Background

Public safety personnel (PSP) serve to maintain public safety and well-being. Occupations included within the definition of PSP include, but are not limited to, border services officers, public safety communications officials (e.g., dispatch or 911 operators), correctional workers, firefighters (career and volunteer), paramedics, and police [1]. PSP and frontline healthcare personnel (FHP, e.g., nurses, physicians, social workers, counselors, and staff in emergency, trauma, surgical, psychiatric, geriatric, and/or intensive care units) are frequently and repeatedly exposed to potentially psychologically traumatic events (PPTEs) [1,2,3]. Consequently, PSP and FHP appear to be at increased risk for posttraumatic stress injuries (PTSIs) [4,5,6,7], which appears to be further exacerbated during the global COVID-19 pandemic [8,9,10,11].

PTSI typically include symptoms of major depressive disorder, panic disorder, generalized anxiety disorder, posttraumatic stress disorder, suicidal ideation and attempts, and substance abuse [1, 4, 12]. In a recent pan-Canadian survey of PSP, 44.5% of respondents screened positive for at least one occupationally mediated PTSI [4, 13]. Furthermore, PSP appear up to four times more likely than the general population to report suicidal behaviors (i.e., ideation, planning, attempts, deaths) [14, 15]. Recent evidence for PTSI prevalence among FHP is lacking; however, the Canadian Federation of Nurses Unions [16] reported that 61% of nurses had experienced abuse, harassment, or assault in the workplace. FHP also report high levels of occupationally mediated compassion fatigue and burnout [17]. The concept of burnout was first proposed in the early 1970s by psychologist Christina Maslach, who explored a phenomenon among care providers involving emotional exhaustion, depersonalization, and diminished personal achievement [18]. Over time, this tripartite construct has become known as burnout [19]. Burnout is currently not a diagnosable mental health disorder but has been formally included as a problematic syndrome in ICD-11 [20].

Evaluations of PPTEs among PSP and FHP have focused mainly on first responders and frontline workers; however, recent evidence indicates that the civilians who work alongside them (e.g., administrative staff, public service employees, victim services) are also frequently exposed to PPTE and report comparable levels of PTSI and suicidal behaviors [21]. The COVID-19 pandemic has also highlighted several less conventional “essential” occupational sectors at increased risk of occupational PPTEs, including public-facing personnel such as transportation workers, grocery clerks, and restaurant workers [22]. Workers in extractive sectors including miners and drillers also regularly face life-threatening operational conditions, increasing the risk for occupationally mediated PPTE [23]. While any study of the effectiveness of a proactive psychological program delivered in an occupational context will qualify for inclusion in the current work, we will focus on PSP and FHP (broadly defined) as the extant literature supports that these occupational groups are most frequently exposed to work-related PPTE.

PTSI symptoms may also negatively impact occupational performance quality, increase absenteeism, increase sleep difficulties, negatively impact interpersonal relationships, increase burnout, and increase early mortality [4, 12, 24, 25]. The economic burden of PTSI among Canadian PSP and FHP is unknown [26]; nevertheless, annual productivity losses from mental disorders experienced by Canadians are estimated to cost between $16.6 [27] and $21 billion [28]. In the USA, health care costs for treating a firefighter, paramedic, or police officer with PTSD are almost five times higher than one without PTSD (~ $10,000/year versus ~ $2000) [26]. The significant costs have prompted several stakeholder organizations and occupational health policymakers to seek proactive approaches, such as implementing psychological and mental health training programs to mitigate the impact of PPTE on workers [29]. Accordingly, psychological interventions that promote well-being have been shown to reduce absenteeism [30, 31]. Proactive measures to support mental health may be particularly relevant for PSP given evidence that stigma is substantially inhibiting care-seeking for mental health challenges [13].

