ABSTRACT
Background Work disability occurs when an injury, illness or other health condition limits the ability of a worker to participate in paid employment. Several lines of evidence suggest that people with work disability are also at increased risk of suicide and intentional self-harm, however the evidence on this relationship has not previously been the subject of systematic review.
Objective This systematic scoping review aims to assess and summarise the research literature regarding the relationship between work disability and subsequent suicide or intentional self-harm.
Methods Peer-reviewed quantitative and qualitative studies published in English from 1st January 2000 were included if they reported suicide or self-harm outcomes in people aged 15 years or older with work disability. Narrative synthesis summarised findings according to the system of work disability income support.
Results Literature search yielded 859 records of which 47 eligible studies were included, including nine set in workers’ compensation, 20 in sickness absence, 13 in disability pension systems, and five from mixed cohorts. Of 44 quantitative studies, 41 reported a positive relationship between work disability and suicidal behaviour. The relationship is observed consistently across nations, work disability income support systems and in people with a range of health conditions. Several factors elevate risk of suicidal behaviour, including presence of mental health conditions and longer work disability duration. There were few quality studies in some nations and no suicide prevention interventions.
Conclusion The risk of suicide and intentional self-harm is elevated in people experiencing work disability. While further observational research is required to fill evidence gaps, this review suggests the need for governments, employers and those involved in the delivery of care and support to people with work disability to focus on identification and monitoring of those at greatest risk of suicidal behaviour, as well as suicide prevention.
INTRODUCTION
Work disability occurs when an injury, illness or other health condition limits the ability of a worker to participate in paid employment. Work disability is a broad concept, encompassing partial and temporary loss of work capacity, episodic changes in work capacity, through to total and permanent incapacity(1). Common diseases and illnesses of working age are the major causes of work disability, and include conditions with high prevalence such as low back pain, depression, anxiety and traumatic injury. The consequences of work disability can be significant for workers. People with prolonged work disability suffer worse mental health(2–4), have shorter life expectancy(5), attend healthcare consultations more frequently with physical symptoms and report higher levels of pain(6), receive more social care(7), and report reduced quality of life(8). There are also gaps in health service delivery. For example one study of disability pension recipients identified that although 69% reported a diagnosed mental health problem, only 16% reported receiving specialist mental health care(9), a finding reflected in similar studies of workers’ compensation cohorts(2, 10).
In addition to treatment for the underlying health condition, people with work disability often require financial support for income losses incurred during periods of time away from work. In nations with developed economies, multiple forms of income support are often available. These can include sickness absence (i.e., sick leave) benefits for temporary work disability offered through employers and sometimes government agencies, and disability benefits for people with permanent or long-term work disability. Some nations also offer financial support through workers’ compensation schemes where work is the cause of work disability. Provision of financial support for people with work disability is very common. For example, a recent study in the authors’ country of Australia reported that for every 1000 working age Australians, an estimated 49.4 received income support for a period of work disability from a government or private sector (i.e., not their employer) benefit scheme during the 2015/16 year (11).
Several lines of evidence suggest that people with work disability are also at increased risk of suicide and intentional self-harm. First, work disability by definition results in a period of detachment from the workplace, and in some cases may result in unemployment. Unemployment both impairs mental health (12) and is associated with an increased risk of suicide, with the greatest risk within the first five years of unemployment (13). Second, several studies show that a substantial proportion of people with work disability have mental health conditions, including people in whom the episode of work disability is linked to a physical injury/condition. Approximately 30% to 50% of people with workers’ compensation claims for musculoskeletal disorders reported moderate to severe psychological distress (2, 3), while mental health conditions are now the most common medical condition among Australians receiving social assistance disability benefits (14). Third, people with work disability are often involved in administrative benefit systems such as workers’ compensation or social assistance schemes. A large proportion of people experience the bureaucratic processes involved in eligibility determination and benefit delivery to be stressful (15, 16). Qualitative studies demonstrate that in some people, these administrative processes may lead to long-term mental health problems including suicidal ideation, and reduced quality of life (8, 17). Fourth, work disability is associated with financial distress. People receiving work disability benefits from workers’ compensation and social assistance schemes report high levels of financial distress (18). Financial hardship is frequently cited as a risk factor for suicidal behaviour. Fifth, long periods of work disability are associated with changes in social support networks, including increasing the burden on caregivers and changing the nature of intimate relationships (19). Social support is a protective factor that reduces the risk of suicide death in adults (20) that may be adversely impacted during periods of work disability. Finally, many people with work disability and concurrent mental health problems do not receive appropriate mental health treatment, and thus the risk of self-harm or future suicide is more likely to go unrecognised (2, 7).
