In the United States (US), increasing mortality among working-age (i.e., 25–64 years) adults contributed to a striking decline in US life expectancy from 2015 to 2017 (Murphy et al., 2018). This decline is largely attributed to rapidly increasing death rates from drug and alcohol poisonings, alcoholic liver disease, and suicides (Case and Deaton, 2015; Dowell et al., 2017; Hedegaard et al., 2021). Coined “deaths of despair” by Case and Deaton (Case and Deaton, 2015), these deaths are observed particularly among middle-aged non-Hispanic white men with less than a high school education (Case and Deaton, 2017). Research has increasingly linked construction, mining and extraction with elevated rates of suicide and overdose relative to other occupations and the general population (Harduar Morano et al., 2018; Peek-Asa et al., 2021; Peterson et al., 2020). In 2016, the suicide rate among males in the mining, quarrying, and extraction industries was 54.2 (95% Confidence Interval (CI): 44.0–64.3) per 100,000 persons (Peterson et al., 2020), making it the highest among all industries and more than double the general US male population rate of 21.4 (Hedegaard et al., 2018). Given that over 650,000 people were employed in the US mining and extraction industries as of 2018 (Employment and wages in mining industries, 2018), these estimates highlight the significant public health burden posed by suicide and overdose within the industry.

Since 2015, mortality rates due to illicit opioids, such as fentanyl, have outpaced those attributed to prescription opioids (Dai et al., 2019), bringing more attention to the role of declining socioeconomic conditions and increasing drug use disorders (van Draanen et al., 2020). A growing body of literature has revealed associations between economic distress and poor mental health and substance use. Of note, economic changes and industrial restructuring are inherently spatial and US regions focused on mining and extraction have been plagued by several unemployment downturns and wage stagnations in recent decades (Monnat, 2018). These macro-level labor market stressors often exacerbate the impacts that job loss already has on mental health (Milner et al., 2018) and consequently, they impose community-level traumas through lower rates of health care usage and social cohesion, and higher rates of substance use and physical injury (Monnat, 2018, 2019). Indeed, in a study of economic recession, job or financial issues preceded 13% of the suicides examined (Coope et al., 2015), while a systematic review found that, for nearly all 48 studies, a period of recession preceded increases in suicide mortality rates (Guerra and Eboreime, 2021). Although job loss and insecurity can worsen mental health, this economic research underscores the significance of structural contexts in amplifying the risk of suicide or overdose beyond the impact of job loss alone, urging ongoing consideration of these broader factors.

Life course and economics scholars stress the importance of considering both the timing of job loss within the social, political, and economic environment and within a person’s life (Gray and Dagg, 2019). Several shocks to the US economy occurred in the early 1980s, when two recessions hit in three years, and the manufacturing and mining industries never quite recovered production levels (Brandon, 2012; Plunkert, 1990). Employment in the overall mining industry dropped nearly 25%, and declined by 7% in non-metal mining specifically (Brandon, 2012; Plunkert, 1990). These boom-and-bust cycles, typical of extraction industries, are known to have adverse health effects (Boslett and Hill, 2022). This literature on the importance of place and local economic conditions on mortality and mental health (Finkelstein et al., 2021) forms the basis of our investigation of mining deaths in the context of the macroeconomic environment.

In this analysis of the Diesel Exhaust in Miners Study II (DEMS II), we compare age-adjusted suicide and overdose rates between men in the mining industry and the general US male population. To better understand the potential impact of leaving work on suicide and overdose mortality among miners, we also investigated the role of age at worker exit. To examine whether the precipitous decline in the non-metal mining industry that occurred post-1980 modified the association between age at worker exit and suicide or overdose, we further stratified our analysis into pre- and post-1980 periods.

Among the entire cohort of 12,315 workers, 84% worked in Wyoming trona mines or New Mexico potash mines, reaching peaks pre-1980 of ~3000 and 1500 employees, respectively (Fig. S1). Shortly thereafter, large decreases in employment occurred, especially in Wyoming, which dropped by 27% from 1980 to 1985. See Fig. S2 for a timeline of exiting workers by state.