Proactive psychological programs have occasionally been integrated into basic training as part of efforts to increase individual resilience before PPTE exposures, as demonstrated among paramedic [32, 33] and nursing students [34], as well as federal and special forces police in Canada and abroad [35,36,37]. Proactive psychological programs have been increasingly offered to experienced workers who have already been exposed to PPTE but are intended to “prevent” or mitigate the development of PTSI rather than treat them. The current systematic review and meta-analysis focuses on a broad variety of proactive psychological program types in order to investigate the degree to which various occupationally mediated PTSIs are impacted by different programming approaches. We are reticent to label “prevention” programs because of rampant misuses of the term in the existing literature and mental health programming. A program can only be deemed preventative with highly rigorous pre- and post-training PTSI clinical screenings among persons who do not already have a PTSI or mental disorder, which would confound the results. The extant literature indicates that any post-training gains (i.e., effect sizes) are small and very time-limited [13, 38]; also, the gains are expected to deteriorate like other learned skills [33, 39, 40], meaning refresher programs are likely critical for maintaining gains. Despite important efforts at summarizing the existing pre- and post-exposure programming options for first responders frequently exposed to PPTE [41,42,43], there are currently significant research gaps regarding the effectiveness of proactive programs designed to mitigate PTSI, especially among FHP.

Objectives

The current study was designed as a systematic literature review to identify published research on proactive PTSI mitigation programs tailored for PSP, FHP, and other workers exposed to PPTE. The effectiveness of such programs for improving outcomes related to PTSI and psychological health will be evaluated with a quantitative meta-analysis. Comparators will include controls in the waitlist, nil treatment, training as usual, or alternative programming groups, and baseline scores for within-subject studies. Results are presented to summarize the various training approaches, durations, and outcomes evaluated in empirical studies of program effectiveness. The current results can assist industrial, organizational, and occupational stakeholders in implementing evidence-based programming for mitigating PTSI among at-risk workers.

Methods

Protocol and registration

The current study was pre-registered with PROSPERO (CRD42019133534) [44]. Systematic literature review procedures followed PRISMA guidelines [45], as illustrated in Fig. 1 and in the PRISMA checklist in Additional File 1.

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Eligibility criteria

The current review was restricted to peer-reviewed English- or French-language studies assessing the impact of any mental health program designed to mitigate the impact of PPTE among adult (aged 18 and older) workers and published since January 1, 2008. To maximize yield, we extended eligibility to any PPTE-exposed group of workers, including counselors, correctional workers, dispatchers, emergency workers, firefighters, nurses, paramedics, police, rail transit operators, and social workers. Eligible study designs included randomized control trials (RCT) and quasi-experimental studies (e.g., pre-post studies). Studies involving participants with one or more identifiable mental disorders (e.g., clinician diagnosis or a positive screen on a validated psychological instrument), non-PPTE occupational stressors (e.g., work-related demands, organizational stress), or non-experimental designs (e.g., protocols, theses, qualitative studies) were excluded.

Information sources

A population-intervention-comparison-outcome (PICO) framework was used to define study variables of interest and keywords entered into our systematic literature searches, which are provided in Table 1. We searched EMBASE, MEDLINE, PsycINFO, PubMed, and Web of Science between 2008 and December 9, 2019. The database-facilitated searches were supplemented for additional studies with hand-searches of the reference lists from included studies, as well as previous review articles and reports. Following the searches, all citations were imported into Covidence—a web-based systematic review manager [46]. There were two independent reviewers who screened articles against the eligibility criteria: first by title/abstract and then in full. Initial screening was verified by having multiple reviewers screen 200 papers resulting in 99% agreement. All discrepancies were resolved by consensus between the two reviewers.

Table 1 PICO literature search strategy
Full size table

Data extraction

There were two reviewers who extracted data independently from published full-text reports of eligible articles. Per the PICO framework (Table 1), population variables included sample size, age, sex, and years of employment. Intervention variables included the duration of the training program, as well as program themes and approaches reported by study authors. Comparison variables included the type and nature of the comparator group. Outcome variables included absenteeism, scores on validated psychological instruments (i.e., General Health Questionnaire [GHQ], Symptoms Checklist 90 [SCL-90], Depression Anxiety Stress Scale-21 [DASS-21], and physiological markers of stress (e.g., heart rate, blood pressure, salivary and plasma cortisol). Absenteeism did not include individuals already on medical leave at the time of the study. Absenteeism was measured for individuals participating in the PTSI mitigation program (versus annual reports of overall sickness absence). Missing data or outcomes reported in incompatible form for the meta-analysis (e.g., ranks, medians, regression results) were requested from corresponding authors. To maximize power for the meta-analysis, program types and outcome variables were categorized, and operational definitions are provided in Tables 2 and 3, respectively.

