Work ‐ related suicide: Evolving understandings of etiology & intervention

Previously published analyses of suicide case investigations suggest that work or working conditions contribute to 10% – 13% of suicide deaths. Yet, the way in which work may increase suicide risk is an underdeveloped area of epidemiologic research. In this


Funding information National Health and Medical Research Council
Commentary, we propose a definition of work-related suicide from an occupational health and safety perspective, and review the case investigation-based and epidemiologic evidence on work-related causes of suicide.We identified six broad categories of potential work-related causes of suicide, which are: (1) workplace chemical, physical, and psychosocial exposures; (2) exposure to trauma on the job; (3) access to means of suicide through work; (4) exposure to high-stigma work environments; (5) exposure to normative environments promoting extreme orientation to work; and (6) adverse experiences arising from work-related injury or illness.We summarise current evidence in a schema of potential work-related causes that can also be applied in workplace risk assessment and suicide case investigations.There are numerous implications of these findings for policy and practice.Various principle-and evidence-based workplace intervention strategies for suicide prevention exist, some of which have been shown to improve suicide-prevention literacy, reduce stigma, enhance helping behaviours, and in some instances maybe even reduce suicide rates.Prevailing practice in workplace suicide prevention, however, overly emphasises individual-and illness-directed interventions, with little attention directed to addressing the working conditions that may increase suicide risk.We conclude that a stronger emphasis on improving working conditions will be required for workplace suicide prevention to reach its full preventive potential.

| INTRODUCTION & APPROACH
2][3] In absolute terms, the majority of suicide deaths occur among people of working age. 2,4Yet the role of work and how it might contribute to suicide risk is a relatively underdeveloped area of research.Given that roughly two-thirds of the working age population is engaged in work, an occupational health & safety (OH&S) perspective aims to elucidate the role that working conditions play in increasing or decreasing the risk of suicide, so as to inform workplace prevention and control strategies.
Defining work-related suicide, on the face of it, is quite simple: death by suicide that is wholly or partly caused by work or working conditions.We define our threshold for determining work-relatedness as adequate evidence to justify policy or practice action, an approach informed by the precautionary principle. 5,6The precautionary principle in public health aims to balance uncertainty of evidence with legal and ethical duties to act to prevent harm.We recognise that inferred causes based largely on observational evidence are always provisional and may change over time, and that causal inference is context-dependent.This uncertainty must be weighed against considerations including the gravity of the potentially preventable outcome and what would be required to reduce the risk in question.[7] Defining work in the context of work-related suicide, however, is in some ways challenging.][10][11] We acknowledge this as a simplification because working conditions and unemployment are not necessarily mutually exclusive contributors to suicide risk.Indeed, job insecurity, underemployment, and unemployment can be conceptualised as being on the same continuum, 12 and evidence detailed below suggests that job insecurity, as well as unemployment, increases the risk of suicide.
When evaluating work-related suicide, the main group to consider is that of persons working for profit or pay at the time of death (e.g., employees, temporary workers, contractors, and selfemployed workers), and their work and working conditions as potential contributors to suicide risk.Consideration of work-related suicides should not be restricted to those happening at the workplace or during the working hours of the deceased, and all means of taking one's own life should be considered.Persons who die by suicide but were not currently working due to work-related injury or illness should also be included, whether they are out of work temporarily (e.g., on Workers' Compensation) or on permanent disability, as detailed further below.Unpaid volunteers are also entitled to safe and healthy working conditions, and are thus another included group.
Finally, because the adverse effects of some working conditions can be delayed, it is also possible that former work or working conditions could contribute to suicide deaths whether among persons working in subsequent jobs, unemployed, or no longer in the labour force.
These groups are outlined in Figure 1, and discussed in further detail in the pages that follow.
We seek to integrate two approaches to identifying work-related suicides to gauge the magnitude of the problem: where workrelatedness is established 1) on an individual case-basis, and 2) using a population-based epidemiologic approach.We argue that these two approaches are complementary, and both are required to understand the scale of the problem and to formulate optimal policy and practice responses (sections II and III).Following an outline of our approach, we present an integrated summary of work-related risk factors for suicide that also highlights recommended areas for consideration in workplace risk assessment and suicide investigations (section IV).
Finally, we outline the strengths and limitations of this review and its implications for policy & practice in section V.The overarching aim is to provide a high-level summary of current evidence on aetiology and intervention so as to inform evolving policy & practice.
We note some acknowledgements and provisos for this Commentary: 1) a previously-published Discussion Paper authored by ADL and TK 13 provided the basis of this Commentary, which was subsequently expanded and refined in collaboration with 17 additional international researchers in this area; 2) we have conducted a narrative synthesis and interpretation of the literature drawing on the collective expertise of the authors; 3) we acknowledge a number of recent reviews on work and suicide which have informed this Commentary [14][15][16][17] ; and 4) this Commentary was informed by, and will be complemented by, a planned systematic review & meta-analysis of working conditions and suicide led by one of the authors of this Commentary.In France, a set of criteria for potential work-related suicides was proposed and tested by the National Public Health agency for population surveillance purposes. 23Each suicide or death of undetermined cause was considered "potentially" work-related when at least one of the following circumstances was present: • The suicide occurred in the workplace; • A suicide note or the testimony of relatives or other close contacts implicated working conditions; • The deceased was wearing work clothes; • Work-related difficulties were otherwise identified in investigations.
When tested in a feasibility study of 1135 suicide deaths in 2018, 10% were deemed potentially work-related. 23 Australia, a 2012 study reviewed 642 suicides identified as "broadly work-related" based on investigations conducted by police or coroners' reports in one state over a 7-year period. 24ese suicides accounted for 17% of all suicides during that time (642/3775)."Work-relatedness," as defined in this 2012 study, was further classified as involving reports of "work stressors" (55%); jumped, laid, or moved in front of a moving train or heavy vehicle (32%); involved a work location (7%); or involved work as a means (6%).Importantly however, this classification captured many people who died by suicide who were unemployed or not in the labour force (NILF) (e.g., retired, full-time students).Conservatively excluding those who were unemployed or NILF at the time of death (which may have been influenced by previous work experiences) leaves 396/3775, or 10.5% where work was a potentially contributing factor.
A 2021 analysis of 89,389 suicides across 35 states in the United States between 2013 and 2017 used death certificates and investigation documentation, and found that 13.5% of suicides were work-related. 25In a further analysis, these investigators estimated work was a "major factor" in one quarter of "work-related" suicides, or rough 3% of all suicides. 26recent New Zealand study of 1678 suicide cases over a 4-year period 27 determined work-related suicide as meeting either of two criteria: • Work-related stressors played a significant role in the person's suicidality, and/or • The means of suicide were distinctly related to the deceased person's work.To summarise, there is no single definition of work-relatedness in the above-reviewed psychological autopsy and other studies.Various indicators or categories of work-relatedness have been applied, some of which have face validity indicating work as a contributing cause (e.g., work-related difficulties) and some of which may not (e.g., wearing work clothes at the time of death).Some determined potential work-relatedness, and others were more definitive Further, following our definition, we exclude (as work-related) the use of the work or workplace of others as a means of taking one's own life (e.g., suicide on rail infrastructure by members of the public) if the deceased was neither working for profit or pay at the time of death nor otherwise eligible for inclusion as "working" per Figure 1.
Case-based approaches have advanced our understanding of work and suicide, as well as justification for developing prevention and control responses both in the workplace (e.g., the 2014 Japanese National Prevention Strategy for Overwork-related Disorders 22 ) and in general community settings (e.g., restricting pesticide access 28 ).Casebased approaches, however, are limited by the lack of standardised data-collection methods on work-relatedness, the limited knowledge of the work circumstances of the deceased in post-mortem interviews, potential biases of persons close to the deceased, and potential biases of investigators and coroners (e.g., prioritising medical diagnoses while understating psychosocial factors when explaining suicides 29 ).For example, an Australian study found that between 1985 and 2007 there were 21 workers' compensation claims for suicide in the state of Victoria, yet review of the coronial records identified work factors or a previous work injury for only 9 of the 21 workers compensation cases. 30Accordingly, it seems likely that work-relatedness is underidentified in studies relying on investigations of individual suicide cases, at least historically.In the section that follows, we argue that the individual case perspective needs to be complemented by a population or epidemiologic perspective.