Suicide and intentional self-harm are major public health problems that affect all ages, especially people of working age. In the author’s home nation of Australia, suicide is the leading cause of death in those aged 15 to 44 years, and the third leading cause of premature death (21). There are more than 33,000 cases of hospitalisation for intentional self-harm annually, with the prevalence higher in younger people and females. People with work disability have multiple risk factors for suicide and suicidal behaviour, some of which are modifiable. While the existing evidence is highly suggestive of a link between work disability and later suicide and self-harm, the evidence-base is disaggregated. Studies have been conducted using a variety of methods, in multiple cultural and societal contexts, using a variety of key concept and outcome definitions, and across nations with very differing approaches to supporting people with work disability. There is a need for a scoping study to examine the relationship between work disability and subsequent suicide and self-harm.
The objective of this systematic scoping review is to assess and summarise the research literature regarding the relationship between work disability and subsequent suicide or deliberate self-harm. The review also seeks to answer the question “What personal, psychological, social, medical, environmental or other factors influence the relationship between work disability and subsequent suicide or deliberate self-harm?”.
MATERIALS AND METHODS
This is a systematic scoping review based on the method outlined by Arskey and O’Malley(22)(22) and extended by Levac et al(23)(23). A protocol for this review was published on the Open Science Foundation website on 29th April 2021 and is available via the following link: https://osf.io/kh5qf/
Eligibility Criteria
Peer-reviewed research studies published in English from 1st January 2000 were eligible for inclusion if they reported suicide or self-harm outcomes in people aged 15 years or older who experienced an episode or episodes of work disability. Work disability was defined as a complete or partial incapacity to work due to an injury, illness or medical condition that can be temporary or permanent, and resulted in receipt of a financial payment from a third party such as an employer, insurance provider or government agency for a period of absence from work. This could include, for example, a workers’ compensation income support benefit, a disability benefit, or a sickness absence payment. Studies reporting samples with injury, illness or medical conditions that affect work capacity but that did not result in a period of work absence were excluded. Suicide and self-harm outcomes were defined as taking or attempting to take one’s own life, intentionally harming oneself including suicide attempts, and suicidal thoughts or ideation. Studies that reported on death, injury or harm that arose from the acts of others, by illness or by accidental causes were excluded. No limits on settings or geographic locations were applied.
Consistent with the exploratory nature of the review, we included both qualitative and quantitative primary studies. For quantitative studies, experimental and quasi-experimental study designs including randomized controlled trials, non-randomized controlled trials, before and after studies and interrupted time-series studies were eligible for inclusion. In addition, analytical observational studies including prospective and retrospective cohort studies, case-control studies and analytical cross-sectional studies were considered eligible. Qualitative studies from a variety of theoretical and methodological approaches were eligible including phenomenological studies, grounded theory, ethnography, qualitative description and action research studies. Case reports, opinion pieces, commentary and literature reviews were excluded from consideration.
Search Strategy
An initial limited search of PubMed was undertaken to identify articles on the topic. The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles were used to develop a full search strategy (Supplementary Table 1). The search strategy included a combination of population terms for work disability combined with Boolean OR operators, and outcome terms for suicide and self-harm combined with Boolean OR operators. These two categories were combined with AND operators in final searches.
Search terms were purposefully broad with limited exclusions to ensure all relevant literature was captured. The search strategy, including all identified keywords and index terms, were adapted for each included database or information source, and strategies for each database were developed and refined iteratively. Final searches were conducted in Medline, Scopus and Pubmed on 13th May 2021.
Reference chaining was also conducted, with potentially relevant references identified by a single author (AC) and included in evidence screening. Forward citation searches of included studies were conducted using academic literature databases Scopus and Medline. Backward citation searches from reference lists of included studies were also conducted.
Finally, ten expert academic researchers from North America, Europe and Australia with publications relevant to the topic were contacted and invited to identify studies for inclusion in evidence screening. These topic experts were provided with a summary of the purpose of the review, research questions and the eligibility criteria to guide their recommendations. An overview of the search and article selection process is provided in Figure 1.
