This systematic review was preregistered in the Open Science Framework (OSF) after the article screening stage and before the quality appraisal, data extraction, and synthesis phases. The registration aims to enhance transparency in the analytical and interpretive phases. DOI: https://doi.org/10.17605/OSF.IO/QESZ2

This study followed a systematic review design in accordance with the PRISMA 2020 guidelines. The review aimed to synthesize empirical evidence on burnout in firefighting professionals by identifying individual and organizational factors associated with the syndrome.

Following PRISMA methodology, an advanced search was conducted in PsycInfo, Web of Science (WOS), and SciELO to retrieve all relevant studies available in these databases on burnout among firefighters on 30 January 2025.

Regarding the search conducted in Web of Science (WOS), it was divided into two sub-searches: one focused on psychosocial factors (which yielded 253 papers) and another on organizational factors (which yielded 27 papers). The same search structure was applied in PsycInfo: one sub-search targeting psychosocial factors (resulting in 58 articles) and another focusing on organizational factors (resulting in 10 papers). These same searches were replicated in Spanish in both PsycInfo and WOS (see Table 1), but no results were obtained. The procedure was also repeated in SciELO: one search focused on psychosocial factors (which returned one result), and another on organizational factors (which returned no results).

As a result of the various searches, a total of 349 papers were initially identified for screening. All records from the aforementioned searches were uploaded into Rayyan, where duplicate articles were removed, reducing the initial sample to 262 papers for further screening.

The screening process was carried out independently by two researchers. Each researcher assessed whether the papers met the inclusion and exclusion criteria (see Table 1). As a result, 89 papers were initially found to meet the inclusion criteria. A second screening was then conducted to exclude articles that did not address burnout as the main topic or whose full text was not available online, yielding a final set of 35 articles selected for systematic review (see Figure 1).

The screening process was carried out in January 2025, using the Rayyan platform. Two researchers independently assessed all abstracts against the predefined inclusion and exclusion criteria. Disagreements were discussed in a joint session until consensus was reached, resulting in a final agreement rate of 100%.

For the systematic review, each study was reviewed to extract information related to burnout, including the instrument used for its assessment (e.g., Maslach Burnout Inventory), the specific dimensions analysed (emotional exhaustion, depersonalization, personal accomplishment), and the reported relationships between burnout and other psychological or contextual variables that the authors explicitly incorporated into their proposed model analysed in the paper (e.g., mediations involving sleep quality or interpersonal relationships within the team).

Additionally, relevant contextual variables were collected to better interpret the findings. These included sociodemographic characteristics such as the country where data were collected, the time period (year of data collection), sample size, and representation of women in the sample (including the number of female participants and, when absent, whether such exclusion was justified).

All available information related to burnout was recorded, including any mediating or moderating effects identified (particularly in cases where burnout functioned as a mediating or moderating variable in the relationship between other variables).

Finally, when specific data on any of the previously mentioned aspects were not available, no assumptions were made; missing information was left uncompleted in order to avoid introducing bias into the study.

Additionally, a quantitative synthesis of the total effects was conducted to estimate the magnitude of the direct associations between job demands, job resources, and burnout. The remaining studies were excluded from the quantitative synthesis because burnout was not assessed using the Maslach Burnout Inventory (MBI) or because they did not report zero-order associations between burnout and job demands or resources, as required by the JD-R framework.

For this purpose, reported Pearson’s correlation coefficients (r) and odds ratios were transformed into Fisher’s Z values and subsequently reconverted into r to obtain a standardised and comparable measure of effect size magnitude. Although 35 studies met the inclusion criteria and were included in the systematic review, only 15 provided sufficient statistical information to compute these associations (see Appendix 3-9 for specific details).

The quality of the included studies was evaluated using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist: the 8-item version for cross-sectional studies and the 11-item version for cohort studies. This tool has recently been updated to improve its applicability and to align its standards with current approaches to risk of bias assessment. These updates allow for the evaluation not only of the internal validity of studies, but also for differentiated judgments at the level of the study, outcomes, and individual results, enabling a more detailed and precise assessment of methodological quality (Barker et al., 2023).

