We demonstrate that in the U.S. adult population, the perception of cSHS harm is a key factor related to having a complete in-home cannabis smoking ban. This was true for the whole target population and, with only two exceptions, in every subgroup examined (see Fig. 1). In nearly all analyses, the perception of harm of cSHS exposure at any level more than totally safe was associated with higher odds of having a complete ban on in-home cannabis smoking than of having no complete ban. One subgroup result of note is the three-fold higher odds of a complete ban on in-home cannabis smoking among past-year cannabis smokers who reported cSHS exposure as extremely harmful vs. totally safe. It is evident that even in households with past-year cannabis smokers, perception of harm is highly positively associated with the presence of a complete ban.

There are very few studies, to our knowledge, that have examined the relationship between the perception of harm of cSHS exposure and a complete ban on in-home cannabis smoking. In one study among young adults (ages 18–34) in the U.S., recruited from social media websites, the likelihood of allowing cannabis use in the home was inversely but not significantly associated with the perception of harm from byproducts of cannabis smoke or vapor (beta = -0.03 (95% CI: -0.07, 0.00) [28]. Similar to the estimated 28% in our study who reported cSHS exposure as totally or mostly safe, a 2018 study focused on the perception of cSHS exposure with respondents from KnowledgePanel® reported an estimated 32% of U.S. adults perceived exposure to cSHS as not at all or a little harmful [29], with younger age, recent cannabis use, recent tobacco use, cannabis, and tobacco co-use, and non-White race/ethnicity being positively related to increased likelihood of perceiving cSHS exposure as not harmful [29]. In studies examining household rules, cannabis smoking was more often allowed inside the homes of young adults, cannabis smokers, tobacco smokers, and cannabis and tobacco co-users than inside the homes of non-smokers [13, 14]. Additionally, in homes of young adults, peer use and perceived social acceptability of using cannabis were positively correlated with allowing cannabis use inside homes [28].

In our study, even among past-year cannabis smokers, who are the most likely to put non-smoker residents of their homes at risk of cSHS exposure, the association between perceiving cSHS as harmful and the presence of a complete ban versus not having a complete ban was strong. Past-year cannabis smokers who perceived cSHS as harmful had a three-fold increase in the odds of having a complete ban, compared to a seven-fold increase in the odds among never-cannabis smokers and an eight-fold increase in the odds among former cannabis smokers. The differences in the relationship may be due to skewed perceptions that past-year cannabis smokers have about harms related to cannabis smoke; cannabis smokers perceive cannabis as less addictive and a “healthier” alternative to smoking tobacco [34]. The perception of cSHS as not harmful has been associated with recent cannabis use [29] as well as regular cannabis use [30]. Additionally, past-year cannabis smokers are unlikely to restrict cannabis smoking inside their own homes, as cannabis users have a high (50%) prevalence of in-home cannabis smoking [16]. Educating past-year cannabis smokers on the harms and risks of cannabis smoke could substantially reduce cSHS exposure, as 40% of those without a ban on in-home cannabis smoking were past-year cannabis smokers.

Perception of harm of cSHS exposure was strongly related to a complete ban on in-home cannabis smoking among respondents with and without children under 5 years living in the home. But unexpectedly, among these respondents, for those with children under 5 there was a weaker association than among those without children under 5 (ORs: 2.6 vs. 6.8). These differences may be due to other drivers of the behavior of parents or those with children living at home. Tobacco SHS studies show that while knowledge of risk and perception of harms are important factors, social norms (communities where the high value placed on social relationships makes changing a guest’s behavior difficult and where smoking functions as a positive shared activity), gender imbalances (women’s lack of agency in affecting rules and others’ behavior), and structural factors (living in someone else’s home, such as parent’s house, so cannot establish household rules) are also barriers to smoke-free homes [35]. Understanding other important drivers is important for cannabis smoke-free home efforts, as decreasing or eliminating cSHS exposure of children could greatly impact children’s health outcomes [19, 21–23].

Lastly, among respondents living in states where cannabis use was not legalized, the odds of having a complete ban among those reporting cSHS as “extremely harmful” (ref: “totally safe”) was lower than among respondents living in states where medical or recreational cannabis use was legalized (ORs: 4.2 vs. 7.6 vs. 7.2). While all three associations are strong and statistically significant, the differential odds by cannabis legalization status may be due to various reasons not measured in our study such as ideological and cultural differences, access to educational resources about cannabis, and differing magnitudes of social desirability bias pertaining to answering questions about cannabis smoking perceptions or behaviors. To our knowledge, there are no studies on reasons cannabis legalization status may affect the relationship between perception of harm and complete ban of in-home cannabis smoking. Future studies are needed, to replicate the results by cannabis use legalization. Replication of our results and the causality of the relationship would indicate that in states where medical or recreational cannabis is legalized, campaigns that aim to shift perception of harm of cSHS will have a strong impact on residents to make changes such as instilling household rules on in-home cannabis smoking. Future studies advancing our understanding of how cannabis legalization status might influence public health campaigns to eliminate in-home cannabis smoking are also needed.

Our study had several noteworthy strengths. The large representative sample and low levels of missing data strengthen the external validity of our findings and allow us to examine effect measure modifications by important covariates. The low level of missingness in our data indicates that selection bias was unlikely in our study, conserving the internal validity of our study findings. Our question from MUES on perceived harm of cSHS exposure asks about the degree of harm from exposure to cannabis smoke at least three times per week. There is no standard measurement of the health risk of cSHS, but surveys such as the National Survey on Drug Use and Health use “once or twice a week” to measure the perceived risk from smoking cannabis [36]. The frequency of “at least three times per week” used in MUES captures more frequent exposure to cSHS, which may be a cause of greater concern for the general population as compared to occasional cSHS exposure (e.g., once a week or once a month).

Limitations of our study included the cross-sectional study design, which provides no information concerning the temporal order of the associated variables, precluding inferences about causality. However, while it is plausible that the implementation of in-home rules for cannabis smoking could cause someone to perceive cSHS exposure as harmful, this seems unlikely. While our study leads us to recommend promoting the adoption of household rules by residents to reduce indoor cannabis smoke, our study is not able to distinguish if the household rules were imposed by the residents of the home or if they were imposed by owners of a rental home or by policies enacted by state or local agencies. This inability to distinguish the source of rules warrants additional investigation in subsequent studies as it may lead to overestimation of the relationship between perceived harm and household rules. The self-reported nature of our data poses another limitation: participants may under-report sensitive information, such as cannabis use, rules related to cannabis use, or use of other (particularly illegal) drugs. Additionally, household rules on in-home cannabis smoking may mediate the relationship between the perception of harm of cSHS exposure and in-home cannabis smoking behavior. Despite limitations, our study’s focused, detailed look at that relationship provides a solid base for future research exploring its mediation by in-home cannabis smoking behavior.