Substance use and sexuality: Comparing sexual identity and attraction using a multilevel multivariate model

Abstract

Previous research has found that sexual minority individuals are more likely than heterosexual individuals to engage in substance use behaviours such as the consumption of illicit drugs, heavy drinking and tobacco smoking. This study innovates by comparing the association of different measures of sexuality with substance use behaviours and exploring whether this association varies by place. We exploit multivariate multilevel models with sexual attraction and sexual identity as exposure measures and Great Britain Local Authorities (LA) as places. This study compares multivariate models for men and women, comparing measures of sexuality using sexual attraction and sexual identity to look simultaneously at cannabis use, current smoking and alcohol use, and variance across LAs. Sexual attraction was a stronger predictor of engaging in substance use behaviours compared to sexual identity for men and women, and LGB women experienced greater odds of engaging in all substance use behaviours than LGB men. After controlling for relevant covariates, there was residual LA level variation only for women cannabis users, although it was not significant. These findings highlight the importance of using different measures of sexuality, and that LGB women are more vulnerable to engaging in risky health behaviours. LGB women in particular may require sexuality-specific interventions, support and policies in LAs with greater cannabis use.

Keywords

• Sexual minority
• Smoking
• Risk behaviours
• Multilevel modelling
• Multivariate