Correlations between caries prevalence and potential etiologic factors in large samples of 4-5-yr-old children

A data base containing information on dental caries, oral hygiene, gingival state, Streptococcus mutans and lactobacillus counts, and salivary flow rate and buffer capacity in some 2800 4-5-yr-old South African children was created in 1984. The children were equally distributed in four groups; rural black, urban black, urban Indian, and urban white. In this study, the data were used for simple linear correlations between all the factors including caries as well as stepwise multiple regression analyses between caries prevalence and the independent factors in each of the four groups. The highest r value obtained in the single correlations was 0.56 for dmfs/lactobacillus count in the white group. The overall pattern of the groups did not vary much. In the stepwise multiple regression analyses, lactobacilli entered first in all groups. S. mutans and oral hygiene interchanged as Nos. 2 and 3, and gingival state or buffer capacity came out as No. 4. The highest percentage of caries variability explained by these regressions was 25%, in the urban black group, while the lowest was 12%, in the indian group. It was concluded that our generally accepted etiologic factors are unsatisfactory to explain variation in caries in observational studies, which among other things might be ascribed to the inadequate way these factors generally are measured or estimated.