CoChat
Sign Up
Cited by 336Suicide and Self-Harm StudiesHomicide, Infanticide, and Child AbuseGrief, Bereavement, and Mental Health

Risk factors for suicide in adults: systematic review and meta-analysis of psychological autopsy studies

Open access
Favril, Louis, Yu, Rongqin, Uyar, Abdo, Sharpe, Michael, Fazel, Seena·2022·Evidence Based Mental Health
DOI 10.1136/ebmental-2022-300549PMC PMC9685708
Open this paper in CoChat — free accountOpen PDF

Abstract

Question: Effective prevention of suicide requires a comprehensive understanding of risk factors. Study selection and analysis: Five databases were systematically searched to identify psychological autopsy studies (published up to February 2022) that reported on risk factors for suicide mortality among adults in the general population. Effect sizes were pooled as odds ratios (ORs) using random-effects models for each risk factor examined in at least three independent samples. Findings: A total of 37 case–control studies from 23 countries were included, providing data on 40 risk factors in 5633 cases and 7101 controls. The magnitude of effect sizes varied substantially both between and within risk factor domains. Clinical factors had the strongest associations with suicide, including any mental disorder (OR=13.1, 95% CI 9.9 to 17.4) and a history of self-harm (OR=10.1, 95% CI 6.6 to 15.6). By comparison, effect sizes were smaller for other domains relating to sociodemographic status, family history, and adverse life events (OR range 2–5). Conclusions: A wide range of predisposing and precipitating factors are associated with suicide among adults in the general population, but with clear differences in their relative strength. PROSPERO registration number: CRD42021232878.

AI snapshot

Written from the paper’s full text.

TL;DR

This systematic review and meta-analysis of 37 psychological autopsy case–control studies from 23 countries found that clinical factors, especially any mental disorder and a history of self-harm, had the strongest associations with adult suicide. Risk factors across sociodemographic, family history, and adverse life event domains were also associated with suicide but generally had smaller effect sizes.

Key findings

  • Across 40 risk factors in 5,633 suicide cases and 7,101 controls, the strongest associations were any mental disorder (OR=13.1, 95% CI 9.9 to 17.4) and a history of self-harm (OR=10.1, 95% CI 6.6 to 15.6); previous suicide attempt had OR=8.5 (95% CI 5.3 to 13.4).
  • Within the clinical domain, depression had OR=11.0 (95% CI 7.3 to 16.5), borderline personality disorder OR=9.0 (95% CI 5.6 to 14.4), schizophrenia spectrum disorder OR=7.8 (95% CI 4.5 to 13.5), alcohol use disorder OR=3.2 (95% CI 2.3 to 4.4), and drug use disorder OR=3.0 (95% CI 1.7 to 5.4).
  • Sociodemographic factors included social isolation (OR=4.0, 95% CI 2.1 to 7.7), unemployment (OR=3.8, 95% CI 2.7 to 5.2), and low socioeconomic status (OR=2.8, 95% CI 1.8 to 4.2); family history factors included family history of mental disorder (OR=5.2, 95% CI 1.9 to 14.1), suicide (OR=3.7, 95% CI 2.3 to 5.7), and suicide attempt (OR=2.8, 95% CI 1.5 to 5.0).
  • Adverse life events included relationship conflict (OR=5.0, 95% CI 3.3 to 7.6), legal problems (OR=4.8, 95% CI 2.4 to 9.4), and family-related conflict (OR=4.5, 95% CI 2.0 to 10.3); events within the past month had OR=10.4 (95% CI 7.1 to 15.3). At death, 71.1% of cases had a mental disorder versus 22.2% of controls, and 28.5% of cases versus 6.0% of controls had a previous suicide attempt.

