Differences of in-home smoking prevalence among the four tobacco-use and cannabis-use groups were assessed using predicted probabilities from mixed effects logistic regression models. All models were adjusted for age and sex with country of origin as a random effect to account for variation in sample sizes of each country and to address the possibility of country-specific influence on associations of in-home smoking by cannabis and tobacco use. Odds ratios comparing cannabis-only users, tobacco-only users, and dual users to non-users were computed separately for in-home cannabis smoking and in-home tobacco smoking. The ORs for in-home cannabis smoking and in-home tobacco smoking were visually compared by observing non-overlapping 95% confidence intervals. Statistical differences in the ORs were assessed using generalized estimating equations by implementing the methods outlined by Horton & Fitzmaurice .
To address possible confounding by participant-level characteristics and country-level characteristics , ORs were computed from models that additionally adjusted for clubbing frequency, drinking frequency, past year drug use, ploy-drug use, area of residence, vertical grow system and country as a fixed and random effect. Logistic regression models were then repeated for each included country to illustrate country-specific differences of in-home smoking among the four mutually exclusive cannabis- and tobacco-use groups. All statistical tests were two-tailed with an alpha of 0.05. Data analysis was performed using R . Global trends in the decriminalization and legalization of cannabis use should prompt increased research that seeks to identify and constrain harm and improve public health. In this large convenience sample of people from 17 countries who used at least one psychoactive drug in the past year, past-year in-home cannabis smoking was slightly more prevalent than past-year in-home tobacco smoking in the overall sample. Approximately 80% of current cannabis-only users reported that cannabis was smoked in their home while in-home tobacco smoking took place in the residences of 68% of current tobacco-only users.
Overall past-year rates of in-home smoking were high among current users of both cannabis and tobacco, and in these groups, in-home cannabis smoke was more prevalent than in-home tobacco smoke. Taken together, these results support our hypothesis that in-home cannabis smoking would be higher than in-home tobacco smoking. Since this is the first study we are aware of that measured the behaviors of in-home cannabis and tobacco smoking, there are no studies to which we can directly compare our results. From studies that focused on household rules, evidence from a convenience sample of US Facebook users reported that of the 54% of respondents who allowed cannabis use on their property, 71% allowed cannabis grow equipment smoking inside their home; unfortunately, rates of in home tobacco smoking were not available . Another study, among US university students, reported that in-home cannabis use was allowed by 36% of tobacco smokers and 59% of cannabis smokers while in-home tobacco use was allowed by 25% of tobacco smokers and 36% of cannabis smokers .
In our study, the generally higher rates among US respondents could be attributed to our assessment of home smoking behavior rather than home smoking policy. They could also be attributed to selection bias from studying self-selected sentinel drug users, though a recent study showed that the age and sex distributions of GDS respondents who use cannabis were similar to probability-based samples in the three countries studied—Australia, Switzerland, and the US . The high rates might also be partly the result of our yes/no inquiry about past-year in-home smoking, which could include rare instances of the behavior that are not representative of usual behavioral patterns.In a recent cross-sectional study in Australia, Canada, England, and the United States, 73.6%, 78.3%, 65.5%, and 60.8% of respondents, respectively, endorsed cannabis smoking as less harmful than cigarette smoking . Decades of tobacco control campaigns and policy likely increased the perceived harm of tobacco smoke, contributing to the lower rates of in-home tobacco use. If true, results from this study suggest that it may be time to develop similar campaigns and policy to correct perceptions of harm of cannabis smoke.