking treatment/emergency services, and eating sweets. In the hierarchical model, the occupation of the father and the level of education of the mother were also retained in the final model.
Conclusion:
hierarchical analysis enabled the identification of distal determining factors, which influence lower level factors and their influence on the outcome, thereby providing a more contextualized discussion of the issue.
Dental caries; Child; Social conditions; Health services; Multivariate analysis
Thumbnail