Methods

  • Systematic review and meta-analysis of psychological autopsy studies identified from five databases searched to 28 February 2022, with no language restrictions; 37 case–control studies from 23 countries were included, covering 5,633 suicide cases and 7,101 controls, predominantly adults aged 18–65 years in general population samples.
  • Crude odds ratios were pooled using random-effects models for risk factors examined in at least three independent samples; risk factors were grouped into sociodemographic, clinical, family history, and adverse life event domains. Heterogeneity was assessed with I², meta-regression was used for risk factors with at least 10 samples, and publication bias was examined with funnel plots and Egger’s test.
  • Risk of bias was assessed with the Newcastle-Ottawa Scale; sensitivity analyses compared crude and adjusted estimates, and post hoc subgroup/sensitivity analyses examined control group type and studies in which controls acted as their own informants.

Limitations

  • The paper states that associations are likely overestimated because the main analysis used crude estimates and did not account for confounding; adjusted estimates were smaller for most factors, with an average reduction of 19% and up to half for clinical risk factors, but inconsistent adjustment across studies limited comparability.
  • The paper states considerable heterogeneity for about one-third of risk estimates, so pooled effects should be interpreted cautiously; it also retained all eligible studies irrespective of methodological quality, could not examine men and women separately, and findings should not be generalised to adolescents or older adults.
  • The paper notes psychological autopsy limitations including control selection/selective non-response and proxy-report recall bias or measurement error; evident from the text, meta-regression found smaller effect sizes in studies with deceased controls for 7 of 21 risk factors, and Egger’s test indicated publication bias for 8 risk factors.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1136/ebmental-2022-300549

Cites

42 of 42 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 64 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Dysthymia is associated with suicide in adults in the general population (pooled OR=2.4, 95% CI 1.4 to 4.1).

    “Dysthymia | 6 | 964 | 1566 | 2.4 (1.4 to 4.1) | 3.2 | 36”
    Quote in the paperSubstantial claim
  • Findingown data

    Family-related conflict is associated with suicide in adults in the general population (pooled OR=4.5, 95% CI 2.0 to 10.3).

    “Family-related conflict | 9 | 1154 | 1364 | 4.5 (2.0 to 10.3) | 3.6 | 92”
    Quote in the paperSubstantial claim
  • Findingown data

    Previous suicide attempt is associated with suicide in adults in the general population (pooled OR=8.5, 95% CI 5.3 to 13.4).

    “Previous suicide attempt | 22 | 2971 | 4107 | 8.5 (5.3 to 13.4) | 9.1 | 77”
    Quote in the paperSubstantial claim
  • Findingown data

    A total of 37 studies reported in 97 publications met inclusion criteria.

    “A total of 37 studies reported in 97 publications met our inclusion criteria (table 1).”
    • Also stated in: § Findings · plain
      “A total of 37 case–control studies from 23 countries were included, providing data on 40 risk factors in 5633 cases and 7101 controls.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Avoidant personality disorder is associated with suicide in adults in the general population (pooled OR=3.9, 95% CI 1.4 to 11.1).

    “Avoidant PD | 5 | 526 | 701 | 3.9 (1.4 to 11.1) | 2.5 | 69”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 41
  • May combine two claims. 33

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    All but three (92%) studies included both men and women.

    “All but three (92%) studies included both men and women.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Eight (22%; n=3540) studies used a control group of people who died from causes other than suicide.

    “Twenty-nine (78%; n=9194) studies recruited living community controls and eight (22%; n=3540) used a control group of people who died from causes other than suicide (mostly accidents and sudden deaths).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The mean age of the suicide group was 39.8 years (SD=10.9).

    “The mean age of the suicide group was 39.8 years (SD=10.9) and 71% were men.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Twenty-nine (78%; n=9194) studies recruited living community controls.

    “Twenty-nine (78%; n=9194) studies recruited living community controls and eight (22%; n=3540) used a control group of people who died from causes other than suicide (mostly accidents and sudden deaths).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Collectively, studies included 12,734 individuals, of which 7101 (56%) were controls and 5633 (44%) were suicide cases.

    “Collectively, studies included a total of 12 734 individuals, of which 7101 (56%) were controls and 5633 (44%) were suicide cases, with an average of 152 suicides per study.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    231 full-text reports were examined for eligibility.

    “Our systematic search of the literature yielded 4284 unique records for screening, from which 231 full-text reports were examined for eligibility (online supplemental figure S1).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    71% of the suicide group were men.