Table 2 Proactive PTSI mitigation program categories and specific interventions included in the meta-analysis
Full size table
Table 3 Outcome categories and specific measures included in the meta-analysis
Full size table

Quality assessment

Study quality was appraised using the Newcastle-Ottawa Scale [47], which evaluates nine items across three domains: outcome, selection, and comparability. Each item received a rating of high, low, or unclear risk of bias; each instance of a low risk of bias counted as one point, for a total possible score of nine. Overall study quality was operationalized using the total score: scores of 9 as “high quality,” scores of 7 or 8 as “moderate to high quality,” scores of 5 or 6 as “moderate to low quality,” and scores of 4 and below as “low quality.”

Synthesis of results

Eligible studies for the quantitative meta-analyses needed to report means and standard error or standard deviation values for study outcomes of interest (see Table 3). A random-effects model was applied to pool effect sizes across studies using standardized mean differences (SMD) and their corresponding 95% confidence intervals (CI). Cohen’s criteria [48, 49] were used to interpret an SMD of 0.2 as “small,” 0.5 as “medium,” and 0.8 or greater at “large.” SMDs were measured at all available post-training and follow-up timepoints.

Assessment of heterogeneity and additional analyses

Heterogeneity was quantified using the I2 statistic [50] and forest plots to graphically display summary effect sizes across studies [51]. Outcomes with at least ten studies were explored for sources using the following pre-specified subgroup analyses: occupation (e.g., firefighters, police officers), intervention (e.g., mindfulness-based, multimodal), and timeline (e.g., post-training, 1-month follow-up, 18-month follow-up). Sensitivity analyses included comparisons of random-effects and fixed-effects model effect sizes, as well as with the leave-out-one technique. For outcomes with at least ten studies, publication bias was assessed using funnel plots, the trim and fill method [52], and Egger’s test of funnel plot asymmetry [53, 54].

Results

Systematic literature review

The systematic review identified a total of 4154 studies. Among the identified studies, there were 224 removed as duplicates, leaving 3930 studies for the title and abstract screening. There were 3815 records removed, leaving 115 studies for full-text review. There were 73 studies excluded at the full-text stage: 25 had a wrong population (i.e., not PSP, FHP, or a PPTE-exposed occupational group), 23 had a wrong study design (i.e., not a pre-post evaluation of outcomes such as qualitative studies or protocols, or non-peer-reviewed dissertations, books, or reports), 20 had a wrong intervention (e.g., post-PPTE service, treatment, or therapeutic intervention), 4 had wrong outcomes (e.g., program acceptability or outcomes unrelated to mental health or wellness), and 1 was a duplicate title. The systematic review process resulted in 42 eligible studies that evaluated the effectiveness of a proactive PTSI mitigation program in workers exposed to PPTEs. Key study characteristics are described below and are summarized in Table 4, including participant summaries, study designs, PTSI mitigation program themes, primary outcomes, and results. A subsequent six studies were excluded from the meta-analysis for failing to report the means and/or standard deviations for their primary outcome measures [72, 86, 89], or reporting the means and/or standard deviations in formats that were incompatible for a quantitative meta-analysis; for example, reporting regression results [71], medians [81], or ranks [77]. All authors were contacted with data requests, but data were not yet provided at the time of submitting the current review. Ultimately, 36 studies were included in a quantitative meta-analysis (Fig. 1).

Table 4 Summary characteristics of eligible studies (n = 42). Studies not included in meta-analyses (n = 6) are marked with an asterisk (*)
Full size table

Study characteristics

The 42 studies represented data from 3182 individuals. Police officers were the most common PSP group (n = 15), followed by correctional workers (n = 2), firefighters (n = 2), and paramedical students (n = 1). There were no eligible studies, including participants from other PSP sectors. FHP occupations represented included nurses (n = 11) and various groupings of FHP (n = 9), including social workers, radiologists, medical examiners, physicians, nursing home employees, Red Crescent/Red Cross personnel, physiotherapists, occupational therapists, and healthcare clerical and administrative staff. The systematic literature search also yielded two relevant studies on educational staff exposed to a PPTE [60] and on miners whose occupational roles pose a realistic and substantial life threat (e.g., workers, blasters, foremen at the mine face) [82].