| EPIDEMIOLOGIC APPROACHES TO DETERMINING WORK-RELATEDNESS
In this section, we focus on epidemiologic studies that assess suicide outcomes in relation to specific occupational exposures that may be risk (or protective) factors for suicide.2][33][34] While studies of occupations and suicidal behaviours may suggest specific workplace exposures or risk factors, they are considered here only where they complement studies of associations assessing specific working conditions or exposures.For example, some studies have found high rates of suicide in construction workers, but this is most likely explained by a combination of compositional factors (e.g., predominantly male workforce) and contextual working conditions (e.g., long working hours, job insecurity, masculine norms discouraging help-seeking). 35ch studies of occupations or sectors are valuable for hypothesis generation and to identify target groups for general suicide prevention intervention, 36 but less informative for assessing specific risk factors or causes.
Potential work-related contributors to suicide identified by epidemiologic studies can be grouped into the following categories: chemical exposures; physical exposures; psychosocial working conditions; long working hours; precarious employment; exposure to suicide, other deaths or trauma on the job; and adverse experiences arising from work-related injury or illness.

| Chemical exposures
8][39][40][41] As far back as the mid-1800s, case reports in rubber and rayon manufacturing documented that occupational exposure to carbon disulphide was associated with increased rates of depression, acute mania, suicide attempts and "commitments to mental institutions" among the exposed 42 (Hamilton 1943, cited in Mancuso and Locke 43 ).Decades later, a United States retrospective cohort study of rayon workers exposed from 1938 to 1948, with follow-up until 1968, found up to fourfold elevated suicide mortality rates compared to the United States working age population. 43me chemical exposures have plausible links to suicide through their association with increased risks of depression and anxiety symptoms, such as certain pesticide, organic solvent, and toxic metal exposures. 31,44The association is further supported by elevated suicide rates in occupations with these exposures, such as farmers and other agricultural workers (independent of using pesticides as a means of suicide) 45 and automotive workers with potential solvent exposures. 44However, a systematic review of occupational pesticide exposure, depression and suicide found mixed evidence of associations, with the studies limited by poor exposure data, limited control for confounding and a dearth of prospective studies. 46One study of inorganic lead exposed workers in Korea found an association between high blood lead levels and suicide mortality among males, which the authors speculated could be mediated by lead-associated depression. 40Finally, a study of United States nuclear weapons production workers found an association between uranium exposure and suicide mortality. 41 summary, while there is some evidence of suicide in association with occupational chemical exposures, the evidence is patchy and for more extensively studied exposures, such as pesticides, inconclusive.Nevertheless, such exposures warrant further research.While many of these exposures are not widely prevalent in high-income countries, and exposure levels are generally lower now than historical levels, they may be of greater concern in lower and middle income countries (LMIC) to the extent that exposures persist or have increased in prevalence or intensity due to the export of hazardous industries from high-income countries to reduce production costs or escape regulation (e.g., for carbon disulphide). 47