    “The mean age of the suicide group was 39.8 years (SD=10.9) and 71% were men.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    21 (57%) studies were in high-income countries.

    “Studies were conducted across 23 countries (21 [57%] in high-income countries) and published between 1994 and 2021.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In five (14%; n=2419) studies, information on the living control group was collected directly from (a portion of) the control subjects themselves, rather than by interviewing proxy informants.

    “In five (14%; n=2419) studies, information on the living control group was collected directly from (a portion of) the control subjects themselves, rather than by interviewing proxy informants.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Studies were conducted across 23 countries.

    “Studies were conducted across 23 countries (21 [57%] in high-income countries) and published between 1994 and 2021.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Studies were published between 1994 and 2021.

    “Studies were conducted across 23 countries (21 [57%] in high-income countries) and published between 1994 and 2021.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Sample sizes ranged from 45 to 1596 (median=216, IQR=194-400).

    “All included studies had a case–control design, with sample sizes ranging from 45 to 1596 (median=216, IQR=194–400).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The systematic search yielded 4284 unique records for screening.

    “Our systematic search of the literature yielded 4284 unique records for screening, from which 231 full-text reports were examined for eligibility (online supplemental figure S1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    71.1% of people who died by suicide had a mental disorder at the time of death, compared with 22.2% of controls.

    “Seven in 10 (71.1%) cases had a mental disorder at the time of death compared with 22.2% of controls.”
    • Also stated in: § Conclusions and implications · plain
      “Absolute rates were also high, with 71% of suicide decedents reported to have a mental disorder at the time of death and 29% having previously attempted suicide.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Egger's test was significant for social isolation (p=0.001), unemployment (p=0.015), living alone (p=0.044), any mental disorder (p=0.009), depression (p=0.071), substance use disorder (p=0.006), any personality disorder (p=0.006), and financial problems (p=0.061).

    “Egger’s test was significant for social isolation (p=0.001), unemployment (p=0.015), living alone (p=0.044), any mental disorder (p=0.009), depression (p=0.071), substance use disorder (p=0.006), any personality disorder (p=0.006) and financial problems (p=0.061).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Income region was a moderator only for the association between depression and suicide, with a smaller effect observed in samples from high-income countries.

    “Income region was a moderator only for the association between depression and suicide, with a smaller effect observed in samples from high-income countries.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Adjusted odds ratios were smaller than crude odds ratios for all but three of 21 risk factors examined, with an average reduction of 19% (range +48% to −61%).

    “Adjusted ORs were smaller than crude ORs for all but three factors examined, with an average reduction of 19% in risk estimates (ranging from +48% to −61%).”
    • Also stated in: § Strengths and limitations · plain
      “In sensitivity analyses, we found that the majority of adjusted estimates were materially smaller than crude ones, with an average reduction of a fifth in effect sizes and up to half for clinical risk factors.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Many included studies exclusively reported significant associations from multivariable models.

  • May combine two claims.

    Findingown data

    The crude OR for a history of self-harm was 18.1 (95% CI 13.2 to 24.6), about two times larger than its adjusted OR of 9.1 (95% CI 6.0 to 13.7).

    “For example, the crude OR for a history of self-harm (OR=18.1, 95% CI 13.2 to 24.6) was two times larger than its adjusted counterpart (aOR=9.1, 95% CI 6.0 to 13.7).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Suicide is associated with many factors across clinical, life events, family history, and sociodemographic domains, and within a particular domain there are differential associations with suicide.

    “The main implication of this review is that suicide is associated with many factors across clinical, life events, family history, and sociodemographic domains—and within a particular domain, there are differential associations with suicide.”
    • Also stated in: § Conclusion · plain
      “We found that a wide range of predisposing and precipitating risk factors are associated with suicide among adults in the general population, but to differing degrees.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Psychological autopsy is a widely used approach to investigating suicide risk factors at the individual level.

    “One widely used approach to investigating suicide risk factors at the individual level is the psychological autopsy method,3”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [3] source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Previous reviews of psychological autopsy studies have been limited in scope (primarily focusing on mental disorders) and samples (including selected populations such as psychiatric patients).

