It's important to point out the failings of this study, since the media were all over it! The first major failing is its retrospective design. Retrospective studies, by nature, rely on existing data. This introduces selection bias, as the data wasn’t collected with this specific research question in mind.
Another potential issue is the definition and identification of cannabis use. If the study uses diagnostic codes (e.g., ICD-10 for the purported ‘cannabis use disorder’) rather than detailed patient histories, it might misclassify casual users as non-users or lump all users into one category, ignoring dose, frequency, or method of use. This lack of granularity weakens any causal inference about how cannabis specifically contributes to myocardial infarction (MI) or cardiovascular risks.
Confounding variables are a third likely failing. Cannabis users often have higher rates of tobacco use, alcohol consumption, or other lifestyle factors (poor diet, sedentary behaviour) that independently increase cardiovascular risk. While a good study adjusts for these, retrospective data might not provide enough detail to control for all confounders adequately. For example, if smoking tobacco and cannabis are conflated, the observed risk might be overstated and misattributed to cannabis alone.
Causality is another weak spot. The title suggests an association between cannabis use and MI/cardiovascular risks, but retrospective studies can’t establish cause-and-effect. Temporal relationships might be unclear—did cannabis use precede the MI, or did patients start using it afterward (e.g. for pain or stress)? Without prospective tracking, this remains speculative, and alternative explanations (e.g. underlying stress or genetic predisposition) can’t be ruled out.
In short, this study’s potential failings include its retrospective bias, imprecise cannabis use classification, inadequate confounding control, data heterogeneity, inability to prove causality and limited generalisability.