| Physical exposures
Occupational exposures to adverse physical conditions at work, including noise and electromagnetic fields (EMF) have been investigated as potential contributors to suicide risk.It has been hypothesised that noise exposure at work may interfere with tasks or other work activities, in turn eliciting negative changes in emotions, psychological effects, and behavioural changes. 48Some research suggests that psychosocial working conditions (e.g., job demands) interact with acoustics and noise exposure in their association with job-related stress. 48A cross-sectional Korean national population survey (N = 10,020), for example, found that occupational noise annoyance, as a proxy for noise exposure, was significantly related to depressive symptoms and suicidal ideation after controlling for individual and socio-demographic characteristics. 49There has been little study in this area to date, but occupational noise exposure is common in the working population and thus warrants further research.
There is mixed evidence of an association between occupational exposure to EMF and suicide.A job-exposure matrix (JEM)-based study of 21,744 electrical utility workers in Quebec found no statistically significant association with suicide, 50 nor did an occupational title-based study of Swedish electricians. 51Conversely, a nested case-control study of suicide mortality in a parent cohort of 138,905 male electric utility workers found an association between increasing years of employment as an electrician or line worker and risk of suicide, including a dose-response gradient. 52This study, however, did not control for important potential confounders such as drug use and mental illness.Although the evidence is inconclusive at present, the role of noise and EMF exposures warrant further research.
Concussion (traumatic brain injury), repeated concussion, and chronic traumatic encephalopathy have been associated with a twofold elevation in risk of suicide. 53,54While this could be a risk for professional athletes in contact sports (e.g., football, gridiron, hockey), investigations to date have shown no excess risk of suicide among professional athletes. 55Further research is warranted.

| Psychosocial working conditions
Ove the last decade, there has been substantial growth in the number of epidemiologic studies on psychosocial working conditions as risk factors for suicide as well as continuing improvements in methodological quality. 56,57e first systematic review of the area 58 presented a synthesis of 22 studies assessing the relationship between psychosocial working conditions and suicide ideation, attempts, and deaths.Results suggested that exposure to psychosocial working conditions was associated with elevated risks of suicidal ideation and death. 58rious psychosocial working conditions were associated with suicidal ideation, but only low job control was associated with increased risk of suicide mortality.The review, however, also acknowledged that studies were available on only a few commonly measured psychosocial working conditions, and most studies were cross-sectional in design.Most studies of ideation were limited by potential dependent misclassification (common method bias) due to the use of self-reported exposure and outcomes measures.All studies were also potentially affected by confounding through health selection: workers who are more vulnerable to suicidal thoughts or behaviours may also be more likely to be employed in jobs with worse working conditions, more likely to report poor psychosocial working conditions, or both.Overall, results suggested the potential for publication bias, and high heterogeneity.Further, there was little control for socioeconomic status or nonwork related stressors, some summary estimates were small and marginally statistically significant, and only six studies included suicide deaths.
We have identified several additional studies on psychosocial working conditions and suicide that have been published since the 2018 review. 58These studies found elevated risks of suicide deaths in association with low social support, 59 low job control, [59][60][61] job strain, [59][60][61][62] isotrain (job strain combined with low social support), 59 passive jobs [59][60][61][62] job insecurity, 63 workplace sexual harassment, 64 workplace bullying, 65 and workplace violence. 66Some studies found evidence of a protective association for active jobs (high control combined with high demands), [60][61][62] however, study results were difficult to compare due to varying reference categories. 59Some studies also found protective associations for high job demands, and no significant elevation of risk for job strain after multiple confounder adjustments. 62The most recent study of bullying 66 presented a small meta-analysis including the earlier positive study, 65 and identified an elevated risk that was attenuated with confidence intervals including unity after adjustment for baseline mental health problems.
Collectively, these studies address most of the shortcomings noted in the 2018 review 58 and strengthen causal inference for the associations between psychosocial working conditions and suicide.
Most were working population-based, all used register-based outcomes and prospective designs.Some used individual and some JEM-based exposure measures, with one study assessing 5-year JEM-based exposure trajectories. 62All studies either adjusted for, or excluded, those with a history of mental illness or previous suicide attempt, and included adjustments or sensitivity analyses addressing a range of other potential confounders.Some of these studies estimated population attributable fractions (PAF), which assume the observed associations are causal and free of residual confounding.Acknowledging these assumptions, in France an estimated 5.3% of suicide deaths among men and 9.1% among women were attributable to job strain. 59In Sweden, an estimated 6% of suicide deaths were attributable to sexual harassment at work. 64In the Swedish job insecurity study (13% prevalence), the PAF for suicide deaths attributable to job insecurity was calculated at 6%. 63 Many psychosocial working conditions cooccur, 67,68 thus, under the assumption that the associations underpinning these estimates are causal, the PAFs could be less than additive but will likely be higher than the maximum PAF for one given exposure. 69We are not aware of any suicide mortality estimates that take these co-exposures into account, nor the extent to which exposures to chronic psychosocial working conditions are represented in the case record-based determinations of work-relatedness outlined in the previous section.Because case-based investigations would likely emphasise acute stressors or recent events, it is likely that a full accounting of work-related suicide that incorporates a range of common psychosocial working conditions could lead to estimates that exceed the case record-based estimates of 10%-13% of suicides in the working population.Whether in the range of 10%-13% or higher, this would represent a substantial and preventable work-related suicide mortality burden.

