Abstract
Firefighters work in uncertain, hazardous conditions where safety protocols do not fulfil complex requirements in a dynamic fireground and real-time crisis actions. The gap calls for an immediate implementation of flexible, pragmatic risk management principles to improve workplace security. This research assessed a workplace risk management model developed for firefighting in Malaysia. The model centred on eight core elements: context setting, risk perception, hazard identification, risk analysis, risk assessment, anticipation, risk treatment, and communication. Using a cross-sectional design, the researchers assessed 331 operational firefighters from 45 fire stations in the Klang Valley area. The data were collected through an administered, validated questionnaire and processed through Confirmatory Factor Analysis and Structural Equation Modelling (SEM) using IBM SPSS AMOS version 26 software. The findings indicated six elements comprising context setting, risk perception, hazard identification, risk analysis, anticipation, and communication significantly and positively affected perceived workplace safety. Conversely, risk assessment and risk treatment did not demonstrate any significant influence. The finding suggests that conventional models lack flexibility in highly dynamic crisis contexts. The research emphasizes real-time communication and flexible and team-based risk approaches and contributes theoretical and practical implications toward improved safety in rescue missions during firefighting activities.
Plain Language Summary
Firefighters often work in dangerous and unpredictable situations. In these fast-changing environments, standard safety rules sometimes aren’t enough to protect them. Because of this, there is a need for flexible and practical ways to manage risks at work. This study looked at a risk management model created specifically for firefighters in Malaysia. The model focused on eight main areas: understanding the situation, how firefighters perceive risk, identifying hazards, analyzing risks, assessing risks, preparing for possible dangers, deciding how to handle risks, and communicating clearly. The research involved 331 firefighters from 45 fire stations in the Klang Valley region. The team used a survey that had been tested for accuracy and reliability, and they analyzed the results with specialized computer software. The findings showed that six of the eight areas are understanding the situation, risk perception, hazard identification, risk analysis, anticipation, and communication that had a clear, positive impact on how safe firefighters felt at work. However, risk assessment and risk treatment did not have a significant effect. These results suggest that traditional risk models might not work well in fast-moving emergencies. The study highlights the need for real-time communication along with flexible and team-based approaches to managing risk. These changes could help improve firefighter safety during rescue operations.
Keywords
Introduction
Firefighting is commonly described as an occupation with high-risk factors since frequent fires occur in volatile and dangerous settings. The 2023 report by the Fire and Rescue Department Malaysia (FRDM) indicated more than 81,000 fire calls for emergency assistance, up from 79,000 in 2022, indicating an increase in demand for emergency services in Malaysia (FRDM, 2024). Concerns about injuries to Malaysian firefighters and accidents during operation have mounted, highlighting the need for more stringent precautions in an urgent manner. An average of 42,037 operational accidents occurred yearly over a recent 5-year period (2014–2018), as indicated by data from the FRDM (FRDM, 2019). “Operational accidents” in the context here are related to injuries or adverse events during fire combat missions at points of emergencies, as opposed to training and station-based accidents. Prevalent causes in accidents during engagement in fighting fires at points of emergencies include collapsing structures (≈17% of fatalities/injuries to firefighters) and slipping or tripping and falling over (≈18% of severe injuries). These statistics in Malaysia correlate with global patterns as seen using the estimated 63,175 firefighter injuries during duty in 2023 by the American National Fire Protection Association and thousands of infectious disease and hazardous exposures (Campbell & Hall, 2024; Li et al., 2023). These statistics highlight that Malaysian heroes on duty face volatile and problematic working conditions with rapidly mounting dangers (Hassanain et al., 2022; Penney, 2019; Poplin et al., 2015). There is an urgent need for an integrated strategy to contain and mitigate associated dangers to enhance firefighter security and operational performance.