    “To our knowledge, five previous reviews have quantitatively summarised findings from psychological autopsy studies,5–9 but these have been limited in scope (primarily focusing on mental disorders) and samples (including selected populations such as psychiatric patients).”
    Quote in the paperSubstantial claim0 of 5 sources checked
    • cites [5] source

      Not checked in the free run.

    • cites [6] source

      Not checked in the free run.

    • cites [7] source

      Not checked in the free run.

    • cites [8] Cho, 2016, Geographical and temporal variations in the prevalence of mental disorders in suicide: systematic review and meta-analysis, J Affect Disord source

      Not checked in the free run.

    • cites [9] Conner, 2019, Metaanalysis of mood and substance use disorders in proximal risk for suicide deaths, Suicide Life Threat Behav source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    Identifying factors associated with suicide can improve risk stratification and help target interventions for high-risk groups.

    “Identifying factors associated with suicide can improve risk stratification and help target interventions for high-risk groups.2”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Premisenone given

    Synthesising the full range of risk factors reported in the psychological autopsy literature by adopting a uniform analytical approach would allow for a direct comparison of their associations with suicide.

    “Synthesising the full range of risk factors reported in the psychological autopsy literature by adopting a uniform analytical approach would allow for a direct comparison of their associations with suicide.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The inclusion of studies with controls who died by causes other than suicide is consistent with previous reviews of psychological autopsy studies.

    “The inclusion of studies with controls who died by causes other than suicide is consistent with previous reviews of psychological autopsy studies.5–9”
    Quote in the paperNot a substantial claim0 of 5 sources checked
    • cites [5] source

      Not checked in the free run.

    • cites [6] source

      Not checked in the free run.

    • cites [7] source

      Not checked in the free run.

    • cites [8] Cho, 2016, Geographical and temporal variations in the prevalence of mental disorders in suicide: systematic review and meta-analysis, J Affect Disord source

      Not checked in the free run.

    • cites [9] Conner, 2019, Metaanalysis of mood and substance use disorders in proximal risk for suicide deaths, Suicide Life Threat Behav source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Post hoc subgroup analyses were conducted by type of control group and sensitivity analyses excluding studies in which controls acted as their own informants.

    “Post hoc, we conducted subgroup analyses by type of control group (living vs deceased) and sensitivity analyses excluding studies in which controls acted as their own informants.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    One researcher screened titles and abstracts, and full-text publications were screened with an additional reviewer, with disagreements resolved by consensus.

    “One researcher (AU) screened titles and abstracts of the retrieved articles for eligibility. Full-text publications were then screened with an additional reviewer (LF) and disagreements were resolved by consensus.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Study design and conductnone given

    All mental disorders were based on standardised diagnostic criteria.

    “All mental disorders were based on standardised diagnostic criteria.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Data extraction was done independently by two authors using a predetermined form.

    “Data extraction was done independently by two authors (LF and AU) using a predetermined form listing relevant study characteristics (ie, publication year, country and sample size), sample details (ie, age, sex and type of control group) and risk factor estimates.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Meta-analysis was performed only for risk factors that were examined in at least three independent samples.

    “To ensure accuracy and reliability of obtained estimates, meta-analysis was performed only for risk factors that were examined in at least three independent samples.14”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [14] Favril, 2020, Risk factors for self-harm in prison: a systematic review and meta-analysis, Lancet Psychiatry source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    No language restrictions were set.

    “No language restrictions were set.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Egger's test is appropriate for variables examined in at least 10 independent samples.

    “The presence of publication bias was examined by visual inspection of funnel plots16 and by applying Egger’s test17 for variables examined in at least 10 independent samples.15”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [16] source

      Not checked in the free run.

    • cites [17] source

      Not checked in the free run.

    • cites [15] Higgins JPT , Thomas J , Chandler J . Cochrane Handbook for systematic reviews of interventions. 2nd ed. Chichester: John Wiley & Sons, 2019. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Five electronic databases were systematically searched for psychological autopsy studies.