| Long working hours
Some studies suggest that long working hours increase suicide risk.A Korean study (n = 14,484) showed 3-4 times elevated suicide mortality risk among workers working 45-52 h and >52 h per week compared to a reference of 35-44 h/week, but this study was limited by the small number of suicide deaths (n = 27). 70In a small study of United States workers (n = 578), working >40 h/week predicted a four-fold increase in odds of moderate to severe suicidal ideation compared to those working < = 40 h/week. 71Some older studies (before 2018) also show adjusted associations between long working hours and suicidal ideation, 72 however further research is needed in this area.

| Precarious employment
Precarious employment is defined in various ways, 73 which makes comparisons between studies difficult.Precarious employment also overlaps to some extent in concept with job insecurity (discussed above).In a very large register-based working population study in Sweden (n = 2,743,764), employment trajectories were characterised with regard to insecurity, income inadequacy, and lack of rights and protections; workers experiencing constant precarious employment showed a 1.2-fold elevated risk of intentional self-harm. 74A South Korean cohort study of 3793 workers who were permanently employed for at least 3 years reported 2 to 4-fold elevated odds of suicidal ideation when transitioning to precarious employment. 75A similar, though smaller magnitude, finding was made in a larger nationally representative cohort of South Korean workers (n = 25,862) wherein workers transitioning from permanent to precarious jobs reported a 1.7 times higher odds of reporting suicidal ideation (OR 1.7, 95% CI 1.3-2.3)than those remaining in permanent employment. 76Further research is urgently needed in this area in light of the growth of precarious employment in many parts of the world.scoping review of 25 studies found that between 32% and 95% of first responders and mental health professionals had been exposed to suicide on the job. 77Exposure to trauma is strongly associated with posttraumatic stress disorder (PTSD) which, in turn, is a risk factor for suicidal ideation and behaviours. 15United States survey of 1048 first responders, crisis workers, and mental health professionals found that increasing exposure to suicide on the job was associated with increasing levels of depression, anxiety, and PTSD. 78In an Australian national crosssectional survey of 14,868 police and emergency responders the career period prevalence of exposure to traumatic events at work was 51% 79 ; further, exposure to stressful events was associated with a 4.5-fold increase in odds of post -traumatic stress (PTS) symptoms (a proxy measure meeting core symptom diagnostic criteria for PTSD). 79These results yield a PAF of 65% of PTS symptoms being attributable to workplace trauma exposure. 79A further analysis of the same data showed PTS symptoms were associated both with suicidal ideation and planning a suicide attempt. 80spite the high prevalence of trauma exposure, PTSD and other mental health problems amongst emergency service workers, a recent examination of more than 13,000 suicide deaths in Australia found there was no evidence of an increased risk of suicide amongst emergency service workers compared to other occupations once the age and gender of the workforce was taken into account, though there was some evidence of excess risk among ambulance personnel in particular. 81Amongst military veterans, the results on risk of suicide are more mixed, with a consistent finding of increased risk of death by suicide for those in some, but not all countries. 82posure to trauma may also directly affect suicidality, independent of PTSD. 83For example, Joiner's interpersonal theory of suicide proposes three key mechanisms to explain suicidality: thwarted belongingness, perceived burdensomeness, and acquired capability. 84quired capability refers to the capability to make a potentially lethal suicide attempt; this capability can be acquired gradually through repeated exposure to pain, fear, trauma, and deathincluding on the job, thus potentially desensitising those exposed and reducing barriers to using potentially lethal force on the self.

| Adverse experiences arising from workrelated injury or illness
There is growing evidence that workplace injury and associated disability is associated with elevated rates of suicide. 85A linked registry-based study of over 700,000 workers in Korea identified that both male and female injured workers had higher suicide mortality relative to the economically active population. 86In the United States, a study drawing on linked compensation data from one state found that work injuries necessitating more than 7 days off work, or with permanent disability benefits, were associated with almost two-fold elevated risk of suicide mortality compared to those with injuries only receiving compensation for medical expenses. 87e pathways underpinning these associations are complex. 88rkplace injuries are associated with a range of adverse outcomes for injured workers, with the consequences of workplace injury not limited to the injury itself, but extending across multiple domains.Compared to nonoccupational injuries, occupational injuries are more strongly associated with depressive symptoms, 89 with depression being a known risk factor for suicide.
In some cases, workplace injury can cause disability (short term or long term), which may reduce income substantially, as well as acute and chronic pain, both well-known risks for suicide. 90Pain medication including opioids are commonly prescribed and drugrelated mortality and suicide is known to be higher among "lost time" injured workers compared to "medical-only" injured workers. 87In a related vein, a United States study grouping segments of the labour force by occupational injury rates found that rising injury rates were associated with rising suicide rates, opioidrelated deaths, and deaths from alcoholic liver disease and cirrhosis (so-called 'deaths of despair'). 91rker's compensation processes may also contribute to distress and suicidality among injured workers.In Australia, elevated rates of hospital admission for self-harm have been observed among workers' compensation claimants. 92Other research from Australia has highlighted the stress of workplace injury compensation systems for injured workers, 93 and Canadian work supports this, identifying that the compensation process contributed to the stress to a greater extent than the injury itself. 94A more recent study of Australian police and emergency services workers found that 8% of employees making a claim reported a positive experience, contrasted with 70% who reported a poor experience; further two-thirds of those who made a claim reported that the process was unsupportive and stressful, and over half reported that it had an overall negative impact on their recovery. 95 summary, there is some, albeit limited, evidence that injured workers are at an increased risk for suicide.Further, evidence suggests multiple pathways through which work-related injury or illness could increase suicide risk, clearly warranting further research but also potentially providing numerous points of intervention to reduce suicide risks.