The use of this tool is especially appropriate in observational studies without control groups, where other instruments such as RoB 2.0 or ROBINS-I present limitations, as they require experimental or comparative designs. Moreover, the checklist format provides a clear guide for evaluating all relevant aspects of methodological quality, from the sample selection process to the adequacy of the statistical analyses performed.

Although the quality appraisal was primarily conducted by a single reviewer, any uncertainties or ambiguities were systematically discussed with a second reviewer until consensus was achieved. While dual independent appraisal is ideal, the collaborative consensus-based approach ensured the methodological soundness and consistency of the quality assessments. The full results of the quality assessment can be found in the Appendix 1 of the Supplementary Material.

Finally, most cross-sectional studies scored 7-8 out of 8, and all cohort studies at least 9 out of 11, indicating generally high methodological quality. The main limitation was that many studies did not control for potential confounders (e.g., age), which should be considered when interpreting the evidence. However, the high quality of the included studies supports confidence in the findings (see Appendix 1 for detailed ratings).

Although a meta-analytic approach is generally recommended, it was not conducted in this review due to a substantial heterogeneity in study designs, measures, operationalization of key variables (e.g., job demands, coping, emotional intelligence), and population characteristics (e.g., professional vs volunteer firefighters, gender representation, organizational context). Instead, a narrative synthesis combined with a quantitative estimation of the total effects were performed to capture the complex, context-dependent nature of burnout in firefighting.

Extracted data were organized in predefined tables (author, year, country, sample, instrument used to assess burnout and findings) and the study selection process was documented through a PRISMA flow diagram (Figure 1; see Appendix 2).

Additionally, to provide a general estimate of the magnitude of associations, total effects were calculated for the subset of studies employing the MBI. Zero-order correlations and odds ratios were extracted and transformed into Fisher’s Z, and subsequently back-transformed into Pearson’s r. Cohen (1988) indicated the following intervals for interpret r as r = .10 to .30 small effect, r = .30 to .50 medium effect, and r = .50 and higher as large effect. This procedure yielded overall effect size estimates separately for job demands and for job resources, according to the Job Demands-Resources framework.

Full study-level data extraction is available in Appendix 1 (methodological quality, item-level JBI ratings) and Appendix 2 (structured extraction of sample information, effect estimates, mediators, moderators, and instrument details). Together, these constitute the complete data-extraction matrix for transparency and reproducibility.

Multiple studies (Dyal et al., 2022; Katsavouni et al., 2016; Khoshakhlagh, Al Sulaie, Mirzahosseininejad, et al., 2024; Khoshakhlagh, Al Sulaie, Yazdanirad, et al., 2024; W. Kim et al., 2019; Mitani et al., 2006; Smith et al., 2018, 2020) have highlighted the role of burnout as a mediating variable in the relationship between stressors (e.g., trauma, occupational stress, or work-family conflict) and various negative outcomes, such as post-traumatic stress symptoms (PTSD), musculoskeletal problems, sleep disturbances, unsafe workplace behaviours, and a decrease in workers’ safety-related behaviours. This mediating role positions burnout as a central node within a network of health- and performance-related consequences.

However, some recent studies have proposed rethinking the assumed directionality in certain explanatory models. In this regard, Liu et al. (2025) identified poor sleep quality as the most central node in their network analysis, suggesting that poor sleep quality could act as an antecedent factor to burnout. In contrast, several studies (Dyal et al., 2022; Khoshakhlagh, Al Sulaie, Cousins, et al., 2024) found that burnout mediated the relationship between occupational stress and sleep quality, rather than being a consequence of poor sleep, thus supporting the classical perspective in which burnout leads to impaired night-time rest. Khoshakhlagh Al Sulaie, Cousins, et al. (2024) also found that burnout functioned as a moderating variable between PTSD and sleep quality.

