A preliminary analysis published in The Lancet Psychiatry of 21 high- and upper middle-income countries indicates that suicide numbers have not significantly increased in the early months of the pandemic compared with expected levels based on the pre-pandemic period.
Internet searches for official government and publicly reported data on World Bank countries and economies with more than 3 million residents (n=135) were completed from September through November 2020. Data (including from countries and regions with fewer than 3 million residents) were also derived from studies in academic literature that reported on suicides before and after the pandemic as well as the World Health Organization (WHO), the International Association for Suicide Prevention, and International COVID-19 Suicide Prevention Research Collaboration (ICSPRC) members, who represent 40 countries.
They sourced government statistics from 16 high-income countries (Australia, Austria, Canada, Chile, Croatia, England, Estonia, Germany, Italy, Japan, Netherlands, New Zealand, Poland, South Korea, Spain, and the United States) and 5 upper-middle-income countries (Brazil, Ecuador, Mexico, Peru, and Russia).
The study authors used interrupted time-series analysis to model trends in monthly suicides before COVID-19 in each included geographical area, accounting for time trends and seasonality when possible.
They found no evidence of an increase in suicides relative to the expected number during the COVID-19 period in any geographical area in the primary analysis, in which they examined data from April through July 2020. They found statistical evidence of decreases in suicides in 12 areas: New South Wales, Australia (rate ratio (RR) 0.81 [95% CI 0.72–0.91]); Alberta, Canada (0.80 [0.68–0.93]); British Columbia, Canada (0.76 [0.66–0.87]); Chile (0.85 [0.78–0.94]); Leipzig, Germany (0.49 [0.32–0.74]); Japan (0.94 [0.91–0.96]); New Zealand (0.79 [0.68–0.91]); South Korea (0.94 [0.92–0.97]); California, USA (0.90 [0.85–0.95]); Illinois (Cook County), USA (0.79 [0.67–0.93]); Texas (four counties), USA (0.82 [0.68–0.98]); and Ecuador (0.74 [0.67–0.82]).
In the first sensitivity analysis, the researchers included all data available in each area through October 2020. They identified significant decreases in 3 additional areas: Victoria, Australia (0.89 [0.80−0.99]); Thames Valley, England, UK (0.82 [0.68−0.98]); and Mexico City, Mexico (0.86 [0.77−0.97]). However, Vienna showed statistical evidence of an increase in suicides (1.31 [1.08–1.59]), as did Japan (1.05 [1.04–1.07]), and Puerto Rico (1.29 [1.05–1.58]).
The second sensitivity analysis, which identified the pandemic period as March 1 through July 2020, indicated a decreased risk of suicide in several additional areas beyond the primary analysis: Manitoba, Canada (0.60 [0.48−0.76]); Poland (0.94 [0.90−0.98]); Las Palmas, Spain (0.69 [0.51−0.94]); and Peru (0.73 [0.64−0.83]). The sole increase was seen in Puerto Rico (1.36 [1.07–1.72]).
When the researchers inflated the suicide numbers in the COVID-19 period by 5% for a supplementary analysis, New Jersey, USA (RR 1.18 [95% CI 1.05–1.34]) and Puerto Rico (1.34 [1.03–1.74]) showed increases.
Limitations of the study included the exclusion of low-income or lower-middle-income countries due to limited data, that low numbers of time points and monthly suicides in given areas might have resulted in models with relatively poorer precision, and that the researchers could not stratify data by demographics.
“We need to understand what has kept suicide numbers down during the pandemic’s early months, and what drives any increases if they do occur,” the study authors said. “We also need to recognize that suicide is not the only indicator of the negative mental health effects of the pandemic; levels of community distress are high and we need to ensure that people are supported.”
Pirkis J, John A, Shin S, et al. Suicide trends in the early months of the COVID-19 pandemic: An interrupted time-series analysis of preliminary data from 21 countries. Lancet Psychiatry. Published online April 13, 2021. doi: 10.1016/S2215-0366(21)00091-2