Aims To describe the pattern of drinking at age 16 and to relate this to outcomes at 16–21 years and 21–25 years across a number of psychosocial domains.
Design A prospective birth cohort study with annual follow-up until age 16 then at 18, 21 and 25 years.
Setting Christchurch, New Zealand.
Participants Of 1265 subjects, 953 were interviewed at age 16.
Measurements Multiple measures of family background were collected from birth to 16 years. Alcohol consumption was measured in terms of frequency, usual or last quantity drunk and most drunk per occasion. Problems were also recorded. Questions about psychiatric symptoms enabled Diagnostic and Statistical Manual (DSM) criteria to be applied. Detailed reports on educational outcomes, employment, sexual behaviours and offending were collected.
Findings Four latent classes were required to describe drinking at age 16, but these appeared to lie along a single dimension which strongly predicted outcomes at ages 16–21 and 21–25 across all domains (alcohol-related, substance dependence, mental health, education, sexual relationships and offending). After controlling for background and correlates only a small number of outcomes were still related consistently to drinking at age 16 over both periods: most alcohol-related outcomes, the number of sexual partners and the extent of violent offending.
Conclusions Drinking at age 16 is a clear indicator of future life-course over most domains in late adolescence and early adulthood. Many of these associations are due to other covariates. Outcomes specific to drinking at age 16 are alcohol outcomes, number of sexual partners and violence.