Among the 11,009 men, there were 4559 deaths by the end of follow-up, including 262 deaths due to suicide (n = 178) or overdose (n = 84) (Table 1). Those who died via suicide were born earlier and worked longer than those who died via overdose. Among the 241 suicides or overdoses with known exit status at the time of death, 82% occurred after worker exit (Table 2).

Age-adjusted mortality rates per 100,000 person-years for suicide (Fig. 2) among the DEMS II male cohort are consistently higher than the US male population overall and the DEMS II mining states. Over the entire follow-up period, the DEMS II age-adjusted suicide mortality rate was 42.4 (95% CI: 35.2–50.5). Over time, the suicide rate increased from 29.3 (1940–1979) to a peak of 67.2 in the decade 1980–1989 before dropping in the next decades to 47.6. This pattern was reflected by the age-adjusted rates for suicide among mine workers at the most populous sites, Wyoming, and New Mexico (Table S2). DEMS II cohort rates for overdose remained above the general populations until 2000 and the peak occurred in the 1990s at 49.1 (95% CI: 31.8–72.7) (Fig. S3). When suicide was combined with overdose (n = 202 among workers 25–64 years), the age-adjusted mortality rate over the entire follow-up period was 65.2 (95% CI: 56.2–75.1).

The adjusted HR was elevated at 2.56 (95% CI: 1.88–3.50) for inactive compared to active workers. Additionally, suicide and overdose risk was most elevated at 1.45 for those with 1–2 years of inactive time compared to zero (95% CI: 0.85–2.48) (Table S3). Table 3 presents results from Cox models restricted to those who exited work and whose employment follow-up was complete (n cases = 187). HRs comparing each age group at worker exit to the reference group of those exiting work at 55 years or older (retirement-age exit), were null (Table 3). Follow-up was then stratified by age at worker exit pre- and post-1980 to examine effect measure modification by the socioeconomic environment. Results are generally null or protective for all age groups among those exiting pre-1980. However, post-1980, results for age at worker exit are monotonic with effect estimates increasing with decreasing age at exit. Results rose to a HR of 1.57 (95% CI: 0.83–2.96) comparing those exiting work before 30 years of age to those exiting near retirement (≥ 55 years). The ratio of HRs consistently indicated greater risk of death for those exiting post-1980 compared to pre-1980, in each age category (Table 3). Similar results are shown in Table S4 with additional adjustment for race/ethnicity.

As a secondary analysis, we examined associations between calendar period and suicide or overdose and suicide only (Table S5). Using 1947–1969 as the reference, suicide or overdose mortality rates by decade of worker exit rose to a statistically significant peak HR of 2.38 (n = 66; 95% CI: 1.28–4.43) in the 1980s before dropping to a lower but still elevated HR of 1.98 (95% CI: 0.90–4.38) in the 1990s (Table S4). Examining suicide only, our results were elevated when comparing each decade to the reference, although the confidence intervals were wider and included the null.

Suicide rates among DEMS II male mine workers were consistently higher than US males nationally or statewide from 1980 to 2015. The DEMS II rates were similar to that reported in 2016 from US mining, quarrying, and extraction industries, indicating that suicide remains a long-standing public health crisis in these industries (Peterson et al., 2020). In addition, fatal overdose rates in DEMS II were generally elevated relative to national and statewide rates. When we investigated the impact of worker exit on the risk of suicide or overdose, we found that exit at younger ages was associated with greater risk. This relationship between exit age and suicide was likely modified by the economic environment. We observed null effects pre-1980; however, post-1980, when the mining industry was struggling to recover from economic declines and facility downsizing (Brandon, 2012; Plunkert, 1990), age-stratified risks rose monotonically to a HR of 1.57 for those exiting younger than 30 years.