| PROPOSED SCHEMA OF POTENTIAL WORK-RELATED CAUSES OF SUICIDE
Based on the evidence synthesised above, we have developed a schematic summary of potential work-related causes of suicide, integrating the case-based and epidemiologic evidence summarised above (Figure 1).The schema also can serve as a framework for workplace risk assessment and management, as well as for investigating the potential work-relatedness of a suicide case.

| Box A-D: Establishing the potential for work to contribute to a suicide
There are numerous forms of "work" that could contribute to suicide risk.These include those who work for profit or pay or as a volunteer at the time of death (Box B).In Australia, for example, there are large volunteer workforces with potential for exposures to trauma and many of the other hazards named above in firefighting (144,000), emergency services (43,000), and surf lifesaving (190,000) alone, equating to ~3% of the working population.Among persons who were not working at the time of death due to a work-related injury or illness (Box D), it remains possible that their death was related to paid or volunteer work as detailed above.Finally, working conditions in past jobs may contribute to suicide risk among those working currently, or who were not working at the time of death, or who were out of the workforce on temporary disability or unemployment (Box C).With regard to investigation of suicide deaths, these various forms of work could effectively be discerned by documenting an employment history over the years preceding death from next-of-kin or other key informants.
Boxes E-G, described in further detail below, outline potential work-related causes.Neither the time nor place of a suicide (i.e., during working hours, or whether at work or not) are criteria for assessing work-relatedness, hence precluding their representation in this schema.

| Box E: Access to means
Access to means of suicide through one's work or workplace is a wellestablished risk factor for suicide. 14,96,97Examples include police, military, and other protective services personnel with access to firearms 98 agricultural workers with access to pesticides, veterinarians with access to euthanising drugs, 99 and healthcare professionals with access to medicines. 100,101Also supporting this association, an Australian national study showed that workers with access to means are more likely to use these means to end their lives than those without access to means. 97An earlier national study from New Zealand made similar findings. 102 should be noted that the use of work-related means might or  As summarised above, current evidence suggests that unintentional occupational exposures to chemical or physical hazards are not major contributors to suicide risk, but this warrants monitoring in the future.The evidence on psychosocial working conditions is the most rapidly expanding area of work and suicide research.In our view, there is adequate evidence to presume that low job control, high job strain, job insecurity/precarious employment, sexual harassment at work, bullying, occupational violence, and long working hours could be work-related causes of suicide.[105] Job-exposure matrices (JEMs) could be used in risk assessment and management, as well as to supplement investigations of individual suicide cases to establish typical exposure levels to hazardous chemicals, 106 physical stressors, 107 psychosocial working conditions, [108][109][110] and other occupational hazards 111 based on job titles and industrial sectors of the deceased.Historically, JEMs have been used mainly in research, but could be a valuable tool in these more applied contexts.Importantly however, JEMs can only be meaningfully used for work exposures that vary to a considerable extent by job group, thus, are only useful for some working conditions. 112

| Box F2: Exposure to suicide or other deaths and trauma on the job
As noted above, these exposures tend to occur mainly in particular occupations (e.g., emergency responders, mental health care providers, protective services workers).We are unaware of any JEMs that specifically cover exposure to deaths and trauma on the job, but there are actively updated inventories of occupational exposure data, including JEMs, that could be consulted in the future (e.g., https:// occupationalexposuretools.net/).

| Box F3: Exposure to normative work environments with high stigma or discouragement of helping behaviours
Work environments vary substantially in prevailing normative influences.Further, the norms within a workplace can vary in the extent to which they inhibit or foster help-seeking behaviour.Here we distinguish these normative influences that operate within workplaces from both macro level normative influences (Box F4) and psychosocial working conditions (Box F1), noting that these may overlap.Mental health stigma within a workplace has been found to be related to suicidal ideation, 95 and it is speculated that perceived mental health stigma may hinder help-seeking and lead to damaging or maladaptive coping behaviours. 95Masculine norms emphasising self-reliance and repudiating help-seeking have been associated with suicidal ideation, 113,114 and it is posited that the enforcement or internalisation of these norms within male-dominated workplaces such as in the construction industry may contribute to high rates of suicide observed in such settings. 115Though stigma is a common barrier to help-seeking, barriers can manifest in other ways as well.
For example, questions about mental illness in medical registration processes have been shown to be a barrier to help-seeking among physicians. 116ews may differ on whether environments with high stigma and discouragement of helping behaviours are working conditions subject to OH&S prevention and control intervention, although discriminatory behaviours are legislated against under disability and discrimination laws in many countries.Nevertheless, the evidence suggests that they are important modifiable work-related risk factors for suicide, and stigma and helping behaviours are a major focus of general workplace suicide prevention strategy which aims to prevent suicide regardless of cause (revisited in Discussion section below).

| Box F4: Exposure to broad cultural norms (macro level) that promote extreme orientation to work
Societies vary in normative orientation to work.Here, we consider what is known about the way broad societal cultural norms related to work are associated with suicide.We note some overlap with Box F1-many of the psychosocial working conditions considered in section F1 are relevant here.The focus in this section is macro level cultural norms that drive exposure to psychosocial working conditions.Japan, with one of the highest rates of suicide in the world, 117 is used as an example, noting that such macro level conditions may operate in many contexts. 21The Japanese term karo-jisatsu translates to suicide by overwork, and is applied to those who have suicided after a sustained period of overwork, characterised by extreme working hours and a heavy workload with little time off work. 19,20Given that there is overlap with long working hours and psychosocial working conditions (Box F1), karo-jisatsu may represent an interaction between psychosocial working conditions and powerful social or cultural norms.Typical of karo-jisatsu is a sense of guilt, shame and self-blame for an inability to accede to work expectations. 118Further, in Japan, a strong emphasis is placed on collectivism, meaning that within the workplace, employees are expected to prioritise and optimise the goals and successes of the organisation over their own wellbeing. 19e way in which these Japanese social norms, steeped in values of respect, perfectionism, and collectivism, 117 are embedded and enacted within work settings is thought to underpin karo-jisatsu. 20tersecting with and compounding work related suicide risk in Japan is stigma over mental health conditions, 119 which means that those in distress may also be reluctant to seek help (see also Box F3).

