The findings from this study illustrate the challenges in estimating the prevalence of medical cannabis use within and between jurisdictions given different ways of identifying it. Jurisdictions vary in the conditions and products that can be used for medicinal purposes [13]. Hence, even standardized questions will be interpreted through these jurisdictional differences, as will the ways in which medical cannabis is accessed. This study shows clearly that in these three jurisdictions people who use cannabis therapeutically differed in their engagement with healthcare providers and their use of pharmaceutical‐grade products. While we consistently find that a relatively small share of people who self‐identify as medical users or acknowledge using cannabis for therapeutic purposes engage with a medical provider, it was even less common in Canada where cannabis is also available recreationally [14].

The findings from this study suggest that if researchers focus on patients who engage with a medical provider about medical cannabis use, this will capture only a small proportion of people who self‐identify as a medical cannabis user or use cannabis therapeutically in every jurisdiction. Despite the regulatory differences between countries, self‐identified medical cannabis use represented 21.3%–36.2% of past‐year use within each country. As in prior research, we found a far greater proportion of the population reported therapeutic use rather than accessing medical cannabis via regulated pathways, even when legally allowed to do so [10, 14]. Studies focused on medically supervised populations (i.e. those who use cannabis through the traditional healthcare system) will accordingly miss a large share of individuals who use cannabis for medical reasons. These variations across measures within populations occur despite the requirement in each jurisdiction for consultation with a medical provider.

The study also revealed that the extent to which the medical use of cannabis overlaps with recreational use depends on how medical use is defined, regardless of country. There was substantial overlap (>70%) between therapeutic use and past‐year recreational use in all three countries, with a substantially lower degree of overlap (<30%) between potentially authorized, Rx use or self‐identified medical use and past‐year use. Differences in federal policies toward medical use and adult use of cannabis are likely to influence the proportion of respondents who reported seeking medical authorization or who self‐identified as a medical user. These differences might also reflect underlying population differences in the willingness to engage with the healthcare system or try pharmaceutical therapies. For example, Australians were far more likely to use a pharmaceutical‐grade product (27.4%) than Americans (13.5%) or Canadians (8.8%).

This study has several limitations. First, the ICPS sample is derived from the Nielsen commercial marketing panels that represent household purchasers, not cannabis consumers specifically. Moreover, they were obtained using both probability‐ and non‐probability‐based sampling due to reliance on Nielsen consumer panels, which could limit the generalizability. For example, the ICPS has a greater proportion of female participants. However, the weighted ICPS results produced estimates of past‐year cannabis use that were generally consistent with that in nationally representative samples of households in Australia, Canada and the USA [8, 12]. Finally, this is a large international study and policy differences may have affected individual responses to specific questions; jurisdictional differences in marketing restrictions, compliance and enforcement may influence motivations for using cannabis medically.