Comparison of Several Count Regression Models on Modeling Decayed Missed Filled Teeth Dental Index in Dentistry

Background Oral diseases are common in many communities and dental caries is the most prevalent disease among children and adults. DMFT (Decayed Missed Filled Teeth) is one of the useful indexes in dental epidemiology. This study aimed to investigate caries epidemiology among students and compare several modeling of DMFT based on real data.
Materials & Methods This cross-sectional study was conducted on school children aged 7-12 years in Khoramabad City, Iran during 2010 to 2011. A total of 920 samples were recruited by multistage random sampling method. Regarding to countable data, right skewness and zero inflated variable of DMFT index, different models such as Poisson regression, negative-binomial regression, and zero-inflated Poisson regression were used for modeling, and the selection of the best model was based on the minimum amount of AIC and BIC. Data analysis was performed using Stata version 12, according to significant level of 5%.
Results In this study, 43% of school children were girls and the rest were boys, so that their Mean±SD age and DMFT were 9.02±1.49 years and 1.02±1.35, respectively. A total of 528 (out of 920) children had dental caries. Zero-inflated Poisson regression, comparing with other models, was of the best model for goodness of fit among the fitted models. This model revealed significant relationships between being at risk of dental caries and variables of age, father’s educational level, and presence of microbial plaque (PConclusion The best regression method for modeling DMFT among all models in this study was zero-inflated Poisson regression. Age, father’s educational level, and presence of microbial plaque were significantly correlated with children’s dental caries.