Some studies also found that specific dimensions of burnout acted as mediators in different relationships (Tao et al., 2022). Specifically, emotional exhaustion and depersonalization mediated the relationship between neuroticism and anxiety, while depersonalization mediated the relationship between neuroticism and depression.

Finally, Khoshakhlagh, Al Sulaie, Mirzahosseininejad, et al. (2024) and Khoshakhlagh, Al Sulaie, Yazdanirad, et al. (2024) found that burnout (specifically the dimension of emotional exhaustion) acted as a mediating factor in the relationship between PTSD and musculoskeletal problems, with emotional exhaustion being the most influential component. In fact, the impact of other variables (e.g., job stress, depersonalization, personal accomplishment, depression) on the PTSD and musculoskeletal problems relationship was also examined, but burnout emerged as the most influential mediating factor.

A total effects analysis was conducted to synthesise the strength of the direct associations observed and evaluated in the included studies, thereby providing quantitative information on the impact of demands and resources on burnout (see Appendix 3-9 for more details).

When data were available, findings were disaggregated by burnout dimension to provide an average estimate of the extent to which each component (emotional exhaustion, depersonalisation, or personal accomplishment) was affected by job demands and resources. However, some studies only reported an overall measure of burnout. In those cases, the reported overall burnout values were also included to compute the global impact of demands and resources (separately) on total burnout levels across studies.

Specifically, zero-order relationships between demands/resources and burnout were examined. Reported Pearson’s correlation coefficients (r) and odds ratios were transformed into Fisher’s Z values and subsequently reconverted into r to obtain a standardised and comparable measure of effect size magnitude. Therefore, this section summarises the direct effects reported across the studies, excluding moderating and mediating mechanisms.

On average, job demands showed a medium positive association with burnout (r = .36, SD = .10, N = 21), indicating that higher exposure to stressors is consistently correlated to higher burnout levels. Oppositely, job resources showed a medium negative association (r = -.33, SD = .13, N = 11), which reflects their protective role in mitigating burnout levels in these professionals.

Specifically, job demands were medium positive associated with emotional exhaustion (r = .31, SD = .12, N = 23), while showing small positive association with depersonalization (r = .25, SD = .13, N = 20), and low personal accomplishment (r = .10, SD = .19, N = 9).

In parallel, job resources were found to have a buffering effect over burnout. Specifically, job resources showed a small negative association with emotional exhaustion (r = -.23, SD = .16, N = 6), followed by depersonalization (r = -.20, SD = .21, N = 4) while showing a marginal negative relationship with personal accomplishment (r = -.01, SD = .41, N = 2).

Overall, these findings are consistent with the Job Demands-Resources framework, whereby job demands uniquely and strongly relate to burnout over time, while job resources primarily buffer demands and relate more strongly to engagement (Bakker et al., 2023).

Overall, the results are aligned with the Job Demands-Resources framework, confirming that burnout among firefighters primarily results from the imbalance between intense job demands and insufficient personal and organizational resources. This pattern was supported by the total effects analysis, which revealed a moderate positive association between job demands and burnout and a moderate negative association between job resources and burnout. Resources showed their strongest protective effect on emotional exhaustion and depersonalisation, while their influence on personal accomplishment was weaker.

However, the non-significant relationship observed between resources and the personal accomplishment dimension should be interpreted with caution since this correlation likely reflects a methodological inconsistency rather than a true absence of association. Specifically, while De Carvalho et al. (2021) conceptualised this dimension as low personal accomplishment, Mitani et al. (2006) assessed personal accomplishment directly. Although both variables capture the same construct, they are expressed in opposite directions; therefore, their effects tend to cancel each other out when averaged. After adjusting for this difference, both studies indicate the same underlying pattern, namely that higher resources are related to greater personal accomplishment (or lower reduced accomplishment). Nevertheless, the direction of the reported effects was not modified to preserve the integrity and comparability of the original data.

Regarding the characteristics of the studies, a high methodological and geographical heterogeneity was found, with the greatest representation of studies conducted in East Asia. In contrast, no study was found to have been carried out in Spain, which clearly highlights a gap in the literature.