    “We systematically searched five electronic databases (PubMed, Global Health, Embase, Web of Science and PsycINFO) for psychological autopsy studies published from inception to 28 February 2022.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Bibliographical searches were complemented by manual searches of reference lists of included studies and previous systematic reviews.

    “Bibliographical searches were complemented by manual searches of reference lists of included studies and previous systematic reviews.5–9”
    Quote in the paperNot a substantial claim0 of 5 sources checked
    • cites [5] source

      Not checked in the free run.

    • cites [6] source

      Not checked in the free run.

    • cites [7] source

      Not checked in the free run.

    • cites [8] Cho, 2016, Geographical and temporal variations in the prevalence of mental disorders in suicide: systematic review and meta-analysis, J Affect Disord source

      Not checked in the free run.

    • cites [9] Conner, 2019, Metaanalysis of mood and substance use disorders in proximal risk for suicide deaths, Suicide Life Threat Behav source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The systematic review and meta-analysis was conducted following PRISMA guidelines.

    “This systematic review and meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines10 (online supplemental table S1).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [10] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

    Premisecitation

    The Newcastle-Ottawa Scale is a valid tool for assessing risk of bias in case-control studies.

    “Risk of bias was assessed using the Newcastle-Ottawa Scale for case–control studies.13”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] Wells GA , Shea B , O'Connell D . The Newcastle-Ottawa scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa: Ottawa Health Research Institute, 2014.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    An I2 value of 75% and over indicates considerable heterogeneity across studies.

    “Considerable heterogeneity across studies is indicated by an I² value of 75% and over.15”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [15] Higgins JPT , Thomas J , Chandler J . Cochrane Handbook for systematic reviews of interventions. 2nd ed. Chichester: John Wiley & Sons, 2019. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Meta-regression is appropriate for risk factors when there are 10 or more samples in the meta-analysis.

    “Meta-regression was considered for risk factors when there were 10 or more samples in the meta-analysis.15”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [15] Higgins JPT , Thomas J , Chandler J . Cochrane Handbook for systematic reviews of interventions. 2nd ed. Chichester: John Wiley & Sons, 2019. source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Risk of bias was assessed using the Newcastle-Ottawa Scale for case-control studies.

    “Risk of bias was assessed using the Newcastle-Ottawa Scale for case–control studies.13”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [13] Wells GA , Shea B , O'Connell D . The Newcastle-Ottawa scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa: Ottawa Health Research Institute, 2014.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    When multiple publications on the same or overlapping study population were available, information on risk factors was extracted from the investigation with the largest sample.

    “When multiple publications on the same or overlapping study population were available, information on risk factors was extracted from the investigation with the largest sample.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The study protocol was pre-registered with PROSPERO.

    “The study protocol was pre-registered with PROSPERO.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    Publication bias was examined by visual inspection of funnel plots and by applying Egger's test for variables examined in at least 10 independent samples.

    “The presence of publication bias was examined by visual inspection of funnel plots16 and by applying Egger’s test17 for variables examined in at least 10 independent samples.15”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [16] source

      Not checked in the free run.

    • cites [17] source

      Not checked in the free run.

    • cites [15] Higgins JPT , Thomas J , Chandler J . Cochrane Handbook for systematic reviews of interventions. 2nd ed. Chichester: John Wiley & Sons, 2019. source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Having a mental disorder is a strong risk factor for suicide.

    “For example, although it is well known that having a mental disorder is a strong risk factor for suicide,2 there were clear differences between diagnoses in their relative strength.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [2] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    A wide range of predisposing and precipitating factors are associated with suicide among adults in the general population, but with clear differences in their relative strength.

    “A wide range of predisposing and precipitating factors are associated with suicide among adults in the general population, but with clear differences in their relative strength.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    The psychological autopsy approach has limitations including issues relating to the choice (and selective non-response) of control subjects and the reliance on proxy reports, which are prone to recall bias and measurement error.

    “In addition, there are several limitations inherent to the psychological autopsy approach,23–25 including issues relating to the choice (and selective non-response) of control subjects and the reliance on proxy reports—the latter being prone to recall bias and measurement error.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [23] source

      Not checked in the free run.