PRISMA Flowchart
Selection of sources of evidence
Records identified through database searching, through reference chaining and from expert recommendations were collated and uploaded into EndNote version 9.3 (Clarivate Analytics, PA, USA). Records were compared across data sources and duplicates removed. Article screening proceeded in two major steps. First, titles and abstracts for all records were assessed against the eligibility criteria by two independent reviewers. Records on which both reviewers agreed were excluded or progressed to the next stage. Disagreements between reviewer eligibility ratings, or records on which one or either reviewer could not reach an eligibility decision, were resolved initially through discussion. In cases where agreement could not be reached through discussion, a third reviewer was engaged to make a final decision.
Second, all records that passed the first phase screening were retrieved in full. The full text of documents was assessed by two reviewers independently against the eligibility criteria. Disagreements between reviewer eligibility ratings, or records on which one or either reviewer could not reach an eligibility decision, were resolved through discussion. Reasons for exclusion of sources of evidence at full text were recorded and are reported in the PRISMA-ScR flow diagram presented in Figure 1.
Data extraction
Data was extracted from eligible records using a data extraction tool developed by the reviewers. The data extracted included specific details about the participants, concept, context, study methods and key findings relevant to the review question/s. Data fields extracted included study title, authors, journal, year of publication, country of origin, study aim/objective, methodological design, inception period, follow-up period, age distribution, sex/gender distribution, sample size, nature of health condition leading to work disability, type of work disability benefit system (workers’ compensation, disability insurance, sickness absence), description of suicide or self-harm outcome/s, statistical analysis method, covariates tested in statistical modelling, whether pre-existing health status was included in statistical adjustment, estimates of statistical effect size, the authors’ conclusions regarding the relationship between work disability and outcomes tested, implications for policy and practice, and any gaps identified by study authors. For qualitative studies the major themes relating to suicide and self-harm were extracted and fields related to statistical analysis and effect size were not able to be extracted. Data extraction was initially completed on five eligible studies by a single reviewer (AC). Extracted data was then reviewed by a second reviewer to confirm that all relevant information was extracted. A single reviewer then completed data extraction on the remaining eligible studies.
Evidence synthesis
A descriptive numerical summary of included studies was developed and presented in a tabular format for discussion amongst the authors. The synthesis tables described the numbers of studies by key study features including by study design, country of origin, nature of condition leading to work disability and type of disability benefit system. Inspection of the summary identified three distinct categories of evidence reflecting three unique contexts in which income support was provided to people with work disability, and also reflecting qualitatively different forms of work disability. First, there were a set of eligible studies set in cause-based workers’ compensation systems, in which the injury or illness leading to work disability must be attributable to employment. Second, there were a set of eligible studies set in disability-based pension systems in which people typically have long-term or permanent work disability. A third set of eligible studies were set in sickness absence systems in which people typically have shorter or temporary periods of work disability. There was also a set of studies that included participants receiving either sickness absence or disability pension benefits and these were combined into a fourth category.
For each of these categories a narrative summary was developed, describing the relationship between work disability and suicide (study main objective), and any factors that contribute to or modify this relationship (key research question). Within each category, studies were grouped and counted according to the study outcome, design, the health condition leading to work disability, by country of origin and other features. Gaps in the current evidence base will also be identified and described in order to identify opportunities for future research.
RESULTS
The literature database search yielded 859 records, with an additional 24 records identified through reference chaining and expert recommendations. After removing duplicates, a total of 524 unique records proceeded to screening. Application of eligibility criteria in title and abstract screening removed 457 records. A further 18 were considered ineligible during full-text screening, and the remaining 47 records proceeded to data extraction and synthesis.
Of the included studies, 35 reported suicide outcomes, 18 intentional self-harm and a further 8 included suicidal ideation, noting that some studies reported outcomes across multiple of these categories. A single study reported a composite outcome that incorporated suicide with intentional self-harm. In nine studies participants were involved in workers’ compensation benefit systems, in 20 participants were receiving sickness absence benefits during their work disability, in 13 participants received disability pension benefits, and in a further five studies participants receiving either sickness absence or disability pension benefits were combined in the exposure group. Thirty-two of the eligible studies reported an objective of assessing the relationship between work disability and later suicide or self-harm. In the remaining 15 studies this relationship was reported in statistical analysis but was not an explicit objective of the study.