| Box G: Adverse experiences arising from work-related injury or illness
We recommend that suicide investigations, in addition to documenting employment status and work history, should collect data on whether the deceased was affected by a work-related injury or illness, whether they were working at the time, on temporary disability, unemployed or no longer in the labour force. 30Such cases could be deemed to be work-related if the consequences of that injury or illness were a contributing cause of the suicide, e.g. through loss of income, status, disability, or pain.

| DISCUSSION
In this Commentary, we have proposed a definition of work-related suicide from an OH&S perspective, assessed the weight of evidence for potential work-related causes of suicide, and reviewed estimates of the proportion of suicides that may be work-related.On the basis of this synthesis, we have developed a schema of potential workrelated causes to summarise current evidence and provide guidance for risk assessment and case investigations of potential work-related suicides.Our arguments have a number of strengths and limitations as outlined below.

| Strengths and limitations
The distinctions between being in work, being unemployed, retired, or otherwise not in the labour force (NILF) are become increasingly blurred in high-income countries, particularly with the rise of precarious and highly insecure employment, with some workers cycling frequently between paid work, unemployment, and NILF.We have attempted to be as comprehensive as possible in considering eligible forms of work.One of the greatest challenges is assessing potential work-relatedness due to previous jobs, regardless of current employment status.Similar challenges occur in the study of workrelated contributions to chronic multi-factorial diseases such as cancers and heart disease; methods to address lagged impacts of exposures, changing occupations, and time-varying exposures, for example, have been addressed in many of the recent studies of psychosocial working conditions and suicide reviewed above.For example, a French study provided evidence of lasting effects of psychosocial working conditions on suicide risk after leaving employment (Tables S1,2 in Niedhammer et al. 59 ).
Suicide deaths usually involve multiple causal factors, 2,3,14 of which work-related factors may be one or more among multiple.
Further, nonwork risk factors are usually stronger in magnitude of association with suicide than work-related factors (e.g., past or current mental illness, history of previous suicide attempt).Mental illness, however, can be work-related, such as psychosocial working conditions as contributors to incident depression. 120Taking a conservative approach by controlling for past or current mental illness (thus excluding potentially work-related depression as a mediator), as well as controlling for various nonwork-related risk factors and confounding by health selection, associations between psychosocial working conditions and suicide persisted in the most recent large-scale register-based prospective cohort studies.In a related vein, in our definition we have not considered whether work was a major, minor, or other contributor to a given suicide.While this is a particular limitation of the individual case-based estimates of the proportion of suicides that are work-related (for all except one study 26 ), the epidemiologically-based population attributable fraction estimates are less affected by this limitation.This argues for taking an integrated approach to workplace mental health and suicide prevention, wherein programs combine strategies to address both work-and nonwork-related risk and protective factors, as discussed further below. 120,121This would include resourcing General Practitioners, psychologists, and other front-line support and mental healthcare providers with the time, knowledge, and skills to conduct occupational histories and consider the role of work in aetiology, treatment, and rehabilitation. 120me previous reviews have highlighted that heterogeneity in results of psychosocial working conditions and suicide studies may be a concern. 16,58Heterogeneity can arise through bias or confounding, but it can also be genuine (without bias), arising through differences in sample composition, design, methods, context, or other factors.
Previous research has demonstrated effect modification of psychosocial working condition-mental health relationships by gender, 122 socioeconomic position, 123 and national labour and social policies. 124Hence, heterogeneity sometimes reveals variation that could inform policy and practice responses.The variety of designs, methods, measures, and analytical strategies brought to bear in assessing the relationship between psychosocial working conditions and suicide is also a positive: overall the evidence has yielded qualitatively consistent results, providing further support, by triangulation, for causal inference. 125 this Commentary, we have focussed on suicide risk factors, taking an OH&S perspective.We also acknowledge that work can be beneficial to health and wellbeing, and certain aspects of work may protect against suicide; these include providing income and promoting sense of purpose and belonging, time structure, meaningfulness, and more. 126These positive or protective aspects of work also warrant attention in workplace mental health and suicide prevention interventions, as further detailed elsewhere. 120,126 have not extensively delved into suicide theory.Working conditions that increase the risk of depression and other psychiatric disorders have clear plausible links to suicide, as these disorders are strong risk factors for suicide. 127But not all who die by suicide have psychiatric disorders.For examples, in certain LMIC many studies show that fewer than half of people who die by suicide have prior mental illness, 128,129 and a recent United States study suggested that an estimated 19.6% of adults had no antecedent psychiatric disorder among first suicide attempts. 130We have touched on Joiner's interpersonal theory above, noting the potential that exposure to trauma on the job could increase suicide risk through acquired capability.Working in isolation, low participation in decision-making (a dimension of job control), poor social support at work, and bullying and incivility at work, for examples, could also contribute to work-related suicide through Joiner's notion of thwarted belongingness. 84A recent qualitative study suggested that influence at work (a key aspect of job control) was a marker of being a valued member of the workplace community, again linking psychosocial working conditions to the construct of belongingness. 131Further, poor supervisor support or supervisor bullying or incivility could contribute to perceived burdensomeness.In short, there are plausible roles for adverse working conditions in each of Joiner's three constructs in the interpersonal theory.Access to means is a factor common to most theories or models of suicide, which has clear relevance when this is enabled through one's job.A more detailed discussion of work-related factors and suicide theories and models is provided in the 2022 UK Health & Safety Executive Expert Committee review 16 ; the authors of this report outline a number of ways in which work-related factors could play roles in the integrated motivational-volitional model, 132 which has some overlap with Joiner's model.Finally, a theory of work-related suicide was proposed in 2021. 15In short, deferring to these and other previously published reviews, [15][16][17] there is consistency between various theories and a role for work and working conditions as contributing causes to suicidal behaviour.
Although we did not conduct a systematic review of the evidence and we may have missed some studies, we are confident that our international and multi-disciplinary expert team of authors identified the major studies linking working conditions to suicide.A major limitation of the current evidence base is that most is based on populations in high income countries.Strikingly, 80% of suicide deaths occur in LMIC, but less than 15% of suicide research is based on LMIC populations. 3This represents a substantial research gap that urgently needs to be addressed.
Finally, there are clearly gaps in the evidence base both in aetiology and intervention.These gaps constitute further limitations and have been highlighted throughout this Commentary.

