“…The authors acknowledge the inability of their study design to attribute causality and the inability to adjust for potential time‐varying confounders. In addition, it cannot be assumed that the smoking cessation medication prescribed was actually picked up from a pharmacy and used by the consumer, given that past studies have found that 20% of people prescribed NRT do not try to fill it [7]. Other limitations to note include the potential for Type 1 error (given multiple tests of significance), and the large effect sizes and extremely wide confidence intervals observed for the analysis exploring the impact of varenicline prescribing on ‘suicide deaths’ (suggestive of a sparse data issue and thus probably a spurious finding).…”