Abstract
BACKGROUND: The choice and definition of a comparison group in alcohol-related health studies remains a prominent issue in alcohol epidemiology due to potential biases in the risk estimates. The most commonly used comparison group has been current abstainers; however, this includes former drinkers who may have quit drinking due to health problems. Lifetime abstention could be the best option, but measurement issues, selection biases due to health and other risk factors, and small numbers in populations are important concerns. This study examines characteristics of lifetime abstention and occasional drinking that are relevant for alcohol-related health studies. METHODS: This study used data from the National Longitudinal Survey of Youth 1979 cohort of 14 to 21 year olds followed through 2012 (n = 7,515). Definitions of abstinence and occasional drinking were constructed based on multiple measurements. Descriptive analyses were used to compare the definitions, and in further analysis, lifetime abstainers (n = 718) and lifetime minimal drinkers (n = 1,027) were compared with drinkers across demographics and early-life characteristics (i.e., religion, poverty, parental education, and family alcohol problems) in logistic regression models. RESULTS: Using a strict measurement of zero drinks from adolescence to the 50s, only 1.7% of the sample was defined as lifetime abstainer compared to a broader definition allowing a total of 1 drink over the lifetime that included 9.5% and to lifetime minimal drinking (a total of 3 drinks or less a month), which accounted for 13.7%. Factors significantly associated with lifetime abstention and lifetime minimal drinking included religion, poverty, having no family alcohol problems, Hispanic ethnicity, foreign-born, and female gender. Importantly, work-related health limitations in early life were significantly associated, but not childhood physical and mental health problems. CONCLUSIONS: Alcohol-related health studies should utilize lifetime classifications of drinkers and abstainers, and, in doing so, should account for early-life socioeconomic adversity and childhood health factors or consider these as unmeasured confounders.