Background Non-drinking among young people has increased over the past decade in England, yet the underlying factors driving this change is unknown. Traditionally non-drinking has been found to be associated with lower socio-economic status and poorer health. This study explores among which sub-groups non-drinking has increased, and how this correlates with changes in drinking patterns, to identify whether behaviours are becoming more polarised, or reduction is widespread among young people.

Methods Using repeated annual cross-sectional data on young people aged 16 to 24 years from the nationally representative Health Survey for England 2005–2015, trends in non-drinking including lifetime abstention and not drinking in the past week were explored using STATA15; 1) Trends in non-drinking among social-demographic and health sub-groups were examined through a test for linear time-trends among sub-groups, adjusting for age. Additionally, an interaction between year and each variable was modelled in sex- and age-adjusted logistic regression models 2) Spearman correlation co-efficients were calculated between the proportion non-drinking by year, and heavy episodic drinking, and the mean alcohol units consumed on the heaviest drinking day. In addition, ordinary least squares regression analyses were used, modelling the proportion non-drinking as the independent variable, and the proportion binge/mean units as the dependent variable.

Results Rates of non-drinking increased from 18% [95% CI 16% to 22%] in 2005 to 29% [25–33%] in 2015 (test for trend; p<0.001), largely attributable to increases in lifetime abstention. Not drinking in the past week increased from 35% [32–39%] to 50% [45–55%] (p<0.001). Significant linear increases in non-drinking were found among most sub-groups including healthier sub-groups (non-smokers, those with high physical activity and good mental health), north and south regions, in full-time education, and employed. Among white participants, non-drinking increased from 14% [12–17%] in 2005 to 20% [17–24%] in 2015. No significant increases in non-drinking were found among smokers, ethnic minorities and those with poor mental health. At the population-level, a one percentage point increase in non-drinking, predicted a 0.22 reduction in mean alcohol units consumed [95% CI −0.32 to 0.12], and a 1.06 percentage point decrease in the proportion binge drinking [95% CI 1.56 to 0.54]

Conclusion Increases in non-drinking were found with variables less commonly associated with non-drinking, suggesting this behaviour may be becoming more normative. This trend is to be welcomed from a public health standpoint and should be capitalised on going forward. Drinking and smoking continue to cluster and could be targeted in tandem. Future research should explore attitudes towards alcohol among young people.