Objective <p>To conduct a detailed epidemiological exploration of the relative contributions of cannabis and ethnicity to US atrial septal defect (ASD) rates (ASDR).</p> Study design <p>State-based ASDR data from the US National Births Defects Prevention Network, substance use, income and ethnicity data analyzed in RStudio.</p> Results <p>Ethnic effects were significant with relative risks amongst African Americans and American Indians and Alaskan Natives of 2·40 (95%C.I. 2·27, 2·54) and 2·31 (2·19, 21·43), Cohen’s D of 1·44 and 1·46 and <i>P</i> values of 2·94 × 10⁻¹⁶⁸ and 3·01 × 10⁻¹⁷² compared to others, respectively. In general, additive models inclusion of ethnicity:cannabinoid and ethnicity:tobacco interactions were significant down to P=zero for cannabis, Δ9THC and cannabidiol. Sequentially doubly robust targeted multiple likelihood estimations confirmed epidemiologically causal relationships under standard assumptions. ASDR amongst Asians and Pacific Islanders in Nevada showed an exponential doubling time of 2.83 years.</p> Conclusions <p>Cannabinoid and cannabinoid:ethnicity interactions drive ASDR and meet epidemiological causal criteria.</p>