Methods were heterogenous, with 31 studies reporting either prospective or retrospective cohort designs, 6 case-control studies, 5 cross-sectional studies, 3 qualitative studies and a further 2 ecological studies. Thirty-five of the studies involved linkage of data between two separate data sources, and thirty-seven included data from more than one source. The most common data sources were administrative records (N=37, e.g., social security payment records, workers’ compensation records) and electronic health records (N=32, e.g., hospital records), while 9 reported data collected via participant or healthcare provider surveys, and 3 from participant interviews. Among the 44 quantitative studies, the primary statistical analysis methods included survival analysis (N=22), logistic regression (N=11), Poisson regression (N=2), linear regression (N=2) and calculation of standardised mortality ratio (N=3). A further 7 studies reported other primary analysis methods including trajectory modelling, cumulative risk estimate, relative risk, descriptive statistics and the chi-square statistic. The 3 qualitative studies reported results of thematic analysis.
The majority of included studies (N=28) did not focus on populations with specific health conditions. Eleven studies were conducted in samples with mental health conditions, a further four in samples with musculoskeletal disorders and four studies in samples with specific medical conditions. In 21 studies an indicator of health status prior to the onset of work disability was included as a confounder in the final statistical model. Forty of the eligible studies were in community-based samples, 4 in groups defined on the basis of their occupation, and 3 in groups defined on the basis of their medical/hospital admission status. Three studies were focused on young adults, one on middle aged adults and the remainder included samples across the working age spectrum.
Studies in workers’ compensation benefit recipients
Nine eligible studies reported relevant outcomes in people with work disability who had received financial support from workers’ compensation schemes. The studies are summarised in Table 1. Five of the studies reported suicide outcomes, four suicidal ideation and a single reported intentional self-harm. Studies were from five countries, including three from each of the United States and Canada, and one each from Australia, Korea and Taiwan. Eight of the studies included samples of people with musculoskeletal disorders (MSD), with three focused solely on this condition and five also including people with accepted workers’ compensation claims for other conditions. A single study reported on people with traumatic injury and specifically excluded those with gradual onset MSD or intentional self-harm as the cause of work disability.
Summary of workers’ compensation studies
Three of the studies were retrospective cohorts involving analysis of administrative data such as workers’ compensation claims records, death registry data and social security data. A single study used electronic health records. There were a further two prospective cohort studies using administrative data. Four of the cohort studies had follow-up periods exceeding 5 years, and the follow-up period in the fifth study was not stated. There was a single cross-sectional study using survey data, and three qualitative studies involving interviews with injured workers. Four studies involved linkage of data sources (e.g., workers’ compensation records and death registry data) and the remainder reported data from a single source.
Two studies from the USA state-based workers’ compensation schemes report greater risk of suicide among injured workers with compensation claims involving time off work compared with workers whose claims were for medical expenses only (24, 25). Applebaum et al (24) reported a retrospective cohort study of 100,806 adults with illness and injury acquired in the course of work between 1994 and 2000. Over a median 16 years follow-up period, those with at least 7 calendar days of lost time following their injury demonstrated a 92% and 72% increase in mortality hazard due to suicide for women and men, respectively. In a similar study, Martin et al (25) reported an 85% increase in suicide mortality hazard among workers with low back pain during a study with an 18-year follow-up period, among workers injured in 1998 or 1999. One Korean and one Taiwanese study reported an approximate doubling of suicide rates among adults with workers’ compensation claims following musculoskeletal conditions (26) or traumatic injury (27) compared with national suicide mortality reference data. A single cross-sectional study of rehabilitation patients with chronic or acute pain conditions receiving workers’ compensation reported significantly higher relative risk of suicidal ideation compared to similar patients not receiving workers’ compensation (28). For example, the relative risk (RR) of “wanting to die because of pain” was 4.22 (2.67,6.68). An Australian death register study identified that 29% of workers who had died by suicide in the state of Victoria were involved in a workers’ compensation claim (29).
The literature search identified three qualitative studies in which workers reported suicidal thoughts (17, 30, 31). These were all conducted in Canadian workers’ compensation settings. The author of one study reports a surprisingly high frequency of discussions of suicide among included workers, and also that other participants in workers compensation such as lawyers also report suicide risk among their workers clients. Workers involved in these studies link suicidal thoughts to their experiences of prolonged and serious disability, precarious personal and financial situations, and also to procedures, policies and attitudes of compensation agencies such as conducting surveillance and complex and dysfunctional administrative processes.
