Table 1 summarizes the characteristics of the 46 studies. A total of 27 studies used suicide counts as the outcome of interest [13, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43]. Of them, the majority examined monthly suicide counts, except for two reports from Nepal not specifying the time window [18, 37], a study examining daily suicides in Maryland, US [21], two studies conducted in Peru assessing biweekly suicide counts [22, 23], and studies examining periods longer than one month [27, 28, 34, 38, 39, 41]. In 20 studies, the outcome of interest was measured as suicide rate – and expressed as monthly suicide rate [29, 44,45,46,47,48,49,50,51,52,53,54], annual suicide rate [55,56,57,58], or suicide rate using a different time window [59,60,61,62].

There was between-study variation in the definition of the COVID-19 period, largely due to geographical variation in the timing of the first local case of SARS-CoV-2. While 22 studies defined March 2020 as the beginning of the pandemic, [18, 19, 21,22,23, 27,28,29, 34, 36,37,38, 40, 43, 46, 50, 51, 54, 58,59,60,61], 6 studies [13, 20, 24, 44, 47, 52] – including Pirkis et al.’s study featuring data from 21 countries [13], used April 2020, 11 studies used January 2020 [25, 30, 31, 35, 39, 41, 42, 45, 56, 57, 62] and 3 studies February 2020 [48, 49, 53]. The studies by Eguchi et al. [26] and Nomura et al. [32] did not report a specific COVID-19 period.

There was substantial heterogeneity in choice of statistical approach to estimate the difference between expected and observed suicide counts or rates. In general, all approaches aimed at estimating the counterfactual outcome (e.g., monthly suicide count, or monthly suicide rate) had the COVID-19 pandemic outbreak not taken place – in order to then compare expected vs. observed outcomes. The majority of studies adopted an ITSA approach based on some specification of a segmented regression – e.g., Poisson [13, 19, 20, 23, 24, 29, 40, 41, 46, 48, 50, 55, 61], quasi-Poisson [26, 30, 32, 34, 36], negative binomial [31, 42, 51], linear [