    • cites [24] source

      Not checked in the free run.

    • cites [25] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    The strongest risk factors in the sociodemographic domain were social isolation, unemployment, and low socioeconomic status.

    “The strongest risk factors identified were social isolation (OR=4.0, 95% CI 2.1 to 7.7), unemployment (OR=3.8, 95% CI 2.7 to 5.2) and low socioeconomic status (OR=2.8, 95% CI 1.8 to 4.2).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Other studies have found a dose-response relationship between the number of events experienced and the risk of suicide.

    “Other studies have found a dose–response relationship between the number of events experienced and the risk of suicide.18”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [18] Fjeldsted, 2017, Suicide in relation to the experience of stressful life events: a population-based study, Arch Suicide Res source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Within the clinical domain, a history of self-harm and a previous suicide attempt were both strongly associated with suicide.

    “Within the clinical domain, a history of self-harm (OR=10.1, 95% CI 6.6 to 15.6) and a previous suicide attempt (OR=8.5, 95% CI 5.3 to 13.4; figure 1) were both strongly associated with suicide.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductreasoning

    The strength of the reported associations is likely to be overestimated because the review did not account for confounding, and the risk factors examined are not independent from one another.

    “First, the strength of the reported associations is likely to be overestimated as we did not account for confounding, and the risk factors examined are not independent from one another (eg, psychiatric comorbidity).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Findings should not be generalised to adolescents and older adults because the review only focused on working-age adults.

    “Fifth, our findings should not be generalised to adolescents and older adults as we only focused on working-age adults.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Risk factors are known to vary by age groups and might have a different impact according to when they occur in someone's life.

    “Risk factors are known to vary by age groups2 and might have a different impact according to when they occur in someone’s life (eg, unemployment and physical illness).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Register-based studies may minimise such biases but are more restricted in scope because many potential risk factors are not captured in routinely collected administrative records.

    “While register-based studies may minimise such biases, they are more restricted in scope since many potential risk factors are not captured in routinely collected administrative records (eg, recent life events, social isolation, religious beliefs and self-harm not resulting in service contact).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Many potential risk factors are not captured in routinely collected administrative records.

  • May combine two claims.

    Study design and conductown data

    The review retained all eligible studies irrespective of their methodological quality, which may have influenced the results.

    “Third, we retained all eligible studies in our analyses irrespective of their methodological quality, which may have influenced the results—though there was little evidence to suggest that study quality was associated with heterogeneity in meta-regression.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    Strengths of this review include the use of a quantitative synthesis and examination of a wide range of risk factors not limited to mental disorders.

    “Strengths of this review include the use of a quantitative synthesis and examination of a wide range of risk factors not limited to mental disorders.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Previous reviews of psychological autopsy studies have not captured the breadth of individual-level risk factors for suicide and mainly focused on the contribution of mental disorders to suicide risk.

    “First, previous reviews of psychological autopsy studies5–9 have not captured the breadth of individual-level risk factors for suicide, and instead mainly focused on the contribution of mental disorders to suicide risk.”
    Quote in the paperSubstantial claim0 of 5 sources checked
    • cites [5] source

      Not checked in the free run.

    • cites [6] source

      Not checked in the free run.

    • cites [7] source

      Not checked in the free run.

    • cites [8] Cho, 2016, Geographical and temporal variations in the prevalence of mental disorders in suicide: systematic review and meta-analysis, J Affect Disord source

      Not checked in the free run.

    • cites [9] Conner, 2019, Metaanalysis of mood and substance use disorders in proximal risk for suicide deaths, Suicide Life Threat Behav source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Any mental disorder and any personality disorder were strongly associated with suicide.

    “We found strong associations for any mental disorder (OR=13.1, 95% CI 9.9 to 17.4) and any personality disorder (OR=6.8, 95% CI 4.7 to 9.8).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    By diagnosis, depression had the strongest association with suicide, followed by borderline personality disorder and schizophrenia spectrum disorder.

