Profiles of suicidality risk: a cross-sectional study of insomnia, PTSD and psychological distress in firefighters Open Access

Emergency service workers are often employed in shift work and/or on-call/standby arrangements (Vincent et al., 2016a, b; Larsen et al., 2015). Shift work, defined as work scheduled outside the daytime hours to cover the entire 24-h day (Moreno et al., 2019), displaces sleep from the biological night, impacting sleep quality and quantity (Kecklund and Axelsson, 2016). Similarly, on-call schedules are also associated with sleep disruption (Sprajcer et al., 2018a, b; Perrin et al., 2019), with 80% of on-call workers reporting less than the recommended minimum amount of sleep during on-call periods, and 70% reporting difficulties returning to sleep after being called (Vincent et al., 2018, 2021). These sleep disruptions can have important health and social consequences (Fernandez-Mendoza and Vgontzas, 2013) which are particularly concerning for emergency services personnel (Dawson et al., 2021). For example, paramedics have higher levels of insomnia than the general population, and insomnia is a major predictor of depression and anxiety symptoms (Khan et al., 2020). In addition to sleep difficulties, emergency services personnel also experience other occupational stressors inherent to their role.

Emergency services personnel are exposed to repeated traumatic events, such as injury or death (Kyron et al., 2021), which can impact sleep. Exposure to traumatic events is related to reports of trouble falling asleep, increased awakenings throughout the night and difficulties returning to sleep after waking (Hegg-Deloye et al., 2014). Importantly, approximately 70% of people with post-traumatic stress disorder (PTSD) have co-occurring sleep problems (Babson and Feldner, 2010). Similarly, rates of PTSD and depression in current serving members of one Australian fire service were comparable to national 12-month prevalence rates in the broader Australian population (PTSD 8.0% in firefighters vs 5.7% in the Australian population; depression 5% in firefighters vs 4.6% in the Australia population (Australian Bureau of Statistics, 2021)). However, the rates for both PTSD and depression rose to 18% in retired firefighters. This is well above the national averages for PTSD and depression prevalence (5.7% and 4.6% respectively), and substantially higher than the reported PTSD prevalence of < 0.4% in Australian adults of retirement age (>65 years; Phoenix Australia – Centre for Posttraumatic Mental Health. Specific Populations and Trauma Types, 2020; Creamer and Parslow, 2008).

Evidence from other regions also indicates substantial mental health burden among firefighters. For example, anxiety disorders have been reported in Brazilian firefighters, PTSD and depression have been examined in Chinese firefighters, mental distress symptoms have been documented among firefighters in Oman and Portuguese firefighters exposed to major wildfires have shown elevated PTSD symptoms and broader psychopathology (Azevedo et al., 2019; Sun et al., 2020; Hashempour et al., 2023; Oliveira et al., 2023). Together, these findings suggest that firefighters across a range of service contexts may experience substantial mental health burden, although the type and prevalence of symptoms may vary across settings and study methods. In USA firefighters, findings revealed that about 40% are facing clinically substantial levels of anxiety and depression, and more than 10% are experiencing clinically significant levels of PTSD, exceeding rates observed in the USA general population (Gibbs et al., 2024). Of significant concern, suicidal ideation is also prevalent in emergency services.

Emergency service workers in Australia report higher rates of suicidal thoughts (5.3% compared with 2.7%) and suicidal plans (2.0% compared with 0.7%) than the general population (Kyron et al., 2021). Further, age-adjusted suicide rates over a 12-year period were 22.4 per 100,000 for males and 7.8 per 100,000 for females working in emergency or protective services (compared with 15.5 and 3.4 for males and females respectively for other occupations) (Milner et al., 2016). Factors contributing to the higher rates of suicidality among emergency services workers include workplace stressors such as exposure to traumatic events, bullying and lack of meaning in their work (Kyron et al., 2021). Understanding contributors to suicidal ideation has often been through the lens of psychological pain associated with mental health disorders like depression, anxiety and PTSD (Di Nota et al., 2020; Baryshnikov and Isometsä, 2022).