Some notable gaps in this literature include a small number of studies with longitudinal designs, undertaken in a small number of jurisdictions; and the lack of studies on workers with mental health condition claims.
Studies in sickness absence benefit recipients
Twenty eligible studies reported relevant outcomes in people with work disability who had received sickness absence benefits. The studies are summarised in Table 2. Fifteen studies reported suicide outcomes, seven intentional self-harm, 2 suicidal ideation and a single reported a composite outcome including suicide, self-harm and ideation. Studies were from eight countries, including eleven from Sweden, three from Norway and a single study from each of Finland, Japan, France, Spain, Slovenia and Brazil. Thirteen studies did not specify a specific condition and included samples of people with sickness absence from diverse causes, six studies included samples of people with Mental Health Conditions (MHC) and one an MSD sample.
Summary of sickness absence studies
Eleven prospective cohort studies and a further two retrospective cohort studies involved analysis of linked datasets such as social security records, death registry data and hospital records. A further two prospective cohort studies involved linkage of employer records with death registry data. Two cross-sectional studies involved collection and analysis of participant questionnaire and employer records. There were three case-control studies and these also involved analysis of multiple linked data sources. Finally, a single ecological study examined municipal level data from multiple, unlinked sources. Among the longitudinal studies, follow-up periods ranged from 2 to 19 years, with 6 studies having follow-up periods longer than 5 years. Overall, these studies demonstrate an elevated risk of suicide and self-harm among people with a sickness absence spell. Nineteen of the 20 eligible studies observed positive associations between sickness absence and later suicide or self-harm.
Eight prospective or retrospective cohort studies showed a higher rate of suicide during follow-up among people with a sickness absence spell during the study inception period, compared to those without sickness absence (32–39). Multiple of these studies observed that people with longer sickness absence spells, and those with multiple spells, are at increased risk of suicide than those with short or no absences (32, 33, 36, 38), and that those whose absence is due to mental health conditions are at increased risk of suicide (34–38). Two of these cohort studies observed similar associations between mental health related sickness absence, and sick leave duration, and intentional self-harm (38, 39), while other studies also observed that there was increased rate of suicide or intentional self-harm in people with sickness absence due to musculoskeletal disorders (38) (35) or other conditions including injury/poisoning, digestive, circulatory, nervous and respiratory disease (38).
Three retrospective case-control studies observed that a prior sickness absence spell was a significant risk factor for hospitalisation for self-harm or suicide (40–42), although this effect was only significant among males in one of these studies (42). Two studies identified predictors of suicide or self-harm among cohorts of people with sickness absence as including absence due to mental health conditions, a history of healthcare due to mental illness, young age, low education, being a single parent and having a high number of sickness absence days (43, 44). One study identified predictors of suicide among adult residents of Stockholm as including longer duration of sickness absence, with a stronger effect in men than women (45). Another study of people receiving healthcare for depression during 2005 identified elevated risk of later self-harm resulting in hospitalisation among people with new, full-time or long sickness absence spells (46). Three studies were conducted in specific occupational cohorts. These identified that physicians and nurses with a high risk of suicide were more frequently on sick leave when admitted to an inpatient treatment program (47); that government employees with more than one sickness absence spell exceeding 3 days had more than 7 times higher mortality from suicide, while employees with more than 15 days sickness absence per annum had 3 times the odds of suicide (48); and that employees of a healthcare organisation with sick leave during 2017 or 2018 had seven times the odds of recording a composite outcome encompassing suicide, self-harm or suicidal ideation, than employees with no sick leave spells (49).
A single population based ecological study in Slovenia observed a univariate relationship between sickness absence duration per capita and the rate of suicide at a local area level, however this effect was not observed in statistical models that adjusted for area level socioeconomic and health service characteristics (50). Finally, a single cross-sectional study reported that taking sick leave was associated with decreased odds of current suicidal ideation in people receiving outpatient treatment for depression (51).
Some notable gaps in this literature include relatively few (N=5) studies outside Nordic countries and no studies in North America or Asia; few studies in specific occupational settings or groups, relatively fewer studies examining outcomes of suicidal ideation or self-harm; and a lack of qualitative studies.
