Adjustment of the M-component of the DMFS index for prevalence studies of older adults.

Cross-sectional studies of caries in older adults report a substantial number of missing teeth, making it difficult to estimate caries experience accurately. The goal of this study was to improve the method of estimating caries experience as expressed by the DMFS index in population groups with missing teeth. The adjustment was demonstrated with reference to the Piedmont 65+ Dental Study conducted on a random sample of 363 community-dwelling older adults in North Carolina who were followed for 5 years. These older dentate adults had a mean (+/-SE) of 11.7 +/- 0.5 teeth missing at baseline, 56.1 +/- 2.5 missing surfaces and a DMFS of 86.7 +/- 2.0. A predicted caries prevalence was determined from the DFS at 5 yrs plus the 5-yr DFS incidence and the baseline DFS of teeth lost during the study period. Then a formula was developed that would estimate the predicted caries prevalence as a function of the observed 5-yr DMFS. This formula provided a good estimation of caries prevalence at 5 yrs (DMFSadj) when compared with the predicted prevalence (paired t-test, p > 0.05), while prevalence was underestimated by the DFS and greatly overestimated by the traditional DMFS index. Subgroup analyses by race, sex, and periodontal status also indicated that the DMFSadj resulted in patterns of estimates similar to the predicted prevalence, while the DFS and the DMFS were likely to result in different findings. The DMFS from the time of tooth eruption also was adjusted using this formula. The resulting analyses of subgroup differences in caries were not different from the previous estimates based on the 5-yr historical data, indicating that the adjustment of all M surfaces avoids the biases inherent in the traditional DMFS and DFS indices. This study showed that predicted caries prevalence could be estimated by adjusting the M component of the DMFS. It appears that this adjustment formula can be used without obvious bias, but additional studies are needed to provide adjustment figures for populations with different caries prevalences.