    “By diagnosis, depression had the strongest association with suicide (OR=11.0, 95% CI 7.3 to 16.5), followed by borderline personality disorder (OR=9.0, 95% CI 5.6 to 14.4) and schizophrenia spectrum disorder (OR=7.8, 95% CI 4.5 to 13.5).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    Comparing suicide cases with controls who have died by other external causes such as accidents may generate smaller or no differences between groups because of overlapping risk factors and because prematurely deceased controls may be at increased risk of suicide themselves.

    “Comparing suicide cases with controls who have died by other external causes such as accidents may generate smaller or no differences between groups because of overlapping risk factors19 and because prematurely deceased controls may be at increased risk of suicide themselves.4”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [19] source

      Not checked in the free run.

    • cites [4] Pirkis, 2020, The case for case–control studies in the field of suicide prevention, Epidemiol Psychiatr Sci source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    For adverse life events, relationship conflict, legal problems and family-related conflict had the strongest associations with suicide.

    “For adverse life events, relationship conflict (OR=5.0, 95% CI 3.3 to 7.6), legal problems (OR=4.8, 95% CI 2.4 to 9.4) and family-related conflict (OR=4.5, 95% CI 2.0 to 10.3) had the strongest associations with suicide.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductreasoning

    Because different studies used contrasting approaches to adjustment, adjusted estimates would be difficult to compare, so the main analysis focused on crude effect sizes.

    “However, because different studies used contrasting approaches to adjustment, which would make adjusted estimates difficult to compare, our main analysis focused on crude effect sizes.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Different studies used contrasting approaches to adjustment, making adjusted estimates difficult to compare.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    In general population settings, many people at elevated risk of suicide do not access healthcare services and links to effective interventions may not be possible or timely.

    “This approach differs from that in general population settings, where many people at elevated risk of suicide do not access healthcare services and links to effective interventions may not be possible or timely.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Many people at elevated risk of suicide do not access healthcare services.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Findings underline the importance of both assessment and treatment of mental disorders in healthcare settings.

    “Our findings underline the importance of both assessment and treatment of mental disorders in healthcare settings.33”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [33] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    Effects might be smaller in prospective studies owing to long follow-up periods, whereas psychological autopsies focus on the time window close to death.

    “Among other reasons, effects might be smaller in prospective studies owing to long follow-up periods, whereas psychological autopsies focus on the time window close to death.9”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [9] Conner, 2019, Metaanalysis of mood and substance use disorders in proximal risk for suicide deaths, Suicide Life Threat Behav source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Effective pharmacological and psychosocial treatments are available for people with depression.

    “Another high-risk group with a treatment gap is people with depression, for whom effective pharmacological and psychosocial treatments are available38”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [38] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Consistent with a stress-diathesis model, suicide should be understood as the result of a cumulation of multiple risk factors, some predisposing and some precipitating.

    “Consistent with a stress-diathesis model,29 this underscores the need to understand suicide as the result of a cumulation of multiple risk factors,2 some of which are predisposing (such as sociodemographic background and family history) and some precipitating (such as incident diagnoses and life events).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [29] source

      Not checked in the free run.

    • cites [2] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Screening and assessment in itself will not improve outcomes unless it leads to effective intervention.

    “Screening and assessment in itself will not improve outcomes, unless it leads to effective intervention.34”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [34] Miller, 2017, Suicide prevention in an emergency department population, JAMA Psychiatry source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Psychosocial treatment following self-harm, particularly cognitive behavioural approaches, can reduce risk of repetition.

    “Interventions to consider include psychosocial treatment following self-harm, particularly cognitive behavioural approaches, which can reduce risk of repetition.36”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [36] Witt, 2021, Psychosocial interventions for self-harm in adults, Cochrane Database Syst Rev

      Not checked in the free run.

66 of 66

  • The paper relies on this without support.

    Many included studies exclusively reported significant associations from multivariable models.

    “many studies exclusively reported significant associations from the multivariable models, which likely results in biased comparisons between estimates.”

  • The paper relies on this without support.

    Different studies used contrasting approaches to adjustment, making adjusted estimates difficult to compare.

