Minimizing double standards in assessing the adverse and beneficial effects of cannabis

Some critics of publicly funded research on cannabis use argue that it has been biased towards finding harms (such as increased risks of car crashes, impaired cognition and educational attainment, a dependence syndrome and psychosis) to justify the prohibition of adult use [1]. Other critics have argued that this research has ignored alternative explanations of associations between cannabis and adverse health effects, such as shared risk factors (e.g. social disadvantage and genetic risks) and reverse causation (e.g. troubled young people using cannabis to treat psychological symptoms) [2, 3].

Since around the turn of the millennium, however, research has focused more upon the putative medical benefits of cannabis that have been used to justify the legalization of its medical use [4]. Some of this research has been of low quality, such as ecological associations between the presence or absence of state medical cannabis programmes and trends in opioid overdose deaths [5], case series [6] and poorly controlled observational studies of patient outcomes in the absence of comparison conditions [7].

In the United States, the legalization of adult cannabis use followed the legalization of medical cannabis use over the next several decades [4]. The arguments that seem to have won public support for adult legalization have been that legalization removes criminal penalties that have been enforced in a discriminatory way against minorities and people of colour; it allows adults to enjoy the effects of cannabis without fear of arrest; it reduces the cost of law enforcement; it creates a new source of tax revenue for government; and it enables cannabis products to be better regulated in the interests of consumer health [4, 8].

These arguments should not preclude research on the harms that may arise from cannabis use, especially the regular use of high-tetrahydrocannabinol (THC) cannabis products, under legalization [4]. Nor should they warrant the use of different evidential standards to evaluate the risks and the benefits of cannabis use; such as, for example, uncritically accepting weak evidence for medical benefits while dismissing epidemiological evidence for causal relationships between cannabis use and adverse health effects as ‘correlational’ (e.g. [5, 9]).

We should also avoid double evidential standards in deciding how to regulate cannabis potency. A popular argument for legalization was that the potency of cannabis products would decline, as high potency varieties and concentrates were thought to be a consequence of the greater risks of trading under legal prohibition. Since legalization in the United States and Canada, however, the THC content of cannabis products has substantially exceeded that in the illicit market, with sales of cannabis extracts, edibles and fortified pre-rolled joints [4]. The average THC content of legal cannabis products is now well above that of the illicit cannabis used by participants in earlier epidemiological studies of the adverse health effects of cannabis [10].

In the face of emerging evidence of increased harm from the use of high-potency cannabis products [10], the legal cannabis industry in the United States strenuously opposes proposals to regulate or differentially tax high-THC products [11]. It does so by setting demands for high standards of evidence to justify the regulation of THC content, with the implied default being no regulation of the THC content of cannabis products [11].

How can we obtain fairer appraisals of the benefits and the harms of cannabis use to improve cannabis regulation in a legal market?

Evidence for medical uses of cannabis should be provided by randomized controlled clinical trials. These study designs reduce the plausibility of alternative explanations of patient improvements seen in uncontrolled studies, such as placebo effects and variations in the severity of a chronic illness or disorder over time. Regulatory and financial barriers that prevent trials of medical cannabis products still need to be removed in many countries [12].

The evidence from clinical trials should be supplemented by well-controlled observational studies that assess whether the benefits in clinical trials reliably translate into routine clinical practice in more representative samples of patients than those participating in clinical trials. Observational studies should not, however, be accepted as sufficient evidence to justify the widespread medical use of cannabis, or indeed any other drug, because of the major inferential problems in interpreting such data [13].

Journals publishing clinical trials and observational studies will need to pay attention to conflicts of interest that may rise from cannabis industry funding. They should apply the same approaches that have become standard in disclosing and managing conflicts of interest in trials funded by the pharmaceutical industry [14]. Research teams could also include members whose a priori views differ on the topic under investigation.

Research on the harms of cannabis use will primarily come from observational epidemiological studies, because ethical issues preclude experimental studies of the harms of regular and long-term cannabis use in humans. We already know a considerable amount regarding one harm that is probably underappreciated among cannabis users: the risk of developing a cannabis use disorder (CUD). A recent systematic review indicated that the risk was higher than older estimates from studies in the early 1990s, when less potent cannabis was used [15]. The prevalence of CUD has also increased among US adults since the 1990s when medical cannabis was first legalized [16].

People with CUDs have higher rates of depression, cognitive impairment and psychoses [17]. Large epidemiological studies of representative samples of adults are needed to evaluate the extent to which these associations are causal. These studies will need to include substantial numbers of people who use cannabis, with varying frequencies and durations of use, and build upon earlier studies in New Zealand and the Netherlands and the ABC study that is under way in the United States. Self-reports of cannabis use in these newer studies may be more accurate when adult use is legal, and it should become easier to collect self-report data on the typical doses of THC and CBD that are used when legal cannabis products have their cannabinoid content labelled [4].

Triangulation of evidence from different types of studies with varying strengths and weaknesses will be very helpful in learning more about cannabis-related health risks [18]. The reduced costs of genotyping and GWAS findings, for example, will allow researchers to use Mendelian randomization to test causal explanations of associations. Animal experiments designed to have greater relevance to patterns of human cannabis use may also contribute. Economic analyses of legal sales data on alcohol, tobacco and cannabis will be able to more effectively assess the extent to which cannabis is a substitute or a complement: an important issue in assessing the aggregate public health impact of cannabis legalization.

Cannabis legalization is spreading beyond the Americas. Cannabis has been legalized in Thailand and the governments of Czechia, Germany, Luxembourg, the Netherlands and Switzerland propose to legalize adult cannabis use in some form. The growing legalization bandwagon increases the urgency of good-quality research on both the benefits and adverse health effects of cannabis. The latter is especially needed if we want to avoid the eclipse of knowledge about the adverse effects of alcohol that followed the repeal of national prohibition in the United States in 1932 [19].

We should take advantage of the greater ease of conducting rigorous research on the benefits and harms of cannabis use after legalization. This needs to be performed in ways that minimize the undue influence that commercial interests have had on research on the adverse effects of alcohol and tobacco. The goal should be to advise governments on how to regulate legal cannabis markets in ways that minimize the harms of the regular use of high-potency products while maximizing the social and medical benefits of adult use under cannabis legalization.

Wayne Hall: Conceptualization (equal); data curation (equal); writing—original draft (equal); writing—review and editing (equal). Eva Hoch: Conceptualization (equal); data curation (equal); writing—original draft (equal); writing—review and editing (equal).

No competing financial interests.