PTSI mitigation program themes identified in the eligible studies overlapped heavily and are not mutually exclusive (i.e., one program may fall under multiple themes). The aggregation of programs into broader categories was conducted to perform the meta-analysis and are defined in Table 2. Program themes included mindfulness (n = 13), psychoeducation (n = 20), psychophysiology (n = 11), resilience promotion (n = 17), stress management (n = 15), building coping skills (n = 7), emotion regulation (n = 4), cognitive (behavior) therapy (n = 3), and psychosocial support or counseling (n = 2). There were 8 studies that evaluated the effectiveness of self-described “multimodal programs” that included multiple themes identified above. There were 6 studies that employed biofeedback with primary resilience promoting program types.

Study designs included RCTs (including cluster, parallel, and crossover RCTs) (n = 26) or prospective cohort studies (n = 15). Comparators included waitlist controls (i.e., offered the program at the end of the study) (n = 9), no training or occupational skills training (i.e., not mental health training) as usual (n = 11), psychoeducation only (n = 2), neutral or negatively valenced versions of the program (n = 2), or alternative control programs such as general wellness (n = 1) or Critical Incident Stress Management (n = 1). There were 7 studies that used online or web-based presentation of their programs, while 34 studies used in-person group sessions. Program durations ranged from a single 90 minute group session [88] or one-day workshop or equivalent (i.e., less than 8 hours) (n = 6) to 4- or 5-day workshops or retreats [35, 36, 40, 87]. Multiple training sessions were distributed over a minimum of two days [79] and a maximum of 9 months [90], which included 15 weeks of mental health programming within nine months of recommended physical exercise. Self-paced programs (n = 8) were predominantly web-based, and studies reported very low levels of program completion and/or adherence [62, 64, 74]. Several programs, predominantly mindfulness-based, also included optional or recommended daily practice, or “homework” (n = 19). Study duration for follow-up evaluations ranged from immediately following the training program (n = 25), 1 week to 3 months post-training (n = 16), 6 months (n = 6), 7 to 12 months (n = 8), and 13 to 18 months (n = 3). A single study with multiple follow-up durations would be included in more than one of the categories (e.g., Andersen et al. [40] conducted pre-, post-training, 6-, 12-, and 18-month evaluations).

Due to the wide variety and limited consistency in PTSI mitigation program types, outcome variables, occupational groups, and follow-up durations across studies, the meta-analysis results will be presented by outcome categories (defined in Table 3) and are summarized in Table 5. Below, we report on the effectiveness of PTSI mitigation programs on reducing symptoms of PTSI and improving general measures of psychological health and wellness. Effect sizes (SMD) and confidence intervals (CIs) will be reported for specific outcomes. Any statistically significant differences in outcomes by program type, follow-up duration, and/or occupational group will then be presented where subgroup analyses were performed. Supporting figures can be found in Additional File 2.

Table 5 Summary of meta-analytic results, subgroup analyses, and publication biases
Full size table

Mental disorder symptoms and absenteeism

Depression

Significant reductions in mental disorder symptoms were observed for depression, with an SMD of − 0.46 [ − 0.71; − 0.21]. Depression effect sizes were largest for resilience promotion programs (SMD = − 1.05; p < 0.01) and immediately post-training (− 0.78) compared with follow-up (p = 0.05) (Figure 2.1 in Additional File 2).

Burnout

Moderate reductions in symptom burden were also observed for burnout (SMD = − 0.45 [− 0.64; − 0.26]), with larger effect sizes observed in resilience promotion versus multimodal programs (− 0.90 vs. − 0.24; p = 0.04) (Figure 2.2 in Additional File 2).

PTSD

Medium reductions were observed for PTSD symptoms with an SMD of − 0.33 [− 0.55; − 0.11]. Significant subgroup analyses evidenced the largest effect sizes at 8-month follow-up (− 1.22; p < 0.01), and among PPTE-exposed educators (− 0.86; p = 0.03) (Figure 2.3 in Additional File 2).

Anxiety

Small reductions were observed across all studies for anxiety symptoms, with an SMD of − 0.20 [− 0.31; − 0.10].

Suicidality

Suicidality was not significantly reduced with the programs considered (SMD = 0.33 [− 0.07; 0.73]).

Substance use

For alcohol, the overall reduction in weekly alcoholic drinks was small and not statistically significant, with an SMD of − 0.08 [− 0.21; 0.06]. The effect was even smaller for other substance use, with an SMD of − 0.05 [− 0.35; 0.24]. Given the small study yield for either outcome (Table 5), subgroup analyses were not conducted.

