Caries experience in adolescents: is there still an association with socio-economic status?

Prevalence and severity of dental caries in adolescents of North- East Italy were estimated and the dependence on socio-economic status (SES) was evaluated. An epidemiological survey was performed among a cluster randomized sample of 560 (290 M, 270 F) adolescents attending secondary schools, to evaluate DMFS following WHO criteria. SES was determined from occupational levels of the parents, classified into 4 levels: high class (managers, professionals), clerks (clerical employees, managerial and technical occupations), self-employed (small employers and own account workers), working class (manual workers, skilled and unskilled, and housewives). [Costa et al: Int J Health Serv 2004; 33.635–67]. The DMFS score was the primary outcome in the study. Each subject was classified as being caries-free (DMFS = 0) or caries-positive (DMFS > 0). The association between caries occurrence and SES was evaluated in a logistic regression model. To account for high proportion of zero scores (DMFS distribution highly positively skewed) data was moreover modelled with negative binomial regression and zero-inflated models. All models evaluated showed a good fit to the data. Caries prevalence was 73.4% with a mean DMFS of 3.0 ± 3.8. Logistic regression showed no association between SES and caries occurrence. The Negative Binomial Regression model proved to be more appropriate than the Poisson model because the dispersion parameter was significantly different from zero (α = 1.5, SE 0.21, 95%CI 1.1,2.0). The zero-inflated negative binomial regression model (ZINB) was, as a whole, statistically significant with and without robust standard errors adjusting for clustering (Wald χ2 (5) = 12.7; p = 0.02), but no variable was able to predict the excess of zero and no predictor showed significant association with DMFS. In conclusion this work failed to prove an association between caries and SES in Italian adolescents