Similarly, within the emergency services sector and among military personnel, suicidal ideation has been associated with symptoms of depression (Kyron et al., 2021; Richardson et al., 2014; Sterud et al., 2008). One study found that within a population of defence personnel, PTSD and major depressive disorder were independently associated with past-year suicidality and further, personnel experiencing both disorders were three times more likely to report suicidality (Ramsawh et al., 2014). However, while studies have shown an association between sleep disorders such as insomnia and suicidal ideation (Dolsen et al., 2021; Ribeiro et al., 2012), this work has typically focused on insomnia co-morbid with mental health disorders such as depression and PTSD.

In the USA, help-seeking behaviour among professional firefighters is low, with 36.7% preferring to handle issues on their own, 25.5% reporting that they do not ask for help and 24.7% citing stigma as a concern (Gibbs et al., 2024). Related evidence from UK firefighters suggests that help-seeking often occurs through informal social networks, particularly spouses, friends and colleagues, rather than through external professional psychological support, with shorter-serving firefighters more likely to seek external professional support and longer-serving firefighters less likely to seek broader social support (Tamrakar et al., 2020). Similarly, recent evidence suggests that shiftworkers report delays of over five years in seeking help for sleep disorders, often attributing symptoms to their work schedules rather than underlying conditions (Reynolds et al., 2024a, b), which may further impede timely diagnosis and intervention (Crowther et al., 2021). Information about the relationship between insomnia and suicidal ideation in emergency services personnel is scant, presenting a potential opportunity to identify at-risk individuals who may not otherwise seek support.

Given the complex relationships between shift work, exposure to traumatic events, insomnia and mental health disorders, the primary aim was to test whether insomnia, PTSD or psychological distress, were associated with suicidal ideation. It was hypothesised that self-reported insomnia would be associated with an elevated risk of suicidal ideation in firefighters and that this association would be independent of the impact of PTSD symptom severity and psychological distress. A further aim was to inform future intervention work by establishing profiles of firefighters who may be particularly at-risk of suicidality based on insomnia, PTSD symptom severity and psychological distress.

This was a secondary analysis on previously collected data from a 2014 cross-sectional survey exploring health and wellbeing in the South Australian Metropolitan Fire Service (MFS) and the prevalence of mental disorder amongst serving members (McFarlane A et al., 2017). Participants completed a 60-min survey about prevalence and exposure to mental health symptoms, psychological distress, physical health symptoms, workplace and occupational stressors. The survey could be completed either via an online version of the survey accessed via a secure link sent by email or hard copy by request. Participants were allocated a unique study number to ensure anonymity. The study protocol was approved by the University of Adelaide Human Research Ethics Committee (H-2014-071).

Participants eligible for the research were from the South Australian MFS and included serving full-time personnel (i.e. salaried firefighters) and part-time on-call firefighters (i.e. retained firefighters). For more information on the roles of salaried and retained firefighters please see (Paterson et al., 2016; McFarlane A et al., 2017). Participants were recruited through the MFS via targeted promotional activities (newsletters and flyers in stations), briefing the MFS leadership and union representatives and a helpline where participants could discuss queries and concerns. The total number of serving members of the MFS who completed the survey was N = 578, which represented just over half of the MFS personnel at the time (McFarlane et al., 2017).

The sample characteristics are summarised in Table 1. At the time of the survey, most participants were in full-time MFS (86.9%), with the rest retained (13.1%). The majority of the sample were male (97.8%). The average age that participants joined MFS was 25.3 years, with most participants between the ages of 24 and 32 years. On average, participants served in the MFS for just over 17 years, with most participants serving for between 7 and 28 years. Participants included station officers (29.7%), senior firefighters (43.9%) and firefighters (21.0%), with few senior managers (1.4%) and commanders (4.0%). Most participants reported rotating day and night shifts (79.2%).

Insomnia was measured using the Insomnia Severity Index (ISI) (Bastien et al., 2001). The ISI is a seven item self-report scale that assesses symptoms of insomnia, as well as the impact of these symptoms on functioning. Participants are asked to indicate how severe their sleep problems were on response scales ranging from 0 (None) to 4 (Very severe), with possible scores ranging from 0 to 28. The scores can be interpreted as no insomnia (0–7); subthreshold insomnia (8–14); moderate clinical insomnia (15–21) and severe clinical insomnia (22–28) (Morin et al., 2011). The ISI has demonstrated adequate internal consistency and has been shown to be a reliable self-report measure to evaluate perceived sleep difficulties (Bastien et al., 2001). For this study, inter-item reliability was good with α = 0.77, and Guttman’s λ = 0.80.

