Figure 2.

Adjusted logistic regression of non-medical and medical cannabis frequency of use before the pandemic on change in mode of cannabis use during the pandemic among a cannabis using population, August-September 2020. Data were collected on those who reported cannabis and/or cannabidiol use in the United States in a single web-based survey to assess cannabis behaviors before and during the COVID-19 pandemic. Multivariable logistic regressions were conducted to assess the association of cannabis behaviors before the pandemic to changes in behaviors during the pandemic. Figure 2 shows cannabis use before the COVID-19 pandemic on change in mode of use during the COVID-19 pandemic. Change in mode of use was dichotomized as change from inhalation to oral or did not change from inhalation to oral (reference). Models for non-medical and medical cannabis use were conducted separately (denoted with a * in Figure 2). Adjusted odds ratio and 95% confidence intervals (denoted as 95% CI in Figure 2) are presented and displayed in a forest plot. State’s cannabis regulatory status, participant’s age, and participant’s sex were adjusted for in the models. Age was recentered at mean age and rescaled per 10-unit increase (denoted as Age** in Figure 2). Sample sizes for the models were: non-medical cannabis (n = 894); medical cannabis (n = 406).