As for female representation in the samples, it was extremely low because of the clear masculinization of this professional sector. However, the exclusion of women from these samples appears to be more of a methodological issue (lack of sufficient statistical power to conduct gender-sensitive analyses) than an ideological one or related to gender inequality.

In turn, although there was broad methodological heterogeneity (use of LBQ, OLBI, SMBM), a clear trend was observed in the use of the MBI as the main assessment tool.

Organizational factors (e.g., workload, role ambiguity, type of job position) have emerged as the main predictors of burnout. Likewise, although various individual and personality variables have proven to play a key role in the development of burnout (e.g., high neuroticism, perceived stress, or reactive coping strategies), others have also been identified as potential protective factors against burnout, by mitigating its adverse effects (e.g., proactive coping styles, self-efficacy, emotional intelligence, religiosity, etc.).

Nevertheless, certain elements that the scientific literature usually considers protective (e.g., vocational calling and self-compassion) have proven to act as risk factors under certain conditions in this professional group.

Finally, in this multidimensional and multifactorial conceptualization of burnout, it has been found that the syndrome produces multiple adverse consequences for professionals in this field (e.g., PTSD, depression, anxiety, musculoskeletal problems, lower safety behaviours, greater absenteeism), which entail significant costs both for the individual and for the organization.

Although many studies have identified the expected patterns between burnout and the variables examined (e.g., workload, coping strategies), divergent results were observed in some cases, which appear to be closely linked to the specific characteristics of the profession. While previous literature predominantly considered social support a protective factor against burnout (Fan et al., 2024; Ruisoto et al., 2021), this review found mixed results. In certain high-tension scenarios, social support may paradoxically amplify stress responses and contribute to burnout, possibly due to overexposure to emotionally charged narratives or unhelpful peer reinforcement mechanisms.

Similarly, results diverging from expectations were also found with regard to professional vocation, which has traditionally been conceptualized as a protective factor (McManus et al., 2011; Nahas et al., 2024). However, the findings of this review indicate that a strong identification with the professional role may increase vulnerability to burnout and post-traumatic stress disorder under conditions of high emotional load. In the case of firefighters, the internalization of roles centred on helping and sacrifice may foster overinvolvement, difficulties in establishing psychological boundaries, and sustained exposure to emotional distress, particularly in contexts of repeated trauma.

Taken together, these findings suggest that factors commonly regarded as protective may operate differently in firefighters, highlighting the importance of interpreting burnout processes within the emotional, relational, and vocational context of this profession. These results prove the need of further research to clarify the conditions under which these factors function as protective or risk-enhancing mechanisms.

First, this systematic review contributes to understanding burnout as a multidimensional and multilevel construct. It is not only a consequence of occupational stressors, but also a central, mediating, and moderating factor that connects psychological demands with both physical and mental health outcomes. In fact, burnout appears as a central node within occupational health networks, mediating between disorders (e.g., PTSD, stress or depression), and physical health issues (e.g., musculoskeletal symptoms or sleep disturbances).

Second, the findings of this review are consistent with the core assumptions of the Job Demands-Resources (JD-R) model (Demerouti & Bakker, 2025; Demerouti et al., 2021), while also introducing important nuances. However, the results suggest that vocational calling, typically considered a protective factor, may under conditions of high emotional strain become a risk factor, thus challenging traditional interpretations of its role within burnout models.

Furthermore, among the dimensions measured by the Maslach Burnout Inventory (MBI), emotional exhaustion consistently emerged as the most predictive and impactful component. This highlights the need for theoretical frameworks to give this dimension greater weight, both in conceptualization and in empirical modelling.

The development of this review gains particular relevance when contextualized within the current institutional landscape many countries (such as Spain), where no structured and sustained policies exist for the prevention of burnout among firefighters. As Cevallos et al. (2024) point out, the absence of concrete preventive strategies within these institutions has contributed to an increase in negative behaviours among personnel, with evident consequences in their professional, family, and social spheres. Therefore, given this lack of systematic institutional policies, the results of the present study provide a theoretical foundation for the development of future interventions aimed at preventing burnout in firefighting bodies.