Studies in disability pension recipients
Thirteen eligible studies reported relevant outcomes in people with work disability who had received a disability pension. The studies are summarised in Table 3. Eleven studies reported suicide outcomes, seven intentional self-harm and 2 suicidal ideation. Studies were from six countries, including seven from Sweden, two from Finland and a single study from each of the United States, Australia, Germany and England. Seven studies did not specify a specific condition and included samples of people with work disability from diverse causes while four studies included samples of people with MHC.
Summary of disability pension studies
Eight prospective cohort studies involved analysis of linked datasets such as social security records, death registry data and hospital records. Two cross-sectional studies involved collection and analysis of participant questionnaire and medical records. There were two case-control studies and these also involved analysis of medical care eligibility and payment files, and key informant interviews. Finally, a single ecological study examined area level data from multiple, unlinked sources. Among the longitudinal studies, follow-up periods ranged from 5 to 14 years. Overall, these studies demonstrate an elevated risk of suicide and self-harm among people with work disability receiving a disability pension.
Four prospective cohort studies showed a higher rate of suicide or self-harm during follow-up among people receiving a disability pension during the study inception period, compared to those not receiving a disability pension (52–55). All of these studies observed that people with work disability due to mental health conditions are at greater risk of suicide, while one also reported elevated risk of self-harm in those with mental health conditions (54), and three reported elevated risk of suicide among people with DP due to other conditions including musculoskeletal disorders (52), and suicide attempt in young people with somatic disorders (53, 54). A single prospective cohort study observed significantly decreased suicide mortality among people with schizophrenia receiving a DP compared to people with schizophrenia not receiving a DP (56), while a further prospective cohort study observed no association between receiving a DP in 2004 and suicide or self-harm among people with multiple sclerosis during a 6 year follow-up period (57).
Two further prospective cohort studies examined predictors of self-harm and suicide among Swedes receiving DP with common mental health conditions (58, 59). These studies observe a higher risk of suicide and suicide attempt among those with prior suicide attempt, living alone and who had been prescribed antidepressants or anxiolytic medicines. Additional risk factors for suicide included female sex, younger age and history of inpatient care, while males had higher risk of suicide mortality. People whose primary DP diagnoses was depressive disorders had the greatest risk of suicide and suicide attempt. Two cross sectional studies observed that people receiving DP were more likely to report a history of suicidal thoughts or intentional self-harm. These included one study of Australian DP recipients (60) and a second study of Swedish DP recipients with Usher syndrome (61).
Two retrospective case-control studies observed greater odds of suicide in low income people receiving DP compared to those not receiving DP (62), and an increased risk of suicide among German adults receiving a DP compared to a healthy community control group (63). Finally, a single ecological study conducted in England observed an increase in rates of suicide in those local areas where a greater proportion of the population had their eligibility for DP benefits re-assessed via a work capacity assessment (64).
Some notable gaps in this literature include few studies outside Nordic countries, few studies in specific occupational settings or groups, relatively fewer studies examining outcomes of suicidal ideation or self-harm; and a lack of qualitative studies.
Studies of mixed samples
Five eligible studies reported relevant outcomes in mixed samples of people with work disability who had received either a sickness absence payment or a disability pension. The studies are summarised in Table 4. Four reported suicide outcomes and three intentional self-harm. Four were from Sweden and the fifth from Denmark. One was in a sample of people with work disability due to MHC, and the remaining four included samples of people with work disability arising from diverse conditions. All five involved analysis of linked datasets such as social security records, death registry data and hospital records. These studies are consistent with those of the SA and DP studies described above, and provide further evidence of an association between work disability resulting in receipt of income support payments and later suicide and self-harm.
