Disparities in self-reported Mental Health, Physical Health, and Substance Use Across Sexual Orientations in Canada

Abstract

Background While prior studies have shown LGB individuals have elevated risk of poor mental health, poor physical health, and substance use, existing study designs may be improved by using representative samples, wider ranges of health outcomes, heterosexual comparison groups, and disaggregated data. The goal of this study is to provide estimates of multiple health disparities across sexual orientations in Canada based on these principles. Methods Using data from 2009-2014 Canadian Community Health Surveys, a sample of 19,980,000 weighted individuals was created. Outcomes included mental health, physical health, binge drinking, illicit drug use, and cannabis use. The study used logistic regression models adjusted by covariates, stratified by sex, to estimate health disparities across sexual orientations. Results Among LGB individuals, there was evidence for elevated risk of poor mental health (i.e. gay men, bisexual men, bisexual women), poor physical health (i.e. bisexual men, bisexual women), binge drinking (i.e. lesbians, bisexual women), illicit drug use (i.e. lesbians, bisexual women), and cannabis use (i.e. lesbians, bisexual women) relative to their heterosexual counterparts. Those identifying as 'don’t know' or 'refuse' showed reduced odds of substance use. Bisexual women exhibited highest disparities in health outcomes, e.g. OR=3.3, 95% 2.58 to 4.22 for poor mental health. Trends over time showed worsening mental health among bisexual women (relative to changes in heterosexual women), and decreasing substance use in gay and bisexual men, and lesbians. Conclusion This study highlights health disparities across sexual orientations in Canada, especially bisexual women, calling for targeted interventions (e.g. increased training of service providers in working with bisexual women and community outreach against biphobia). Future research should aim to explore these disparities longitudinally while also including the use of administrative-linked health data to reduce potential bias in self-reported data.