Firefighters encounter dangers on the fireground during an emergency response and in non-emergency environments (such as training exercises, equipment runs, and station house duties). However, “workplace safety” in the context of this study generally describes the working environment in which emergencies are resolved. Unlike a typical workplace, an incident scene is dynamic and uncontrolled, so risk management for firefighting must be highly responsive (Campbell & Hall, 2024; Heidari & Jabbarpoor, 2024). Firefighters must often make almost instantaneous risk decisions with incomplete information. An example is an incident commander who might have moments to determine whether a building is about to collapse or contain hazardous materials when deciding whether to commit crews into a burning building (Zanatta & Amaral, 2012). Thus, there is a need to size up risk under urgent time pressure, and ambiguity is a constitutive aspect of work involving an emergency response (Taheri et al., 2023). Established models of occupational safety based on more time for contemplation and mitigation might not be directly transferable or must be recalibrated to apply to firefighting efforts. While firefighting organizations do have standard operating procedures (SOPs) and safety protocols in place, actual implementation in real time is mainly based on human judgment and communication (Fraser-Mackenzie & Dror, 2011; Kocher et al., 2019; Skinner & Parrey, 2019).
Workplace Safety Factors
In safety science, the workplace is a concept based on the physical space (Smith et al., 2019), organisational safety climate (Luo, 2020), and following procedure (Le et al., 2018), all factors that affect operational conditions. For effective risk management practice in firefighting, all these factors must be synchronized to reduce hazards and establish a safety culture.
Firefighting is widely recognized as one of the most hazardous professions, demanding continuous physical exertion in inherently risky environments (Coca et al., 2010; Smith et al., 2019). The physical conditions firefighters face is shaped by constantly changing factors such as equipment, machinery, layout of workplace safety factors, and ambient environmental conditions (Hemmatjo et al., 2017; Teixeira et al., 2024). These elements directly impact task execution, stress levels, and the potential for accidents. For example, heat and heavy smoke exposure could severely degrade cognitive capacity and physical performance (Coca et al., 2010; Teixeira et al., 2024). Malfunctions in equipment and poor layout planning could impede flow during work, producing greater chances for traffic congestion and on-site accidents (Fialho et al., 2024; Fortunato et al., 2012). While more personnel might temporarily balance inefficiency, it usually creates overcrowding and further diminishes security (Fialho et al., 2024). Therefore, adopting ergonomic design principles and regular maintenance of firefighting equipment are guaranteed ways for lessening risk (Arifin, Juhari, et al., 2024; Sobhani et al., 2017).
Organisational structure and safety governance form the backbone of a systematic approach to implementing and sustaining risk control measures in firefighting (Berhan, 2020). Successful operations depend heavily on well-defined management systems, efficient communication channels, and robust mechanisms for ensuring regulatory compliance (Al Mazrouei et al., 2020; Luo, 2020). These structures usually rest on policies and formal documents that establish responsibilities, set out procedural requirements, and capture the organisation’s general commitment to workplace safety. Ambiguous and poorly communicated policies create room for misunderstandings, violations, and recurring mishaps (Kim et al., 2019; Zarei et al., 2021). Conversely, concise, openly accessible, and regularly disseminated safety guides improve accountability, foster an agreed-upon understanding of workplace safety practices, and coordinate teamwork (Boustras et al., 2015; Kim et al., 2019).
Standard Operating Procedures (SOPs) comprise the procedural framework of firefighting safety, providing organised workflows, simple-to-implement safety checklists, and well-delineated emergency response protocols (Berhan, 2020; C. F. Chen & S. C. Chen, 2014). SOPs enable consistent action, encourage safety behaviour, and improve coordination in high-pressure and time-critical operations. Standardising responses minimises variability in decision-making and increases overall operational efficiency. However, compliance with SOPs is insufficiently researched despite their utility in field settings characterised by uncertainty and rapidly changing conditions (Duncan et al., 2014; Le et al., 2018). Fatigue, on-the-spot creativity, and deep-seated organisational habits often fuel procedural drift and non-compliance. These issues must be addressed through ongoing training programs, regular system drills and tests, and intense supervisory routines (Penney et al., 2024; Weinschenk et al., 2008). Routine audit and investigations involving policy review, observation and interview with staff also play an important role in detecting procedural gaps and promoting an active, compliance-driven safety and quality system (Asbury, 2019; Macpherson et al., 2021).