Summary of mixed sickness absence / disability pension studies
One population based prospective cohort study observed elevated risk of intentional self-harm and suicide in Swedish adults with and without MHCs receiving SA or DP benefits in 2005, compared to adults not receiving benefits (65). This study also observed a positive association between outcomes and longer durations of work disability. Niederkrodenthaler et al (66) observed elevated risk of self-harm among hospitalised adults with long-term SA, and also elevated risk of self-harm or suicide among hospitalised adults in receipt of DP. A retrospective cohort study of adults who completed suicide between 2007 and 2010 observed five different trajectories of SA/DP receipt in the five-year period preceding suicide (67). The two largest trajectory groups included nearly half of the sample who had received on average a few months/year of SA/DP benefits, and approximately a further third of the sample who had received more than 10 months/year of SA/DP benefits. A prospective twin study observed that twins with MHC who received either SA or DP benefits between 2005 and 2010 had greater risk of self-harm than their adult twins without SA/DP receipt (68). Finally, a Danish retrospective case-control study observed that receipt of SA benefits was a significant risk factor for suicide in men, but not women, but that receipt of a disability pension was not associated with suicide in either men or women.(69)
DISCUSSION
The risk of suicide and intentional self-harm is elevated in people experiencing work disability. This relationship is observed among people receiving work disability income support from workers’ compensation, disability pension and sickness absence systems; across multiple nations; and in people who are work disabled due to mental health conditions, musculoskeletal disorders, traumatic injury and other medical conditions including neurological disorders. Of the 44 quantitative studies included in this review, 41 showed a statistically significant positive relationship between work disability and suicide or self-harm. Only one study showed a negative (protective) relationship, and two showed no effect. While the relationship between work disability and suicide or self-harm is robust, in that it occurs in multiple settings and cohorts, we also observe that a range of health, social, demographic and administrative factors are associated with elevated risk of suicidal behaviour among people with work disability. These include longer durations of work disability, work disability arising from mental health conditions, younger age and female sex, a history of poor health, and living alone. Three qualitative studies in workers’ compensation systems and an ecological study of disability pension eligibility re-assessment suggest that burdensome administrative processes in income support systems may also contribute to risk of suicidal behaviour in work disabled people.
Overall, our findings suggest an opportunity for reducing suicide and self-harm among working age people, via preventive interventions focused on those experiencing work disability. In developed nations with established work disability support systems such as workers’ compensation and disability pension schemes, there are multiple opportunities for identification of those at risk, and delivery of preventive interventions along the pathway between work disability and suicide. These include at the onset of work absence, upon application for income support, during delivery of treatment and workplace rehabilitation programs, and at periods of review such as during the assessment of entitlement to ongoing benefits which are a feature of most work disability support systems in developed nations (11, 64). At each of these points workers engage with formal systems operated by government, employers or government-appointed organisations involved in case management and healthcare delivery. These represent opportunities for identification of those most at risk, and delivery of services and supports to reduce risk (70). For instance, to detect and monitor individuals at increased risk of suicide, organisations providing case management services could establish follow-up protocols for those whose work disability is due to mental health conditions. Our findings also indicate the need for more careful consideration of psychological supports and risk mitigation strategies during the process of worker rehabilitation, when there may be opportunities to address modifiable risk factors. These approaches could be adapted, for example, from existing screening and early intervention programs that seek to identify people with risk factors for long periods of work disability(71, 72).
Many of the studies included in this scoping review had features of rigorous observational study design. Multiple studies included using population-based data capture, linkage of multiple data sources across health care and work disability systems, incorporated long follow-up periods beyond the period of work disability, utilised healthy community or other comparison groups, or adjusted statistical estimates for multiple potential confounding factors including health status prior to work disability. There are also quality issues in some studies, for example studies reporting only descriptive statistics(29), cross-sectional studies (49, 51, 60, 61) and small sample sizes (29, 47, 51, 61). Studies used a range of different outcome measures, and applied variety of statistical analysis techniques. As a consequence, study outcomes were reported variously using odds ratios, hazards ratios, risk ratios, standardised mortality ratios, trajectory models and descriptive statistics. While this makes it challenging to compare the strength of relationship between work disability and suicide/self-harm outcomes between studies, some broad trends were observed.
People with work disability due to common mental health conditions were observed in multiple studies to be at higher risk of later suicidal behaviour than people with work disability due to other conditions. For example, a population based register study by Wang et al (38) reports highest hazards of both suicide and self-harm in both male and female participants with MHCs than participants with other conditions. This pattern is repeated in multiple studies in sickness absence (36, 37, 49) and disability pension systems (52–55), and for self-harm outcomes (43, 54). These findings are broadly consistent with the extensive literature demonstrating elevation of suicide risk in people with mental health conditions, and suggest that people with work-disability due to mental illness should be screened for suicide risk. We also observe elevated risk of suicide and self-harm among people with work disability due to other physical conditions, including those with musculoskeletal disorders, injury and neurological conditions. Notably, our review identified studies reporting elevated risks in people with musculoskeletal disorders in sickness absence systems (38), for women receiving the disability pension (52), in people with different MSD diagnoses (35), and for self-harm outcomes (38). It is possible that comorbidity between mental health conditions and MSDs helps to explain these findings. These studies also demonstrate the importance of identifying and monitoring groups with common physical health conditions such as MSD to identify individuals who may benefit from intervention.