Suicidal ideation and/or behaviour in the past 12 months was assessed using four items asking specifically about suicidal thoughts, plans and attempts, adapted from the National Survey of Mental Health and Wellbeing (Australian Bureau of Statistics, 2008). The participants were asked to answer yes or no to have they “ever felt that life was not worth living”, “felt so low they thought about committing suicide”, “have made a suicide plan” or “attempted suicide”. Due to the low frequency of reported attempted suicide (n = 1), all four items were combined so that a response of yes to any of the four questions was categorised as having experienced suicidality.

The 20-item PTSD Checklist for DSM-5 was used to assess self-reported PTSD symptom severity (Weathers et al., 2013). Participants were asked how bothered they were by a set of feelings over the previous month using response scales from 0 (not at all) to 4 (extremely). PTSD symptom severity was calculated as a sum of item responses, with a higher score indicating greater PTSD symptom severity. A score of 33 or higher was used as an indicator of probable PTSD rather than a formal diagnosis (Wortmann et al., 2016). Given the small number of participants scoring above this threshold, the variable was dichotomised into probable PTSD versus below-threshold symptom levels. This measure has demonstrated strong psychometric properties including robust internal consistency, test–retest reliability, convergent and discriminant validity and sensitivity to clinical change (Blevins et al., 2015). For this study, interitem reliability was good with α = 0.94, and Guttman’s λ = 0.96.

The study assessed levels of psychological distress using the 10-item Kessler Psychological Distress Scale (Kessler et al., 2002). Participants were asked to consider the last 4 weeks when responding to questions about how often they experienced psychological distress using scales from 1 (none of the time) to 5 (all of the time). Total psychological distress was calculated as the sum of responses to items with a range of 10–50, with higher scores indicative of greater symptom severity (Snyder, 2013). The K10 has shown to be a valid and reliable measure of psychological distress (Andrews and Slade, 2001). For this study, interitem reliability was good with α = 0.89, and Guttman’s λ = 0.90.

Data missingness was 6.6% for insomnia, 7.3% for suicidal ideation, 9.2% for PSTD symptom severity and 7.6% for psychological distress, all profiled as missing at random. Therefore, missing data were imputed with means for normally distributed data and medians for data with non-normal distributions. Insomnia was highly positively skewed (skew = 1.50) and assumption testing revealed high risk of undue influence from outliers, so it was truncated at the 75% interquartile range (=7.00) for analyses, which corrected skew to 0.01.

Data analyses were conducted using R version 4.2.3 and RStudio (R Development Core Team, 2008; RStudio Team, 2020). To test whether insomnia was associated with suicidal ideation, a binary logistic regression was estimated with suicidal ideation regressed onto insomnia, controlling for the covariates of sex, years in MFS, age when joined MFS and rank (rank categories in Table 1). To test whether any association between insomnia and suicidal ideation was present above and beyond PTSD symptom severity and psychological distress, these predictors were included in a second step of the model. A two-step cluster analysis (Milligan and Sokol, 1980; Ward Jr, 1963) of standardised insomnia, PTSD symptom severity and psychological distress was conducted to partition firefighters into profile groups. Euclidean distance was used as the similarity method and the resultant number of clusters was selected based on hierarchical cluster dendrogram, comparisons of the Dunn index (Dunn, 1974) and silhouette width for internal validity (Rousseeuw, 1987), as well as the stability of alternative cluster solutions. To test which profiles of participants represented the highest risk of suicidal ideation, an ANOVA with post-hoc Tukey difference tests were conducted.

Descriptive statistics are reported in Table 2. Participants reported sub-threshold insomnia on average, with few experiencing moderate (5.4%) or severe clinical insomnia (0.5%). Nine percent of the sample reported experiencing suicidal ideation in the past 12 months, and less than 3% reported PTSD symptoms at the level of a likely disorder. Psychological distress was low, with the average score of 14.2 on the 10–50 range scale.