The physical safety environment, organisational safety climate, and procedural compliance dimensions collectively influence the operational conditions in which firefighting workplace safety factors are assessed. Aligning these elements with comprehensive risk management practices is essential for mitigating hazards and fostering a resilient safety culture within emergency response operations.
Risk Management in Firefighting
Risk management in firefighting is grounded in the ISO 31000:2018 Risk Management Guidelines, which provide the core theoretical structure for establishing context, identifying, analysing, evaluating risks, applying treatments, and maintaining ongoing monitoring. In this study, ISO 31000 serves as the primary conceptual framework, ensuring that each element is anchored in internationally recognised best practice. To translate these principles into a fire service context, the study adopts the structured, participatory RM model developed by Poplin et al. (2015a), which operationalises ISO 31000 through a three-phase process of hazard scoping, risk assessment, and control implementation. This adaptation highlights flexible, iterative procedures and actively involves firefighters in hazard detection and control, reflecting ISO’s inclusion principle. Further, High-Reliability Organisation (HRO) principles are incorporated as a contextual enhancement, emphasising vigilance, adaptability, and deference to expertise in dynamic, high-risk environments. This layered approach positions ISO 31000 as the structural base, Poplin et al.’s model as the operational translation, and HRO concepts as performance enhancers for unpredictable firefighting scenarios.
Context Setting (Risk Context): The situational context and organisational environment lay the foundation for effective risk management. Clear safety policies and strong leadership commitment help create a setting where firefighters are encouraged to prioritise safety (Arifin et al., 2011; Kim et al., 2019). In high-risk environments such as active fire scenes, context-setting may include pre-incident planning, safety briefings conducted at the incident command post, the integration of effective technologies, and a culture that empowers personnel to raise concerns about hazards (Dror & Charlton, 2006; Fraser-Mackenzie & Dror, 2011; Penney et al., 2020a). A strong safety context ensures firefighters have a broader operational landscape view before taking action in the fireground. This involves identifying internal factors, such as crew experience and training levels, and external influences like building design or weather conditions, all of which may affect operational decisions (Penney et al., 2020a; Svensson, 2002). Research suggests that a well-established context, supported by policies, organisational culture, and thorough preparation, has a positive impact on firefighters’ workplace safety because it supports all subsequent risk management actions (Dror & Charlton, 2006; K. Khan et al., 2018; Kocher et al., 2019; Skinner & Parrey, 2019).
Risk Perception: Risk perception of individual firefighters and commanders refers to their ability to recognise dangers in a given situation, which is fundamental to their decision-making and behaviour (Gutnik et al., 2006; T. Liu & Jiao, 2018). A strong sense of risk perception has been associated with more cautious actions, such as consistently using personal protective equipment (PPE) and adherence to safety protocols (Dror & Charlton, 2006; Fraser-Mackenzie & Dror, 2011). Risk perception often varies with experience. For example, seasoned firefighters may intuitively detect when a building is at risk of flashover, while those with less experience might not recognise the warning signs (Fialho et al., 2024; Larsen et al., 2021; T. Liu & Jiao, 2018). When firefighters can accurately assess risks, they are more likely to take proactive measures to prevent harm and avoid unnecessary exposure (Bronfman et al., 2020; Larsen et al., 2021). Based on this, it can be suggested that higher levels of risk perception contribute to a safer working environment, as individuals aware of potential hazards are more inclined to act in ways that reduce those risks.
Risk Identification: Risk identification refers to systematically detecting hazards before or as they emerge. In practical contexts, this process may involve checklists and formal inspections. Firefighting typically includes an initial scene size-up and continuous hazard scanning throughout the operation (Salmon et al., 2020). Early identification of risks, such as signs of structural collapse, flammable substances, or exposed electrical hazards, allows teams to address them proactively (Jaafar et al., 2018; Thompson et al., 2020). Fire services often emphasise conducting a thorough 360-degree size-up upon arrival to identify key threats (Widagdo & Cahyono, 2020). Prior research underscores that risk identification is a critical proactive step in reducing hazards and increasing operational readiness. The effectiveness of firefighters and commanding officers in identifying risks is positively associated with better safety outcomes. Known hazards can be managed effectively, while unidentified dangers are more likely to catch crews off guard and lead to accidents (Frost et al., 2015; Penney et al., 2020b).