| Implications for policy & practice
The findings of this Commentary have several implications for policy and practice.

| Workplace prevention & control
To the extent that specific working conditions contribute to suicide risk, they should be managed, controlled and where "feasible" or "practicable," prevented in work settings.This follows the precedent and principles of OH&S intervention and regulation in high income counties. 133The caveat of where "feasible" (or 'practicable') acknowledges that in some circumstances, adverse exposures are unavoidable, such as exposure to trauma among emergency responders.Standard OH&S intervention principles would apply: preferencing work-directed elimination of exposures, followed by reduction of exposures, then worker-directed measures to moderate the impacts of exposures, and finally illness-directed measures to minimise the adverse impacts of illness. 134ychosocial working conditions, long working hours and precarious employment are the most prevalent work-related risk factors for suicide when considering their overlap with normative work environments and culture.05]120 The association of these risk factors with suicide powerfully reinforces the already established need for exposure prevention and control efforts from an OH&S perspective.
Further, best practice strategies are available to improve psychosocial working conditions, and prevent exposure-associated adverse impacts on health 120,[135][136][137][138] These strategies would apply equally to preventing exposure-associated suicide risk.
There has been massive growth in workplace mental health programs over the last two decades, including in workplace suicide prevention.A 2015 systematic review 142 updated in 2018 14 identified 13 published examples of workplace suicide prevention programs, only five of which included some form of evaluation.Evaluated programs provided evidence of improvements in suicide awareness and literacy, attitudes towards suicide (including stigma), and helping behaviours, and one large-scale program in the United States Airforce showed an implementation-associated decline in the suicide rate. 143A separate systematic review of programs for emergency and protective services workers included a metaanalysis of five studies, showing an approximate halving of the suicide rate over an average follow-up period of 5.25 years. 144There were few programs in which restrictions to access to means have been evaluated.In a Montreal police force program, 145 improvements were reported in supervisor's willingness to remove an officer's service revolver as a suicide prevention measure; further, an estimated 15% of supervisors who intervened in a crisis reported removing the officer's service revolver.Such measures, however, must be balanced with the risk of potentially stigmatising the affected worker and discouraging workers in distress from seeking help.
Nevertheless, this multi-component program has shown sustained reductions in suicide rates compared to non-intervention comparison departments over more than two decades of follow-up. 146th respect to access to means, pesticides warrant particular mention as a means to take one's own life, for people accessing them through work or otherwise.Intentional self-poisoning with pesticides accounts for approximately one in five suicides globally, pesticides are the most common suicide method used in LMIC, and the majority of suicides worldwide occur in LMIC. 3,129Hence efforts to eliminate, reduce the use of, and restrict access to pesticides is of paramount concern for both occupational health and general population health, particularly in LMIC. 28,129The data from trials of safe storage approaches are not as positive as anticipated but workplaces may prove a more effective setting. 147This aligns with OH&S authorities or regulators in some jurisdictions including an extended purview to protect the public from risks arising from work or workplaces.
In summary, most workplace suicide prevention programs to date have utilised the workplace as a setting for general suicide prevention, usually without addressing working conditions.These programs contribute importantly to suicide prevention independent of cause, and could work in concert, or in parallel, with work-directed approaches to reduce adverse working conditions. 121,139The importance of addressing working conditions has been acknowledged in policy advice as early as the 2006 WHO Guidelines for Workplace Suicide, 83 but has yet to be realised in prevalent practice.Improving the focus on working conditions should be a priority for workplace mental health in general, and for workplace suicide prevention in particular.There are growing calls to address the upstream social, structural and other determinants of suicide 148 -addressing working conditions as one such determinant would be a positive step in this direction.
Post-incident response, or "postvention," entails interventions following a suicide or other critical incident or traumatic fatality to alleviate the distress of co-workers and others who have been affected by the death, to reduce the risk of further suicides in the same group, and to promote the recovery of the affected group. 149A 2021 rapid review identified five critical incident response programs in workplace settings. 149Due to the small evidence base and other limitations, the review did not provide evidence of the impacts of postvention interventions on adverse outcomes for workers or organisations.While the need for such programs is clear, particularly in sectors with anticipated exposure to trauma, care is also needed to avoid repeating previous mistakes made with post-incident debriefing, where a wellmeaning intervention proved to be unhelpful and possibly harmful due to the disruption of people's usual coping mechanisms. 150Pearce et al. 149 provide principles-based recommendations for postvention planning for pre-incident, incident, and post-incident aspects.
Finally, the schema of potential work-related causes (Figure 1) can be applied in the development of workplace programs, particularly at the risk assessment stage.