    “because different studies used contrasting approaches to adjustment, which would make adjusted estimates difficult to compare”

  • The paper relies on this without support.

    Most primary studies did not stratify analyses by sex.

    “most primary studies did not stratify analyses by sex”

  • The paper relies on this without support.

    Many potential risk factors are not captured in routinely collected administrative records.

    “many potential risk factors are not captured in routinely collected administrative records”

  • The paper relies on this without support.

    Many people at elevated risk of suicide do not access healthcare services.

    “many people at elevated risk of suicide do not access healthcare services”
5 of 5

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Social and cultural factors can contribute to suicide risk.

    “In line with early theories, recent studies found that social and cultural factors can contribute to suicide risk (perhaps to a lesser extent compared to the presence of any mental disorder, see Favril et al., 2022), particularly in the way that individuals perceive themselves within broader contexts.”

    From Toward objective characterizations of suicide risk: A narrative review of laboratory-based cognitive and behavioral tasks (2023) · cites this paper as Favril et al., 2022

  • Decades of epidemiological and clinical research have established a reasonably consistent inventory of variables associated with elevated suicide risk at the group level.

    “Decades of epidemiological and clinical research have indeed established a reasonably consistent inventory of variables associated with elevated suicide risk at the group level.”

    From Suicide risk assessment: clinical implications of the unpredictability of suicidal behavior (2026) · cites this paper as [5]

  • The possibility of suicide usually comes into focus when individuals are exposed to significant and/or multiple stressors and/or suffer from a mental disorder.

    “The possibility of suicide usually comes into focus when individuals are exposed to significant and/or multiple stressors (separation, death, unemployment, financial loss, physical illness, trauma (27);) and/or suffer from a mental disorder (28);”

    From Ambivalence model of suicidality [ABS-model]: an orientation model for the treatment of suicidal individuals (2024) · cites this paper as [28]

  • NSSI might often be underestimated and underreported.

    “NSSI, which might often be underestimated and underreported [68, 69]”

    From Association between adverse childhood experiences and suicidal behavior in affective disorders: A systematic review and meta-analysis (2025) · cites this paper as [68]

    The Strengths and limitations section (page unavailable) says proxy reports can have recall and measurement error and that administrative records miss self-harm not resulting in service contact. I searched for NSSI, non-suicidal self-injury, underreporting, and underestimation, but found no passage saying NSSI is often underestimated or underreported. (LLM, conf=0.72)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Favril, L., Yu, R., Uyar, A., Sharpe, M., & Fazel, S.
Affiliations
  • Ghent University (BE)
  • University of Oxford (GB)
Corresponding author
Fazel, Seena
ORCID
Venue
Evidence Based Mental Health
Publisher
BMJ
Published
2022-09-26
Volume
25
Issue
4
Pages
148-155
ISSN
  • 1362-0347 (ISSN-L)
  • 1468-960X
DOI
10.1136/ebmental-2022-300549
PMC
PMC9685708
OpenAlex
W4297265271
Cited by
336
References
42
Citations by year
2026: 87 · 2025: 107 · 2024: 93 · 2023: 46 · 2022: 1 · 2021: 1
FWCI
10.36
Type
journal article
Language
en
Open access
Licence
Locations
  • Evidence-Based Mental Health (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Ghent University Academic Bibliography (Ghent University) (submittedVersion, cc-by)
Funders
Wellcome Trust (202836/Z/16/Z)
Topics
  • Social Sciences › Psychology › Clinical Psychology › Suicide and Self-Harm Studies
  • Social Sciences › Sociology and Political Science › Homicide, Infanticide, and Child Abuse
  • Social Sciences › Psychology › Clinical Psychology › Grief, Bereavement, and Mental Health
Keywords
suicide risk factors, suicide mortality, adults, psychological autopsy, mental illness, self-harm, sociodemographic factors, family history (medicine), life events, suicide prevention
MeSH terms
Suicide Prevention, Adult, Autopsy, Mental Disorders, Humans, Risk Factors, Self-Injurious Behavior

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review