Risk Analysis: Once hazards are identified, risk analysis involves assessing an event's likelihood and potential consequences (Chemweno, 2020; Roughton & Crutchfield, 2016). In a fire incident, firefighters must conduct a real-time risk analysis. Although formal methods of risk analysis, whether qualitative or quantitative, are well established in safety engineering (Finney, 2005; Marsden-Smedley & Whight, 2011), decisions on the ground are typically made through experience-based judgement. Fire officers often need to determine acceptable levels of risk within minutes (Frost et al., 2015; Kocher et al., 2019). For instance, an incident commander might decide it is acceptable to perform an interior search if the likelihood of rescuing a victim is high, but would likely avoid entry if the structure appears unsafe. Quick and sound risk analysis ensures that resources are allocated effectively and that dangerous tactics are avoided when the expected benefit is low (Gutnik et al., 2006; Hsu et al., 2012). Strong risk analysis capability, even when carried out mentally rather than through formal calculation, is anticipated to improve workplace safety by guiding firefighters to take actions that achieve a favourable balance between risk and reward.
Risk Assessment: Risk assessment involves combining hazard identification and risk analysis, followed by a comparison against established risk criteria to determine whether risks are tolerable (Fialho et al., 2024; Hassanain et al., 2022). In this context, the study defines risk assessment as deciding which risks are acceptable and require intervention. In firefighting, this often translates to a go or no-go decision, such as evaluating whether it is acceptable to enter a building based on prevailing conditions and the potential benefits (Arifin, Ahmad, et al., 2023, 2024; Heidari & Jabbarpoor, 2024). Clear criteria should guide these decisions. For instance, Incident Command policies might state that withdrawal is mandatory if flashover conditions are imminent (Juhari & Arifin, 2020; Oliveira et al., 2018). Practical risk assessment helps ensure that responders do not overextend themselves in hazardous situations (Bahr, 2015; Hassanain et al., 2022). However, applying risk assessment in real-time operations can be difficult due to constraints such as limited information and time pressure (Hopkin, 2018; Poplin et al., 2015). This study considers risk assessment as a critical factor in evaluating whether, in practice, firefighters' ability to assess and respond to risks meaningfully contributes to improved safety outcomes.
Risk Anticipation: Fireground conditions can change rapidly, so firefighters need to predict how an incident might develop based on what they see, what they have experienced before, and a clear understanding of risk (Ali et al., 2022; Bronfman et al., 2020; T. Liu & Jiao, 2018). Experienced officers often develop an instinct for outcomes, such as thinking, “if we ventilate here, the fire might move in that direction.” This ability to look ahead is risk anticipation (Okoli, 2020; van den Bosch & Helsdingen, 2002). It involves learning from previous incidents and understanding fire behaviour, such as recognising signs of a possible flashover by observing smoke. Mohd Zahari et al. (2024) suggests that using past data and scenario knowledge helps officers anticipate risks more accurately and make quicker decisions. When firefighters can anticipate risks effectively, they are more likely to take proactive steps such as ordering an evacuation or requesting more support, which helps keep the operation safer (Bakx & Richardson, 2013; Ferguson et al., 2024).
Risk Treatment: Risk treatment means taking action to reduce or remove hazards. In industrial settings, this might involve using safety equipment, following rules, or changing tasks (Arifin, Ali, et al., 2023; Reeve, 2005). In firefighting, many of these controls are already part of everyday practice. Examples include wearing SCBA gear, setting up collapse zones, and having Rapid Intervention Teams ready to help. The success of these safety measures depends on how well they are used during an emergency (Pollack et al., 2017; Snedaker & Rima, 2014). For example, are firefighters following safety rules like the two-in, two-out policy? Are they pulled out quickly if a building starts to collapse? This study examines whether people think these actions make things safer (Poplin et al., 2015; Van Coile et al., 2019). Risk treatment can prevent injuries and save lives (Hong et al., 2012). However, safety steps might be missed or not followed in fast-moving and stressful situations (Paterson et al., 2022), which this study will explore.