Absenteeism

There was no significant improvement in absenteeism—defined as the number of sick days taken by employees in the study—with an SMD of 0.01 [− 0.19; 0.21].

General measures of general psychological health, stress, resilience, and well-being

General psychological health

There was a significant medium reduction in general psychological symptom burden across studies, with an SMD of − 0.70 [− 1.14; − 0.26] (Figure 2.4 in Additional File 2). There were significant differences in effect sizes across timepoints (p < 0.01) – with small positive effects (i.e., increases in general psychological symptoms relative to pre-training) noted at 18-month follow-up (0.34) relative to large decreases in symptoms (reflected by negative effect sizes) at 1-month (− 0.95) and immediate post-training (− 0.91) timepoints—and by program type (p < 0.01), with the larger reductions associated with multimodal programs (− 1.09) relative to resilience promotion programs (0.34).

Stress

Stress symptoms were associated with small-to-medium reductions in symptom burden overall, with an SMD of − 0.35 [− 0.51; − 0.20] and larger effects among hospital staff (− 0.84) compared with police officers (− 0.29; p = 0.02).

Well-being

Among measures of broader mental health status, effect sizes were largest for well-being with a medium SMD of 0.46 [0.26; 0.66]; larger effects were observed among educators (1.95) compared with other occupational categories (p < 0.01).

Coping

There was evidence of a medium SMD of 0.41 [0.02; 0.80], with larger effect sizes at 18-months (0.93) versus other timepoints (p < 0.01), with resilience promotion programs (0.93) relative to other program themes (p < 0.01), and among police officers (0.73) and radiation therapists and nurses (0.70) relative to paramedical staff (− 0.01; p < 0.01) (Figure 2.5 in Additional File 2).

Resilience: Overall improvement in resilience was small, with an SMD of 0.27 [0.13; 0.42] (Figure 2.6 in Additional File 2). Effect sizes were largest at immediate post-test (− 0.46; p = 0.02) and larger with resilience-promoting strategies (0.98) relative to other modalities (p < 0.01).

Biological measures of stress

There was no evidence that the investigated programs reduced serum biomarkers of stress, such as cortisol, antithrombin, dehydroepiandrosterone (DHEA), and prolactin (Table 5). While there were no significant reductions in blood pressure across studies, there were significant improvements in overall and average measures of heart rate, with a small SMD of − 0.27 [− 0.40; − 0.14]. The greatest reductions in heart rate were seen at 12-months of follow-up (− 1.52, p = 0.04), with resilience promotion programs (− 1.00; p < 0.01), and among nurses (− 0.55) and police officers (− 0.45) compared with correctional officers (− 0.15; p = 0.04).

Publication bias

There was evidence of a publication bias for anxiety (p = 0.0061) and depression (p = 0.03) (Figure 2.7 in Additional File 2). The trim-and-fill method was used to account for potential outcome effect size estimate inflation (Table 5). For anxiety, the effect size changed from an SMD of − 0.20 to − 0.10 [− 0.23; 0.02]. For depression, the effect size changed from an SMD of − 0.46 to − 0.17 [− 0.47; 0.12]. For each outcome, correction for publication bias significantly reduced the effect size, meaning that anxiety and depression outcomes are likely associated with publication bias.

Quality assessment

Quality assessment ratings for all studies in the current systematic review are illustrated in Fig. 2. Individual study ratings are reported in Table 4, and detailed study ratings are presented in Table 3.1 in Additional File 3. Overall, only one study was of “high quality” [63], 11 were “moderate to high quality,” 24 were of “moderate to low quality,” and six were of “low quality.”

Fig. 2
figure 2

Quality assessment using Newcastle-Ottawa Scale. Full sample (n = 36 studies) summary of strength of evidence from systematic review and meta-analysis

Full size image

Outcome

All studies were rated at low risk of bias because all used empirically validated self-report mental disorder screening tools or objective physiological data. There were 14 studies that were rated at high risk of bias for follow-up periods; outcomes were only assessed immediately after program delivery, prohibiting evaluations of program effectiveness following subsequent occupational exposures to PPTE. Concerning adequacy of follow-up, more than half of all studies (n = 25) received high risk of bias ratings for either failing to report post-training sample sizes and any possible participant attrition and/or failing to provide analyses of retained participants to those lost at follow-up.

