John Strang

Background and aim The rate of drug poisoning (or overdose) deaths in England and Wales has risen annually since 2010. We aimed to measure seasonal and other cyclical changes in these deaths within years. Methods We used the daily count of deaths due to drug poisoning in England and Wales between 1 January 1993 and 31 December 2018 to investigate variation by season, weekday, week‐of‐month and public holiday. We used Poisson regression to estimate the count of deaths per day for each of these variables and peak‐to‐low ratios. We also stratified the analysis by time period and whether an opioid was mentioned on the death certificate. Results 78 583 deaths occurred between 1993 and 2018, increasing from 5.50 (95% confidence interval [CI] = 5.24–5.77) per day in 1993 to 13.18 (95% CI = 12.66–13.72) per day in 2018. The rate peaked in Spring and was 1.07 (95% CI = 1.04–1.09) times higher in April than in October. This seasonal pattern emerged in the past decade and was only present for opioid‐related deaths. The rate at New Year was 1.28 (95% CI = 1.17–1.41) times higher than on non‐holidays; and this peak was only present for deaths that were not related to opioids. The rate was higher on Saturday than on other weekdays. We did not find evidence that the number of deaths varied by week‐of‐month. Conclusions Deaths due to drug poisoning in England and Wales are seasonal and peak in Spring and briefly at New Year. This suggests a role of external triggers. These seasonal variations are small compared with long‐term increases in drug‐related deaths.