Table 2.

Summary of analytical studies tackling suicide-related outcomes during major international outbreaks of respiratory viruses.

a. General characteristics and methodology of the included analytical studies.
Study	Type	Country	Outbreak	Methodology			Conflict of interest
				Recruitment strategy	Recruitment site	Eligibility criteria
Lee (2020)	Cross-sectional	United States	COVID-19	Online survey data collected from March 11 to March 13, 2020	Participants recruited through Amazon MTurk in exchange forpayment ($0.50)	Participants were eligible if they provided consent and complete information, followed the directions to a validity item, had spent at least one hour during the past two weeks thinking about and/or watching media about COVID-19, and had experienced significant anxiety, fear, or worry about the disease outbreak.	None
Okusaga et al. (2011)	Cross-sectional	United States	Influenza A, B, and coronavirus	Participants recruited from two studies of environmental influences on mood disorders and suicidal behavior (studies not mentioned within the manuscript)	Participants recruited from the University of Maryland, Johns Hopkins University, and the Sheppard Pratt Health System	Patients were eligible if they met criteria for major depressive or bipolar disorder according to the Structured Clinical Interview for DSM-IV Disorders. Suicide attempts were recorded using the Columbia Suicide History Form.	None
						Patients meeting criteria for substance dependence, cognitive disorders, or primary psychotic disorders were excluded.
Yip et al. (2010)	Cross-sectional	China	SARS	Chart review of death case records in the year 2003	Death records obtained from the Coroner’s Court in Hong Kong	Case records were eligible if they contained suicide notes and witnesses’ descriptions of the suicide deaths (ICD10: X60-X84). Wherever SARS was mentioned as being crucial to the suicide act and was included in the police death investigation, the suicide death was defined as a SARS-related case. Non-SARS-related cases were then randomly selected for comparison.	Not mentioned
Cheung et al. (2008)	Cross-sectional	China	SARS	Chart review of suicide case records for the period 1993–2004	Suicide records obtained from the Coroner’s Court in Hong Kong	No reported inclusion or exclusion criteria.	None
						Case records contained sociodemographic, medical, and psychosocial data gathered from police investigations and medical institutions. Information about suicide was obtained from suicide notes, interviews, and witness reports.
Chan et al. (2006)	Cross-sectional	China	SARS	Chart review of suicide case records for the period 1986–2003	Suicide records obtained from the Census and Statistics Department of the Government of Hong Kong Special Administrative Region	Deaths labeled as “of undetermined cause” were excluded from the analysis.	Not mentioned
Huang et al. (2005)	Cross-sectional	Taiwan	SARS	Retrospective chart review from March 14 to August 31, 2003	Charts obtained from the emergency department of Taipei Veterans General Hospital, a tertiary referral and teaching medical center in northern Taiwan	Patients younger than 14 years were excluded.	Not mentioned
						Patient information was reviewed and compared for different stages of the SARS epidemic (pre-epidemic, early epidemic, peak epidemic, late epidemic, and post-epidemic stages).

b. Outcomes of the included analytical studies.
Study	Sample size	Gender	Age	Suicide-related outcomes	Other relevant outcomes
Lee (2020)	775	446 males		Passive suicidal ideation (measured by the item: “I wished I was already dead, so I did not have to deal with the coronavirus”):	In terms of significant anxiety or worry about the coronavirus during the past two weeks, most participants spent several days feeling elevated anxiety (36.5%), followed by more than seven days feeling elevated anxiety (27.0%), less than a day or two feeling elevated anxiety (22.8%), and nearly every day feeling elevated anxiety (13.7%).
		329 females		1.58 ± 1.46 (reported as mean ± SD)
Okusaga et al. (2011)	257	95 males	32.72 ± 9.35 18 to 65	A more significant percentage of individuals in the suicide attempt group were seropositive for influenza A (p = 0.006), influenza B (p < 0.0001), and coronavirus (p < 0.0001) in comparison to healthy controls.	There were statistically significant associations between seropositivity for influenza A, influenza B, coronaviruses, and the diagnosis of a major depressive disorder.
		162 females		Among individuals with a history of mood disorder, seropositivity for influenza B, but not for influenza A or coronaviruses, was significantly associated with a history of suicide attempt (p = 0.001). The odds of having attempted suicide were increased in influenza B seropositive individuals (OR = 2.53, CI 1.33-4.80).	The odds of having a history of mood disorder were increased with seropositivity for influenza A, influenza B, and coronaviruses.
					Seropositivity for influenza B, but not for influenza A or coronaviruses, was significantly associated with the presence of a history of psychotic symptoms in mood disorder patients.
Yip et al. (2010)	66 (SARS-related: 22, Non-SARS-related: 44)	33 males 33 females (For each: SARS-related: 11, Non-SARS-related: 22)	SARS-related: 74.9 ± 5.72 Non-SARS-related: 74.75 ± 6.77	Sociodemographic factors, employment status, and medical and psychiatric profiles were insignificantly different between the SARS and non-SARS groups.	None
				Disconnection (p = 0.002) and fear of contracting SARS (p<0.001) were significantly more common in the SARS compared to the non-SARS group.
Cheung et al. (2008)	Not applicable	Not applicable	65 and above	There was a decreasing trend of elderly suicide rates from 1993 (39.19 per 100,000) to 2002 (28.44 per 100,000), whereas the rate drastically climbed up to 40.35 per 100,000 in 2003 and maintained a high in 2004 (33.95 per 100,000).	The overall severity of illness (p < 0.001), level of dependency (p < 0.013), and worrying about having sickness (p < 0.021) among the older adult suicides were found to be significantly different in the pre-SARS, peri-SARS, and post-SARS periods.
				The monthly older adult suicide rate in April 2003 was significantly higher than those in April 1993, 1997, 1998, 2001, and 2002. It was also significantly higher than that in June 2003, which suggested the disappearance of the usual summer peak and that some of the suicides might have been brought forward.
				The annual older adult suicide rate in 2003 was significantly higher than in 1996 and 1998-2002. The annual older adult suicide rate in 2004 was also significantly higher than in 2002 (yet lower than in 2003), suggesting that the suicide rate did not return to the level before the SARS epidemic.
Chan et al. (2006)	Not applicable	Not applicable	Stratified as below and above 65	Suicide rates of elders aged 65 and above in 1986–1997 were significantly higher than in 2002, with an Incident Rate Ratio (IRR) of 1.34 to 1.61. Suicide rates in 1998–2001 did not differ from 2002, representing the stabilization of rates for four years after 1997. Elderly suicide rate increased to 37.46/100,000 in 2003, with an IRR of 1.32 relative to 2002 (p < 0.0019).	The peak number of suicides in elders aged 65 and above occurred in April 2003, a month after the SARS outbreak began.
				Such a trend remained significant when female elderly (but not male elderly or the age group below 65) suicide rates in 1993–2003 were analyzed.
Huang et al. (2005)	17,586	Reported percentage in males: 62.7 ± 3.8 to 65.6 ± 3.9, according to epidemic stage	54.6 ± 3.2 to 57.0 ± 3.1, according to epidemic stage	The mean numbers of patients attending the emergency department with a principal diagnosis of suicide attempt via drug overdose were higher during the peak-epidemic stage than all other epidemic stages. However, these differences were not statistically significant.	The total percentage of patients in the “Psychiatric problem/disease” diagnostic category who attended the emergency department did not significantly change across the various stages of the epidemic.