Selection

All but seven studies were rated at a high risk of selection bias for failing to demonstrate sample representativeness, limiting results’ generalizability. All studies were rated at a low risk of bias due to clear selection criteria for control groups except for one sample from a different police district [85]. Most studies adequately ascertained participation in the programming being assessed (n = 34); however, three studies were either unclear about program completion or participation [68, 77, 86] and five studies of self-paced online programs reported very low participation or completion [62, 64, 74,75,76].

Comparability

The purpose of the current systematic review was to identify the effectiveness of proactive psychological programs designed to mitigate PTSI and limit the decline of psychological symptoms among workers at high risk of exposure to a PPTE. Therefore, studies that included individuals with diagnosed PTSI or mental health disorders were excluded at the title and abstract screening phase, as these would be considered PTSI treatments or services. Accordingly, the comparability criterion pertains to controlling for the most important factor in the study design, which in the case of the current review is the presence of a pre-existing PTSI (i.e., before program onset) and/or exposure to a PPTE following program onset, both of which would significantly confound investigations of program effectiveness. Only 12 of the included studies received a low risk of bias rating for the comparability criterion, either reporting or controlling for mental disorder symptom severity at baseline or pre-training measures, or by reporting PPTE exposures following program onset and before any follow-up evaluations. All remaining studies (n = 25) were at a high risk of bias.

Conversely, all but five studies received a low-risk rating for demonstrating that the outcome of interest was not present at the start of the study by providing pre-training baseline measures for reported outcomes. Roughly half of the eligible studies (n = 22) controlled an additional factor in the study design or analysis, including age or years of service/employment, which are known correlates of mental health among PSP and FHP [4, 14, 92].

Discussion

The effectiveness of various organizational programs designed to “prevent”—or more accurately to proactively mitigate—PTSI and improve psychological health indicators among PPTE-exposed occupational groups remains unclear. The current systematic review identified 42 empirical research studies measuring the effectiveness of organizational training programs designed to proactively mitigate PTSI among PSP, FHP, and other workers exposed to PPTEs (Table 4). A great deal of heterogeneity was indicated across program themes and durations and study designs, durations, and follow-up periods. Self-directed or web-based programs also suffered from poor participant adherence and completion. The quality assessment indicated a high risk of reporting bias for several study elements (Fig. 2), including failure to demonstrate sample representativeness (83% of studies), evaluate or report on the presence of a PTSI and/or mental disorder before the study and program onset, and/or participant exposure to PPTEs before follow-up evaluations (71%). The identified factors would significantly confound investigations of program effectiveness and limit the generalizability of results. Most studies also reported high attrition rates at follow-up evaluations (60%), and several collected post-training measures before the newly acquired skills could be practiced or applied in work conditions (33%).

Evaluation of 36 study outcomes with a quantitative meta-analysis provide evidence that all programs (i.e., collapsed across program type) resulted in statistically significant reductions in PTSI after training (Table 5), including symptoms of general psychological health, depression, burnout, stress, PTSD, and anxiety, as well as significant improvements in measures of well-being, coping, and resilience (see Tables 2 and 3 for operational definitions of program types and outcomes, respectively). Consistent with previous literature [13, 41, 42], post-training improvements are of a medium (SMD < 0.8) or small (SMD < 0.5) effects in magnitude and time-limited. Subgroup analyses indicated large (SMD > 0.8) effect sizes for sustained improvements in PTSD symptoms and coping for up to 18 months. Resilience promotion programs appeared to reduce symptoms of depression and burnout and improved coping and resilience measures. Multimodal programs that combined therapeutic approaches (e.g., mindfulness, stress management, emotion regulation, resilience promotion) appeared to improve measures of general psychological health. Police appeared to report the greatest improvement in coping measures, likely due to the overrepresentation of this population across studies (n = 15). In contrast, combined groups of FHP indicated the greatest reduction in stress symptoms, and PPTE-exposed educational staff indicated the greatest decrease in PTSI symptoms [61]. Significant and sustained improvements in coping may be promising, but coping is not yet a clinically validated construct for PTSI. Similarly, wellness and resilience are general health indicators that vary in operational and theoretical definitions between individual studies. Substantial barriers to evaluating program effectiveness identified in the current review include inconsistency in outcome measures, outcome reporting (i.e., mean and standard deviation values required for meta-analysis), and follow-up durations across studies. The limitations precluded more detailed subgroup analyses and data synthesis for the current meta-analysis.

