Patterns and correlates of workplace and non-workplace cannabis use among Canadian workers before the legalization of non-medical cannabis

•A non-trivial proportion of workers use cannabis before or at work.•Workers reporting workplace use frequently report use of high THC products.•Workers reporting workplace use also frequently use for medical or mixed purposes.•Work factors related to likelihood of detection are associated with workplace use.•Having a safety-sensitive job or supervisor role are associated with workplace use. Little information exists about cannabis use and its correlates among workers, particularly use before or at work, which may impact occupational safety. This study explores overall and workplace cannabis use patterns before legalization among Canadian workers and estimates the associations of personal and work-related characteristics with workplace and non-workplace cannabis use. Cross-sectional data were collected from 1651 Canadian workers in June 2018. The primary outcome was past-year cannabis use pattern: use, including before/at work (past-year workplace use); use, but not before/at work (past-year non-workplace use); no past-year use (non-past-year use). The associations of personal (sociodemographic, health) and work-related factors with workplace and non-workplace cannabis use were estimated using multinomial logistic regression. A quarter of respondents reporting past-year cannabis use used cannabis before and/or at work. Respondents reporting workplace use were more likely to report more frequent cannabis use, use for medical or mixed purposes, and high THC cannabis use than workers reporting non-workplace use. Several personal factors were positively associated with workplace and non-workplace use (e.g., younger age, lower education). A safety-sensitive job, drug testing, supervisor role, less job visibility, lower perceived ability of supervisors to identify use/impairment, and less restrictive workplace smoking policies were positively associated with workplace use only. A non-trivial proportion of workers reported workplace use, but the nature of this use is complex. Work-related factors addressing the likelihood of detection and being in a safety-sensitive job were associated with workplace use. Worker education on lower risk use appears warranted.