Table 1.

Select meta-analyses and large cohort studies examining associations between mental health and substance use disorders and suicide

Source	Study design	Inclusion criteria	Exposure	Measure of Association (95% CI)	Effect Measure
Meta-analyses
Harris et al. 1997 (11)	International Meta-analysis 249 studies	• Cohort must have a defined disorder with two years of follow up  • Less than 10% lost to follow up  • Published observed suicide number and expected number, or sufficient information to estimate the expected number	Major depression	20.35 (18.27, 22.58)	SMR
			Opioid use	14.00 (10.79, 17.88)
			Any SUD	5.47 (5.41, 6.09)
			Anxiety neurosis	6.29 (5.33, 7.38)
			OCD	11.54 (2.38, 33.72)
			Panic disorder	10.00 (4.57, 18.98)
Chesney et al. 2014 (12)	International Meta-review 20 studies	• Included the most recent meta-analysis or review, or if not available then the most recent large (N>1000) study reporting risks for suicide	Depression	19.7 (12.2, 32.0)	SMR
			Opioid use	13.5 (10.5, 17.2)
			Anxiety disorders	3.3 (2.1, 5.3)	OR
Too et al. 2019 (15)	International Meta-analysis of linkage studies 13 studies	• Comparisons of individuals with and without mental disorders based on a general population sample	Mood disorders	12.3 (8.9, 17.1)	IRR
			SUDs	4.4 (2.9, 6.8)
			Anxiety disorders	4.1 (2.4, 6.9)
Conner et al. 2019 (40)	Meta-analysis: Psychological autopsies 35 studies	• Living or dead controls  • Descriptive data on mood disorders and/or alcohol or drug use disorders  • In-person proxy interviews using a diagnostic instrument  • Controls can be interviewed or a proxy can be interviewed for controls  • Must include all suicides in a population or an age, sex, and/or location subgroup	Major depression	26.87 (15.58, 46.36)	OR
			Asia Europe North America	7.95 (3.30, 19.16) 7.68 (2.60, 22.70)
			SUDs
			Asia Europe North America	3.35 (2.00, 5.61) 6.54 (3.76, 11.39) 3.97 (1.99, 7.90)
Large Cohort Studies
Mortensen et al. 2000 (21)	Retrospective longitudinal record linkage case-control study (1980–1994)	• Population: Danish psychiatric inpatients  • 811 cases, 79,871 controls  • Exposure ascertainment: Psychiatric diagnosis (ICD-9) at most recent inpatient discharge from the Danish Psychiatric Central Register	Reactive psychosis	1.80 (1.28, 2.53)	IRR
			Drug abuse	1.08 (0.82, 1.42)
			Other psychiatric disorder	1 (reference)
Sonderman et al. 2014 (73)	Prospective longitudinal cohort survey-based study (2002–2009)	• Population: Black and White residents in Southeastern US  • N=85,000  • Exposure ascertainment: “Has a doctor ever told you that you have depression or have you been treated for depression?”	Depression	3.05 (1.70, 5.48)	HR
Boggs et al. 2018 (29)	Retrospective longitudinal health-systems-based case-control study (2010–2013)	• Population: 8 US health systems (Mental Health Research Network)  • 2,674 cases (1,298 firearm cases), 267,400 controls  • Exposure ascertainment: ICD-9 codes within the previous 12 months of suicide for cases and matched index date for controls	Any psychiatric disorder		OR
			Firearm Other	5.51 (4.92, 6.14) 9.04 (8.06, 10.15)
			Anxiety disorder	6.53 (5.68, 7.50)
			Firearm Other	7.74 (6.84, 8.75)
			Depression
			Firearm Other	7.29 (6.46, 8.23) 12.28 (10.96, 13.74)

OR=odds ratio; SMR=standardized mortality ratio; IRR=incidence rate ratio; SUD=substance use disorder; OCD=obsessive compulsive disorder