| OH&S authority investigation & surveillance of work-related suicide attempts and deaths
Suicide attempts or deaths that occur at work should trigger immediate investigation by OH&S regulators in similar fashion to any other traumatic injury or critical incident; this is standard practice in many high income countries. 151When suicide attempts or deaths occurring outside of work may be work-related, investigation by OH&S regulators should also be triggered. 152In the latter instance, suspicion of work-relatedness would need to be determined by police, coroners or other investigators, and OH&S regulators would need to be actively notified.The focus of the OH&S regulator would be on potential work-related causes of suicide (Figure 1), as suggested by findings from preceding death investigations.OH&S regulator investigations would, ideally, precede and inform final determinations of death by intentional self-harm by coroners or other authorities, and their focus on work-related hazards would complement both police and coronial investigations (e.g., could draw on JEMs to determine exposure history for some exposures).OH&S regulator investigations would also be essential to closing the workplace health surveillance loop, wherein the occurrence of work-related injury or death, should feed back to reducing exposure to the identified occupational hazards.This could occur through jurisdiction-based enforcement actions, educational campaigns, or other intervention strategies, and would not be solely restricted to the affected workplaces. 153llowing potential work-related suicide investigations, those deemed work-related should be reported to national occupational injury or disease registries and surveillance systems to facilitate national monitoring of trends in relation to policy and practice developments, and on-going research. 152,154

| Workers' compensation
Work-related suicide is already acknowledged and eligible for workers' compensation in many jurisdictions around the world.These include France, Japan, Canada, and Australia. 22,23,30,155Detailed discussion of the compensability of work-related suicide is beyond the scope of this review, but to the extent that work-related causes contribute to a given suicide, it would seem appropriate that such deaths are covered by workers' compensation.Based on the evidence detailed above, workers compensation systems also require reform to decrease the potential for harm due to adversarial aspects of claims processes as a risk factor for suicide and adverse impacts on mental health. 85

| Conclusions
Work-related suicide likely represents a substantial preventable burden in the working population.We have summarised our review findings in a schema of potential work-related causes of suicide that can be applied in risk assessment and prevention programs as well as case investigations.The evidence linking psychosocial working conditions to suicide, in particular, has grown stronger in recent years.Given that these are common exposures, psychosocial working conditions could be substantial contributors to preventable workrelated suicides.Multiple intervention strategies, based on both principles and evidence, are available to address this problem.Some of these strategies have been shown to improve suicide prevention literacy, reduce stigma, enhance helping behaviours, and in some instances have been associated with reduced suicide rates.Current workplace suicide prevention policy and practice, however, overemphasises individual-and illness-directed interventions and devotes too little attention to addressing working conditions that increase suicide risk.A stronger emphasis on improving working conditions will be required in order for workplace suicide prevention to reach its full preventive potential.

2 |
CASE-BASED APPROACHES TO DETERMINE WORK-RELATEDNESS Deaths by suicide are routinely investigated and documented by police, healthcare personnel, and coronial authorities.Depending upon the jurisdiction, medical examiners, coroners or police may make the final determination of a death as occurring through intentional self-harm, 3 often generating extensive documentation which can be investigated for contributing causes.Work-relatedness can be suggested by many factors such as attribution to work by the deceased (e.g., in a suicide note) or by family, friends or co-workers of the deceased; a work location of the death; suicide being a consequence of a work-related injury or illness; or accessing means of suicide at or through work.Based on case records, various categories of work-relatedness have been proposed.F I G U R E 1 Schema of potential work-related causes of suicide.[Color figure can be viewed at wileyonlinelibrary.com]Work-related suicide has long been recognised in Japan, usually in relation to excessive working hours or job stress.Suicide due to overwork or job stress is termed karo-jisatsu, as distinct from sudden death from overwork, termed karoshi. 19-21A Japanese National Police Report (National Police Agency, Government of Japan, cited in Yamauchi et al. 22 ) stated that there were 24,025 deaths by suicide in Japan in 2015.The reasons for suicide were determined by police in 75% of those deaths, and 12% of those involved "work-related issues."

"
Stressors" in this context referred to any form of adverse exposure arising from work, including workplace injuries.They found that 11.7% of suicides were work-related.Of work-related cases, most involved stressors related to work (86%), 21.8% involved workrelated means, 19.3% were not employed at the time of death, and 10.2% occurred in the deceased person's workplace and met one of the two criteria above (only half of the cases occurring in the deceased's workplace were deemed to be work-related).

3. 6 |
Exposure to suicide, other deaths, or trauma on the job Occupational exposure to trauma entails witnessing bodily injury or death, including suicide, in the course of one's work.Such exposures commonly occur among emergency responders, members of the military, social workers, mental healthcare providers, and others.A might not be precipitated by work-related factors.In some, but not all, contexts, work-related means and other work-related risk factors co-occur, such as among trauma-exposed emergency responders with access to firearms or medicines, and veterinarians accustomed to relieving suffering with euthanising drugs and access to such drugs.Whether combined with other work-related risk factors or not, control of access to work-related means warrants careful attention in workplace risk assessment and suicide prevention.

4. 3 |
Box F1: Exposure to adverse chemical, physical, or psychosocial and other working conditions