Use of Regression Models for Estimating DMFT Index in 12-year-olds

Objectives: To determine the representativeness of each permanent tooth in the DMFT index in 12-year-olds and develop regression models to estimate the DMFT using teeth most attacked by dental caries. Methods: The random probabilistic sample was composed of 1,763 (824 examined in 2001 and 939 in 2005) schoolchildren (public and private) in Piracicaba, SP, Brazil. The examinations were performed by two calibrated dentists in an outdoor setting, under natural light, using mirror and ball point probe, in accordance with the WHO recommendations and after tooth brushing. The DMFT distribution (absolute and relative frequencies) was calculated as a function of each permanent tooth. The mean and the standard deviation were determined for all teeth (DMFT– real) as well as for the most affected ones (DMFT– partial). Regression models were estimated as a function of the most affected teeth using the data collected in 2005 and were validated using the data collected in 2001. Results: The mean (SD) for DMFT–real was 1.7 (2.08) in 2001, and 1.3 (1.9) in 2005. The DMFT-partial, which was estimated by the first molars, was 1.4 (1.6) in 2001, and 1.1 (1.4) in 2005. From the status of the first molars the partial-DMFT could be estimated at 82.0% and 81.5%, in 2001 and 2005, respectively. The regression equations estimated a DMFT of 1.67 in 2001, which represents 98.2% of the real-DMFT. Conclusion: The status of first molars was determinant for estimating the DMFT index in 12-year-old individuals. This study was supported by FAPESP (2004/06033-9), São Paulo, Brazil.