The Virginia Twin-Family Study of Adolescent Behavioral Development: assessing sample biases in demographic correlates of psychopathology.

The Virginia Twin-Family Study of Adolescent Behavioral Development (VTSABD) is a current longitudinal study of psychopathology in 1412 pairs of 8-16-year-old Caucasian twins and their parents. The primary aim of the study is to evaluate family-genetic and environmental risk factors for major domains of psychopathology in families representative of the Virginia Caucasian population. In this report, we utilize census-derived indices of neighbourhood income and urban residence to identify departures from population representation arising at the time of family enrollment in the twin registry and family participation in a psychiatric interview. Furthermore, we consider whether demographic sample biases influenced prevalence rates of adult psychopathology (including major depression, generalized anxiety disorder, alcohol dependence, phobias, and panic disorder) in the VTSABD. Results indicated that families that enrolled in the twin registry (83% of those identified by Virginia schools) and that participated in the home interview (75% of those targeted) resided in urban and rural communities with a range of per capita income levels representative of the Virginia population. However, participation biases operated throughout the study and were primarily characterized by losses of families living in low income, urban communities. There was also a smaller number of families living in high income neighbourhoods that did not enroll in the twin registry or that indefinitely postponed the psychiatric interview. These biases had small effects on prevalence rates of adult psychopathology in the VTSABD sample, even though neighbourhood income was significantly related to a subset of adult diagnoses. We emphasize the usefulness of the census methodology in evaluating sample biases in population-based psychiatric genetic studies.