As shown in Table 3, insomnia presented with similar proportions in males and females, although suicidal ideation was notably higher among females. Across years of service, insomnia ranged from 13.7% in those with 0–10 years to 31.2% in those with more than 30 years, while suicidal ideation was highest in the group who served between 11 and 20 years. Across rank, insomnia was highest among senior management and commanders and lowest among firefighters, while suicidal ideation was most frequent among mid-ranking personnel and lowest among senior management.

The results of the logistic regression assessing the association between suicidality and insomnia is presented in Table 4. Insomnia was positively associated with suicidality, with the odds ratio indicating that for every one unit increase in insomnia, the odds of suicidality increased by a factor of 1.42 (top of Table 4). However, as revealed in the second model (bottom of Table 4), the association between insomnia and suicidality did not remain when accounting for PTSD symptom severity and psychological distress. Rather, it was psychological distress that was significantly, positively associated with suicidality, with an odds ratio of 1.27. The addition of PTSD symptom severity and psychological distress significantly increased the fit of the model (χ² = 71.13, p < 0.01).

The cluster analyses revealed participants could be partitioned into 2 profiles based on insomnia, PTSD symptom severity and psychological distress (Dunn = 1.31, Silhouette = 0.16; AD = 7.81). The first profile, which represented just under half of the sample (n = 266, 46.0%) was represented by high insomnia (M = 0.94), higher PTSD symptoms (M = 0.16) and high psychological distress (M = 0.55). The second profile, which represented just over half of the sample (n = 312, 54.0%) was represented by lower insomnia (M = −0.80), lower PTSD symptom severity (M = −0.14) and lower psychological distress (M = −0.46). The ANOVA revealed that the group with high insomnia, PTSD symptom severity and psychological distress had a higher risk of suicidality than their counterparts (F = 39.15, p < 0.01, 95% CI Δ = −0.20 to −0.11).

This study investigated the association between insomnia and suicidality and whether the association was independent from PTSD symptom severity and psychological distress in a large sample of Australian firefighters. It was found that there was no unique impact of insomnia on suicidality above and beyond PTSD symptom severity or psychological distress. Additionally, profiles based on insomnia, PTSD symptom severity and psychological distress were tested with the goal of identifying people at-risk for suicidality. The cluster analyses findings showed the shared impact of these risk factors on suicidality, such that higher insomnia, PTSD symptom severity and psychological distress was associated with greater risk of suicidality.

Insomnia was positively associated with suicidality such that one unit increase in insomnia increased the odds of suicidality by 1.42. Previous literature has shown a consistent association between insomnia and suicidal ideation (Bishop et al., 2020; Dolsen et al., 2021; Ribeiro et al., 2012; Littlewood et al., 2016; Morgan et al., 2018; Richardson et al., 2014) with some research suggesting that sleep difficulties, including night-time awakenings, can make life “feel harder”, limit problem-solving opportunities and reinforce negative thoughts, potentially fostering suicidal ideation (Littlewood et al., 2016). However, the association between insomnia and suicidality in the current study was not significant after accounting for PTSD symptom severity and psychological distress. This finding suggests that the impact of insomnia on suicidality is not above that of PTSD symptoms and psychological distress. This finding is not surprising given the high comorbidity of these factors. Indeed, psychological distress was positively associated with suicidality with an odds ratio of 1.27, in line with previous research (Kyron et al., 2021; Richardson et al., 2014; Sterud et al., 2008). One possible explanation for the strong association between psychological distress and suicidality is that the presence of distressing thoughts and emotions may drive individuals towards considering suicidality as an outlet (Kyron et al., 2021).

The cluster analysis revealed that the group with high levels of insomnia, PTSD symptom severity and psychological distress had a significantly higher risk of experiencing suicidality than those classified as low for these factors. The identification of distinct profiles based on insomnia, PTSD symptom severity and psychological distress may facilitate the tailoring of interventions to better address the specific needs of firefighters who are at higher risk for suicidality. A recent analysis showed self-reported insomnia and self-reported delayed sleep timing (e.g. sleep period starting later than recommended) were both correlated with suicidal ideation (Rumble et al., 2020). Sleep health measures may be a useful predictor and/or target for early intervention in firefighters less likely to seek support for other mental health issues.

