Are substance use, abuse and dependence associated with study participation? Predictors of offspring nonparticipation in a twin-family study.

OBJECTIVE: Although epidemiologic studies have reported that problem drinking is associated with nonresponse to surveys, it is unclear whether parents' alcoholism is associated with nonresponse in their offspring. This question is particularly important to family studies of alcoholism. In the current study we constructed a model of offspring nonparticipation in a twin-family design and computed weights to recapture the distribution of offspring alcohol abuse and dependence.
METHOD: In 1999, the first wave of a longitudinal study of offspring of alcoholic twins was conducted via telephone interview with members of the Vietnam Era Twin Registry. The target offspring sample consisted of 2,096 male and female children, of whom 1,270 were successfully interviewed. Offspring response status was classified as participation, refusal or unavailable/no consent. Stepwise logistic regression models were used to identify variables that were significantly associated with one or both types of offspring nonparticipation. A multinomial logit procedure with backward deletion was then used to build a model of the three levels of child response.
RESULTS: Paternal alcoholism was not significantly associated with offspring nonresponse, although offspring nonparticipation because of not being located, or being deceased, disabled or unavailable was associated with current paternal smoking, paternal divorce and paternal marital status (after adjustment for other predictor variables).
CONCLUSIONS: The most important conclusion to be drawn from current results is that the alcohol abuse and dependence history of fathers should not bias analyses in family studies of alcoholism when data are collected via telephone interview. Study limitations and directions for future research are discussed.