The profound psychiatric consequences of economic instability have been particularly observed among younger individuals who often work at lower wages and possess less experience and skill (Bell and Blanchflower, 2011). Job losses due to international trade have highly localized impacts, with layoffs, unemployment, and lower wages concentrated in specific labor markets (Autor et al., 2016). Region-specific industries are particularly exposed (Ellison et al., 2010). The loss of such jobs may pose problems for people of young ages given that prolonged youth unemployment may lead to reduced self-efficacy, feelings of hopelessness, and negative impacts on future plans – risk factors for deaths of despair (Special Issue on the Impact of the Global Economic Crisis on Youth, 2010). In short, trade-related jobs can be extremely valuable (Davis and Harrigan, 2011). However, even when new employment is found after job loss, as the age at job loss decreases, the resulting dislocation and economic shock may lead to an increased risk of suicide and overdose.

The current research on work-related psychosocial factors is limited, especially in large-scale prospective cohort studies (Kim and von dem Knesebeck, 2016). Further understanding how occupational conditions affect suicidal behavior is important, as a large proportion of all suicidal ideation, attempts, and deaths occur among working-age individuals, with especially high risk among those in construction and extraction (Han et al., 2016). Stressors, including job insecurity, are assumed to contribute to an increased risk of mental health problems due to wear on body regulation functions and/or increases in risky behaviors (Kim and von dem Knesebeck, 2016). Due in part to recent economic recessions, advancements in technology, labor market deregulation and offshoring manufacturing, US job insecurity has risen (Johnston et al., 2020). Although we were unable to measure stress or risky health behaviors as mediators, we assessed potential effect measure modification via the broader economic environment (i.e., period of increased layoffs).

Implementation of suicide and fatal overdose prevention programs in the mining industry could provide counseling and education to workers and increase awareness in this high-risk occupation. The 2012 revised National Strategy for Suicide Prevention encourages community-based settings, such as workplaces, to implement wellness and suicidality prevention programs (Office of the Surgeon & National Action Alliance for Suicide, 2012). Importantly, age differences were identified in suicide literacy in the construction (King et al., 2019) and energy (Ross et al., 2020) sectors. Younger workers reported a lack of awareness of the warning signs of suicidal ideation; however, they were also more likely to consider mental health a workplace health and safety issue. Therefore, the effectiveness of a workplace suicide awareness program may be highest if it is implemented soon after workers are hired.

The World Health Organization suggests a comprehensive workplace approach by raising suicide literacy, preventing and reducing job stress, early detection of mental health difficulties, and early intervention through employee health and assistance programs (Tiesman et al., 2015). Such workplace suicide prevention programs have demonstrated efficacy and could be highly beneficial for workers of any age. For example, a large law enforcement cohort study found a 79% decrease in suicides in the 12 years since the implementation of a suicide prevention program (Mishara and Martin, 2012). Although little is known regarding the details of and exposure to suicide and overdose awareness programs in the US mining industry, the results of this study suggest there are opportunities to implement or improve these programs, identify workplace factors that induce strain and contribute to psychological distress in employees, and restructure safety policies and practices to include mental health.

Our study has important strengths, including a large occupational cohort with nearly 70 years of follow-up. Limitations include the lack of information about the specific reasons for worker exit or any employment information beyond 1999. Suicide and overdose mortality are outcomes resulting from a complex confluence of socioeconomic, social, and individual factors. The interpretation of our results is constrained by lack of information on injury, alcohol and substance misuse, social isolation, and depression or anxiety prior to leaving work. Finally, the lack of available information on race/ethnicity limited our ability to explore how the impact of age at worker exit on suicide or overdose may differ by membership in a socially racialized and/or marginalized group.

The capacity of policy makers to identify mining-related mental health issues is limited by an absence of empirical investigation. Building on a scarcity of research, we found an association between job loss at younger ages and suicide or overdose death among male mine workers. In particular, risk was elevated post-1980, a period of decline in US mining when worker exit was likely related to facility downsizing and closure. Our findings support the need for counseling and educational awareness programs for suicide prevention in the mining industry. Such efforts should incorporate the impact of economic circumstances on suicide risk. Based on our retrospective cohort analysis, we expect that future economic recessions may have similar impacts on US mine workers if effective mental health awareness programs are not established and maintained. Finally, we recommend more comprehensive research on the impacts of workplace practices and policies such as shiftwork, contract work, salaries, and safety protocol on mine worker health. A view that covers these structural aspects of the workplace will ensure more effective development of targeted education and interventions to promote mental health in mining.