Cannabis legalisation and cardiovascular hospitalisations in the United States: nationwide difference-in-differences analysis, 2012-2018

This study examines a large sample of over 50 million hospitalisations that are representative of all US Census regions.

The difference-in-differences analysis of cannabis policy mitigates the effect of individual-level confounders that are common in other observational analyses.

The difference-in-differences model is still susceptible to population-level confounding by secular trends that may have coincidentally begun to diverge at the time of legalisation.

Due to limitations of the National Inpatient Sample, our design treats the entire Western region as exposed from the first date of recreational legalisation, possibly biasing results towards the null.

While this study had adequate power to detect changes in hospitalisation rates for myocardial infarction, heart failure and atrial fibrillation, it may have lacked power to detect changes in less frequent diagnoses.