Background Drinking can be a social activity when partaken in certain environments; however it is not well understood how social capital may affect drinking patterns within the population in England. One study showed no association with social trust and support with consuming more than 2 units per day however did not assess consumption at greater levels. In better-off areas social capital may offer a protective effect against harmful levels of consumption (binge drinking) by promoting moderate consumption whereas in deprived areas drinking appears to be polarised between non-drinking and harmful levels. This study explores the relationship between individual and area-level social capital and drinking, and whether social capital mediates the association between deprivation and alcohol consumption.

Methods Using the Health Survey for England 2001–2006, a cross-sectional nationally representative sample of England, multilevel multinomial regression models on the probability of being a non-drinker/moderate (reference)/excess/binge drinker were assessed for men (N = 14,397) and women (N = 17,536). Key exposures include index of multiple deprivation quintiles and measures of individual (social trust, civic participation, social support) and area-level social capital (views of the neighbourhood). Models adjusted for age, education, ethnicity, income and marital status. Among drinkers, linear regression on the number of drinking days drank in the last week was carried out using the same variables.

Results At the individual-level, social gradients in non- and binge drinking were found. Those living in the most deprived areas had greater odds of being a non-drinker (men: OR = 1.62 (95% CI = 1.35–1.94); women: 1.59 (1.40–1.81)) and binge-drinker (men: 1.78 (1.56–2.03); women: 1.84 (1.60–2.12)). There was only slight attenuation after including measures of individual-level social capital. Severe lack of social support was associated with non-drinking for men (1.27 (1.07–1.50)) and women (1.26 (1.10–1.45)), and high civic participation had an inverse association (men: 0.78 (0.67–0.91); women: 0.78 (0.60–0.75)). Among men high civic participation (1.25 (1.12–1.39)) and weak social trust (1.16, (1.04–1.29)) were associated with binge drinking. Among women, high levels of civic participation decreased odds of binge drinking (0.77 (0.68–0.87)) and was associated with drinking fewer days in a week among women (-0.23, CI = -0.33 to -0.14).

Conclusion Social capital was associated with consuming alcohol within recommended limits, compared with non- or binge drinking in men and women. Men who binge drink may belong to groups where norms of heavy drinking are reinforced. Social capital enhances our knowledge of the psychosocial motivations for drinking. Further research is required to assess the effects of area-level social capital on drinking pattern.