These findings also have practical implications at both service and national levels. At the service level, the results suggest that firefighter wellbeing initiatives may benefit from screening approaches that consider insomnia, PTSD symptom severity and psychological distress together, rather than as separate risk factors, to better identify those at increased risk for suicidality. This may assist with earlier identification of firefighters who would benefit from more comprehensive support, particularly where sleep-related concerns can be more readily disclosed than other mental health difficulties. At a national level, the findings support the inclusion of sleep and mental health indicators within firefighter health and wellbeing frameworks, as well as suicide prevention strategies implemented across emergency service organisations. Collectively, these results reinforce the value of integrated approaches to prevention and support that recognise the overlapping contribution of sleep and psychosocial factors to suicide risk.

Interestingly, less than 3% of firefighters in the current study recorded a score that was above the clinical threshold for PTSD. This appears to be in contrast to previous studies involving firefighters and emergency service personnel where the prevalence of PTSD has been reported at almost 10% (Harvey et al., 2016). One possible explanation is that the current sample comprised serving metropolitan firefighters, whose work context may differ from volunteer, rural, disaster-exposed or retired firefighter cohorts included in other studies, including in patterns of trauma exposure and access to organisational support. It is also possible that firefighters experiencing more severe symptoms were less likely to participate in the study, which may have contributed to a lower observed prevalence. Further, the appropriateness of the self-report measure (PCL-5) for assessing PTSD prevalence among firefighters is a key consideration. The nature of firefighters’ work involves repeated exposure to extremely traumatic events, aligning with the revised stressor criteria for PTSD (Harvey et al., 2016). A more focused assessment of PTSD within firefighters may provide better insight into prevalence and severity of symptoms, the impacts of rumination on sleep and the subsequent relationships between insomnia, PTSD and suicidality.

Although this study provides important insights into the relationships between insomnia, PTSD, psychological distress and suicidality, there are some limitations that should be acknowledged.

First, the generalisability of the results is somewhat limited as although reflective of the demographics of MFS personnel, 98% of the study sample was male.

Second, the cross-sectional design of this study does not allow for causal relationships to be identified. Longitudinal studies may assist in identifying factors associated with short-term fluctuations in suicidal ideation to understand how symptoms may change over time which may provide greater insight to predicting those at risk (Kyron et al., 2021). A recent study found that burnout and depression may modulate the impact of post-traumatic stress on sleep quality (Khoshakhlagh et al., 2024). Therefore, future research is needed to explore targeted interventions addressing these modulating factors to enhance sleep and mental health outcomes in firefighters. Longitudinal studies beginning at the outset of a career in emergency services doing shiftwork will help identify risk factors more accurately to inform workplace risk mitigation strategies (Crowther et al., 2022a, b; Meadley et al., 2022).

Third, the measures used in this study assessed the participants’ thoughts and feelings over different time periods. The PTSD and psychological distress scales captured the participants’ experiences over the preceding month and four-week period, respectively. The insomnia scale focused on the previous two weeks, along with other sleep-related variables collected at the time of survey completion and the questions related to suicidality covered the previous 12-month period. The varying timelines of measures (insomnia, PTSD, psychological distress) compared to suicidality make it challenging to determine whether these risk factors are transient experiences, or if a longer-term association exists.

Finally, further exploration of other constructs known to be associated with suicidal ideation such as workplace bullying, substance abuse and stigma (Kyron et al., 2021) would be beneficial to generate greater insight to suicide risk factors and inform more robust intervention and policy directions.

This study highlights the relationship between insomnia, PTSD symptom severity, psychological distress and suicidality among firefighters. While insomnia showed an association with suicidality, this association did not remain significant when accounting for PTSD symptom severity and psychological distress. Psychological distress emerged as a significant risk factor for suicidality, suggesting that addressing overall psychological distress levels may be critical in suicide prevention strategies. The identification of distinct profiles based on these risk factors can inform targeted interventions tailored to the specific needs of firefighters at higher risk for suicidality.

We would like to thank all of the firefighters who participated in this project.