Despite claims for reduced absenteeism as a justification for funding and implementation of mental disorder “prevention” programs, evidence provided by the current meta-analysis does not support reduced absenteeism due to inconsistent or insufficient reporting, and especially for distinguishing missed work as a result of a psychological injury sustained at work or due to physical illness. Similarly, the results did not evidence statistically significant reductions in substance use or suicidality, both of which are prevalent among PSP [4, 14, 21] but remain under-investigated among FHP and other at-risk workers. There was also no evidence for post-training improvements in physiological biomarkers of stress except for heart rate (Table 5), which was the third most common outcome measure and included several studies that condition adaptive through heart rate variability biofeedback training (HRV-BF). Despite the limited evidence supporting the effectiveness of HRV-BF as a possible tool for modulating stress physiology, doing so requires substantial resources such as medical-grade cardiovascular equipment, software, repeated training sessions, and qualified personnel psychophysiology to guide psychoeducational modules [35, 36, 40, 79, 80].

The studies identified by the current systematic literature review provide modest evidence for very time-limited improvements in several PTSI symptoms. Variability in program types and durations also represents variability in organizational cost and feasibility, which place limitations on implementing and repeated investigation of PTSI mitigation program effectiveness either separate from or part of a longitudinal research study (see also [93]). Web-based or self-directed programs may be considered more cost-effective to implement than multiple in-person group sessions; however, the current systematic review results demonstrate a critical and substantial lack of adherence, as well as very low completion rates for several online program protocols [62, 64, 74,75,76]. Inconsistent and poor-quality study designs precluded more conclusive recommendations directed at organizational stakeholders to inform PTSI mitigation programming or training tailored to PSP and FHP.

Limitations

Despite a relatively high number of eligible studies (n = 42), the main limitation of the current systematic review is the high heterogeneity across studies, which precluded the inclusion of six studies in a quantitative meta-analysis and more detailed subgroup analyses. The quality of available studies was also highly variable (Fig. 2 and Table 3.1 in Additional File 3), with 71% (30 of 42) of studies scoring moderate to low quality (≤ 6 out of 9). The geographical variability of participants also makes generalizability difficult because an effective program in one political, cultural, social, economic, and epidemiological context may not be relevant, applicable, or effective elsewhere. Nevertheless, the substantial impact of PPTE exposures on the mental health of PSP and FHP appears broadly accepted, as does the need to develop effective evidence-based PTSI mitigation programming for all at-risk workers to minimize personal, social, and economic costs [26, 29].

Publication bias was high for studies with outcome measures assessing anxiety and depression, and adjustments for methods biases rendered the results no longer statistically significant (Table 5). As smaller trials are generally analyzed with less methodological rigor than larger ones, the resulting asymmetrical funnel plot suggested that selective reporting may have led to an overestimation of effect sizes in smaller trials (Figure 2.7 in Additional File 2). Other limitations included the search strategy and criteria process (Table 1), which was restricted to English- and French-language studies published after 2008 from five indexed electronic databases.

Conclusions

Especially during the current global coronavirus pandemic, there is an urgent need to identify effective organizational training tailored for PSP and FHP and designed to mitigate the psychological impact of PTSI that can result from occupational PPTE exposures [26, 29]. The extant literature identified by the current systematic literature review indicates broad variety in sampled occupational populations, implemented programming approaches, and measured outcome variables. Heterogeneity across studies precludes identifying a proactive PTSI mitigation program type that is superior to others and effective for diverse PSP, FHP, and other at-risk workers exposed to PPTE. Nonetheless, we have synthesized the available evidence on proactive programming effectiveness in reducing specific PTSI symptoms. Based on our meta-analytic results, resilience promotion and multimodal programs that combine a variety of therapeutic and skill-building approaches appear to produce modest time-limited reductions in symptoms of general psychological health, depression, burnout, stress, PTSD, and anxiety, as well as promoting well-being, adaptive coping, and resilience. By identifying significant research gaps and practical study limitations, we intend to help inform future high-quality research evaluating program effectiveness within the context of PSP and FHP working environments. The current results provide organizational stakeholders and policymakers with numerous options for developing innovative mental health solutions tailored to the unique occupational challenges faced by those who serve to maintain and protect public health and safety.