Multiple studies also report that people with longer durations of work disability were at significantly elevated risk. For example, Lundin et al (45) observed that the odds of suicide over a five year follow-up period in people sickness absent for more than 62 days during 1992-93 was 8.68 for women, and 4.41 for men, relative to people with 0 to 15 days sickness absence during the inception period. Similar trends were observed in healthcare workers (49), in males by Tang et al (42), in Finnish local government workers (73), in disability pensioners (65), and for self-harm outcomes (44). Several authors propose that the duration of work disability does not only reflect the chronicity or severity of the health condition underlying work disability, but may also reflect other factors that can increase suicide risk, such as poor quality medical treatment, delayed help-seeking, presence of adverse health behaviours such as substance use and social isolation (44). Importantly, our findings also suggest that work disability prevention interventions that reduce duration of disability, and thus both promote engagement in work and potentially limit exposure to these other adverse factors, may also contribute to reduced risk of suicide and self-harm. The common features of effective work disability prevention interventions are well-documented, with multiple studies reporting effective interventions in workers with mental health conditions and musculoskeletal disorders (74). Longer-term follow-up of participants enrolled in work disability prevention trials appears warranted.
Our evidence synthesis grouped studies by the nature of the work disability system via which people were accessing income support. We chose this approach so that study findings would be relevant to those involved in administering, and providing care and support, within those systems. This approach also demonstrates that there is a stronger evidence base in some systems and nations than others. We included the largest number of studies in sickness absence and then disability pension systems, with most of these studies located in European nations. In contrast there were fewer studies included from workers’ compensation systems, and also from North American and Asian nations where such systems are more commonly a feature of government work disability support. The evidence base is larger when suicide is the measured outcome, than for self-harm as an outcome, where there are fewer studies. We identified even fewer studies reporting suicidal ideation in work disabled samples. There is value in filling these evidence gaps in order to better understand the relationship between work disability and suicide or self-harm within work disability systems and within nations. Further studies on suicidal ideation would provide critical information to explore the pathway from work disability through suicidal ideation to completed suicide, which may inform preventive interventions.
This scoping review provides an overview of the evidence relating to work disability, suicide and self-harm. We sought to draw together a diverse evidence base from across different systems of work disability support, to examine this relationship. As per accepted scoping review methods, we included a wide range of study designs, utilised systematic and transparent processes for searching and mapping retrieved studies. Our study is subject to the known limitations of scoping review methods, including that we did not appraise the quality of included studies, and that the broad scope of the research questions made it difficult to draw definitive conclusions regarding particular populations and subgroups. The studies included did not examine intermediate factors between work disability and mortality, which may include things such as opioid prescriptions, or the quality and nature of treatment, presenting an opportunity for further studies. Specifically, we did not identify any studies of suicide prevention in work disabled samples. This is a substantial gap in the evidence base given the findings of our review.
CONCLUSIONS
The risk of suicide and intentional self-harm is elevated in people experiencing work disability. Within work disabled samples, several factors may place individuals at greater risk of later suicidal behaviour, including disability due to mental health conditions and longer duration of disability. However, the relationship between work disability and suicidal behaviour is observed consistently across nations, across work disability support systems and in people with a range of underlying health conditions. While further observational research is required to further understanding and fill gaps in the evidence base, this review suggests the need for governments, employers and those involved in the delivery of care and support to people with work disability to focus on identification and monitoring of those at greatest risk of suicidal behaviour, as well as suicide prevention in the highest risk.
Data Availability
Most data produced in the present work are contained in the manuscript - see data extraction tables. Other data extracted from studies included in the review are available upon reasonable request to the authors.
ACKNOWLEDGMENTS
The authors would like to acknowledge the contribution of Ms Karki Kajal to article searching and screening. Professor Alex Collie is supported by an Australian Research Council (ARC) Future Fellowship (FT190100218). Funders played no role in the scoping review.
REFERENCES
