Objectives Substance use among young people is associated with poor health outcomes. Interventions to reduce consumption of tobacco, alcohol and illicit substances have been targeted on disadvantaged areas, and many address self-esteem or other aspects of subjective wellbeing. The evidence for this approach is conflicting. The aim of this paper is to explore the relationship between measures of subjective wellbeing, substance use and living in a low income household, for school children in England.

Design Analysis of a cross sectional survey of pupils from schools in two English local authorities in 2009 (Tellus4), using multivariate logistic regression. Eligibility for free school meals was used as a proxy measure for residence in a low income household.

Setting Schools in two local authorities in north west England

Participants 3903 schoolchildren in school years 6, 8 and 10 (aged between 10 and 15 years).

Main outcome measure Composite indicator of self-reported regular substance use, including smoking more than once per week, and being drunk and/or using drugs in the preceding 4 weeks.

Results Data were available from 3093 respondents, accounting for one third of registered pupils in these age groups. Respondents were distributed evenly across year groups, (1354 (34.7%) in year 6, 1266 (32.4%) in year 8, 1283 (32.9%) in Year 10). Each year group contained similar numbers of males and females (χ²=1.332, df=2, p=0.51) with 18.2% of the sample entitled to free school meals. Increasing age (year group) was strongly associated with regular substance use in multivariate analysis. Free school meal eligibility was associated with a greater likelihood of being a regular substance user, adjusting for age and subjective measures of wellbeing (adjusted OR=1.78, 95% CI=1.36 to 2.34). Two subjective measures of wellbeing, (happiness and the ability to talk to parents if worried), were inversely associated with substance use (adjusted OR=0.67, 95% CI = 0.52 to 0.86 and OR=0.51, 95% CI =0.39 to 0.65). A positive association was found between substance use and the ability to talk to friends if worried (OR=1.37, 95% CI=1.04 to 1.82).

Conclusion Living in a low-income household increases the risk of substance misuse for young people, though older age is the most important predictive factor. Subjective happiness and open communication with family rather than friends appear to help young people to avoid being regular users. Policies aimed at maintaining household income and stability in disadvantaged areas may be more effective than those targeting individual self efficacy at reducing inequalities in health.