P2-184 Measures of health inequality applied to dental inspection data

The aim of this study was to apply tests of inequality to dental inspection data from 3-year-old children in Greater Glasgow and Clyde and to compare and contrast their usefulness. Dental inspections were repeated from 2006/7 to 2009/10. The measures of caries experience used were d3mft >0 and mean dmft, while the measure of socioeconomic status was “quintiles” of the Scottish Index of Multiple Deprivation. Analyses of each year's data included: raw summaries by SIMD score; logistic regression (LR) models with ROC plots and c-index; Significant Caries Index (SIC); Population Attributable Risk (PAR); Concentration Index; Gini Coefficient and Lorenz Curves; Slope Index of Inequality (SII) and Relative Index of Inequality (RII). We inspected 10 022 three year old children (19% of the population). The prevalence of obvious decay was 26%, 25%, 18% and 17%, in the 4 years. Raw analyses revealed a drop in absolute inequality, but little change in relative inequality. The SII values over the 4-year period were −1.61, −1.38, −1.16 and −0.54, respectively. ORs from the LR models provide relative estimates (backed up by the ROC plots), the SIC and the PAR are good at picking up the absolute reduction, the Concentration Index does not add much information, the GINI coefficient does pick up inequality but is counter-intuitive in this context, and the SII and RII are useful. We found that standard analyses such as LR and ROC are good for estimating dental inequalities, but other methods are needed to more graphically demonstrate reductions in absolute inequality.