This application responds to the April 12, 2004 program announcement entitled """"""""N1DCR Small Research Grants for Data Analysis and Statistical Methodology (R03); PAR-04-091"""""""". The Florida Dental Care Study (FDCS) is a prospective observational study of oral health and dental care based on a diverse, population-based sample including many subjects at increased risk for oral health disparities (e.g., African Americans, persons of low socioeconomic status, and persons whose approach to dental care is problem-oriented, rather than based on regular attendance). FDCS is the first study to use an innovative approach that directly links clinical exam data, self-reported dimensions of oral health, use of specific dental services, and diverse patient characteristics in a population-based sample of subjects who may or may not be regular attendees at dental clinics. This data set, which includes longitudinal observations across 72 months of follow-up, provides a unique database that is unprecedented in adult oral health disparities research, oral epidemiology, and dental health services research. So far, important hypotheses about the relationship between dental disease measured clinically, self-reported dimensions of oral health, health-related quality of life, and use of dental care have been tested. While the randomized trial has traditionally been considered the gold standard study design for evaluation of medical treatment effects, observational trials have important strengths. Observational studies allow evaluation of the effects of treatments for which randomized assignment is impossible, such as those for which subjects self-select into treatment groups, or for which a representative sample of the population of interest is not attainable. However, because subjects are not randomly assigned to receive treatment or not, observed group effects are likely to be confounded by characteristics, observed or not observed, other than the group difference that is of interest. This confounding leads to a potential for biased estimation of treatment differences, known as sample selection bias. The goals of this 2-year project are (1) to quantify the effects of sample selection bias on conclusions made about the effectiveness of specific types of dental services, using a broad range of oral health outcomes and treatment types, and (2) to measure whether these effects of selection bias in Specific Aim 1 differ by race, socioeconomic status, or typical approach to dental care (regular versus problem-oriented attendance). ? ? ?
