Validation of a method for collecting annual, population-based oral health data for the MCH Title V Block Grant

This study sought to determine whether estimates of oral health-related indicators generated using a relatively small purposive sample of schools were reasonable.

We collected and analyzed data from two open-mouth school oral health surveys conducted in Ohio in 2004-2005. In each survey, data were gathered for nine indicators using direct observation and a questionnaire. The first survey used unweighted data from 23 sentinel schools purposively selected to represent the state based on the findings of a large 1998-1999 oral health survey (337 schools). The second survey used weighted data aggregated from 374 schools randomly selected to represent each of Ohio's 88 counties. The point estimates from the first data set were compared with 95% confidence intervals (CIs) from the second and Chi-square measures were estimated to determine statistically significant differences.

For six of nine indicators, point estimates for the sentinel schools fell within the relatively narrow 95% CIs generated for the 374 schools sampled at the county level. Only one indicator (history of tooth decay) was found to be significantly different according to Chi-square analysis.

Ohio's sentinel schools approach provided reasonable estimates suitable for annual reporting as required for the Title V Maternal and Child Health Block Grant. Data collected in this manner are likely to be reasonable population estimates and less costly than larger surveys.