Specifically, the conceptual framework derived from the results of this project can guide both occupational health professionals and responsible institutions in designing intervention programmes tailored to this workforce. These programmes should target both prevention and health promotion, at both individual and organizational levels.

These recommendations are aligned with international best practices. For example, in Canada, standardized protocols for emotional debriefing (Critical Incident Stress Management, CISM) are applied, along with psychological resilience programmes designed specifically for fire services. In particular, the CISM model has become one of the most effective strategies for managing post-traumatic stress in emergency services (Mitchell, 2006). Transferring such initiatives to the Spanish context is not only viable but particularly necessary, given the current lack of structural measures in the country.

Finally, to ensure the effectiveness of burnout prevention measures, it is essential that institutions go beyond merely complying with legal requirements regarding psychosocial risk (Ley 31/1995, 1995) and move toward a proactive and integrated management of these risks. Although the law mandates regular evaluations, these are often carried out reactively, without the systematic implementation of corrective plans. Therefore, the elaboration and execution of effective action plans is recommended, to identify risks as well as to promote real organizational changes in the work environment, leadership practices, and human resource allocation.

In conclusion, this study not only advances understanding of the factors involved in the onset and persistence of burnout, but also provides a solid empirical and theoretical basis for designing public policies, organizational strategies, and clinical interventions tailored to the operational realities of firefighters, promoting their mental health, quality of work life, and sustainable professional performance.

Based on the findings of this systematic review, several key priorities for future research on burnout in firefighters can be identified.

First, there is a clear need for longitudinal and experimental studies that go beyond correlational designs and allow for the analysis of causal relationships between individual, organizational, and contextual factors. Most of the studies included in this review used cross-sectional methodologies, limiting the ability to determine temporal dynamics or directionality between variables. In contrast, longitudinal approaches would help to better understand how burnout evolves over time and how it interacts with related conditions such as post-traumatic stress disorder (PTSD), insomnia, or work capacity decline.

Second, it is essential to empirically evaluate the effectiveness of psychosocial interventions designed to reduce burnout among firefighters. These could include programmes focused on the development of emotional competencies (e.g., emotional intelligence, regulation), resilience-based interventions, adaptive coping strategies, or organizational changes aimed at improving workplace climate, perceived equity, or access to social support. In this context, the role of transformational leadership also warrants further exploration, given its potential to mediate the impact of occupational stressors on psychological health.

Thirdly, there is a need to expand sample diversity. The underrepresentation of female firefighters in most studies hinders gender-based analyses and the design of tailored interventions. Including women more actively in future research is essential from both an equity and a scientific perspective.

Furthermore, the findings of the total effects analysis justify that future research should place greater emphasis on examining protective resources, rather than predominantly focusing on job demands. Specifically, the total effects analysis revealed that, although resources have been less frequently studies (there are fewer data available when compared to demands), their overall impact on burnout is comparable in magnitude to that of job demands. This finding underscores the importance of adopting a preventive and resource-oriented approach to occupational health in firefighting. In this sense, examining how job and personal resources (e.g., social support, resilience, or emotional competencies) can be strengthened is crucial to designing effective interventions that not only mitigate existing burnout symptoms but also prevent their onset.

Finally, future studies should broaden the geographical scope of the current literature, which remains concentrated in East Asia and selected European countries. Greater attention should be paid to underrepresented sociocultural contexts such as Latin America, Africa, and Mediterranean countries like Spain, where structural differences may shape the manifestation and prevention of burnout.

Moreover, once methodological standards are better aligned across studies, future reviews should aim to conduct meta-analyses to quantify effect sizes, identify moderators, and strengthen the empirical basis for intervention design. Statistical integration will be especially valuable for clarifying the magnitude of associations between burnout and its most critical predictors and outcomes.