Monitoring and Communication: Continuous monitoring and communication are essential for safety during fast-changing firefighting operations (Arifin, Ali, et al., 2023; Griffin & Hu, 2013; Hallowell et al., 2013). Monitoring can involve a safety officer keeping track of the environment or each firefighter staying alert to new dangers (Hedlund et al., 2016; Y.-J. Liu et al., 2013). Communication helps ensure that if one person spots a risk, like a crack in the wall, the whole team is immediately warned. Strong communication up and down the command chain is often mentioned in incident reports as a key factor in preventing accidents (Gao et al., 2019; Zarei et al., 2021). When communication fails, it can lead to serious consequences, such as the incident commander not knowing that a team has moved to a new location. Sharing information quickly allows teams to respond together to new threats (Ahmed Naji et al., 2020; Dasgupta et al., 2014; Zarei et al., 2021). This study suggests that good monitoring and clear communication improve firefighter safety.
Effective firefighting operations require continuous and informed decision-making in dynamic, high-risk environments. This is in line with the ISO 31000:2018 Risk Management Guidelines (Table 1). The study must clearly distinguish risk-related concepts to ensure a common operational language and consistent application in the field. However, based on the literature, the distinctions between risk analysis, risk assessment, and risk anticipation remain conceptually and empirically ambiguous, with the terms often used interchangeably. The table below clarifies these constructs by presenting ISO-aligned definitions, key decision points, and practical firefighting examples. This structured view enables operational leaders and crew members to progress from understanding “what could happen” to deciding “what should be done” and anticipating “what is likely to happen next.”
ISO 31000:2018 Aligned Definition.
Several studies support participatory risk management models and ISO 31000:2018 in high-risk industries, but their suitability for fast-paced firefighting is questioned. Ahmadi et al. (2020) found that linear hazard assessment and treatment cycles can be too rigid when conditions change within minutes. Alternative approaches, such as the Dynamic Risk Assessment (DRA) model (Sarvestani et al., 2021) and High-Reliability Organisation (HRO) theory (Dwyer et al., 2023), stress distributed authority, adaptive decisions, and continuous situational scanning. These emphasise safety through rapid information exchange and real-time adjustments rather than strict procedural adherence. This study adopts ISO 31000 as the core framework but acknowledges that firefighting may require hybrid models combining adaptability with structure. Most prior research addressed single factors due to small or unrepresentative samples, incomplete construct coverage, and context-specific tools, limiting integrated model testing. Evidence suggests each risk management element can enhance safety, yet its combined effects remain unclear. Existing studies often isolate one or two factors, such as safety climate or communication. This study fills the gap by empirically testing a unified eight-element model to show that stronger risk management practices improve firefighter safety during emergency operations.
Research Hypotheses
The study is guided by the question: How do key risk management elements—context setting, risk perception, risk identification, risk analysis, risk assessment, risk anticipation, risk treatment, and monitoring and communication—affect the safety of the firefighting workplace during emergency operations? The objective is to develop and validate a structural model quantifying these relationships. In line with this, this study proposes the following hypotheses for empirical testing:
These hypotheses reflect the expectation that improving risk management will lead to a safer working environment for firefighters (e.g., fewer accidents or better safety compliance during operations). Figure 1 illustrates the proposed workplace accident risk management model for firefighting operations. The model posits eight latent independent variables (the risk management elements) and one latent dependent variable “Workplace Safety” which may be reflected by factors such as the physical safety of the environment, the organisational safety climate, and adherence to safe work procedures on scene.
A conceptual model of risk management elements influencing firefighter workplace safety factors in firefighting operations.
Method
Study Design and Participants
This study used a cross-sectional survey design with quantitative data analysis to evaluate the proposed risk management model. The target population comprises operational fire officers based in the Klang Valley region of Malaysia, including Selangor, Kuala Lumpur, and Putrajaya. This region represents a high-demand area for firefighting services, with 45 fire stations and 2405 operational firefighters (FRDM, 2019). A stratified random sampling approach was adopted to ensure proportional representation across regions and ranks. Krejcie and Morgan (1970) formula set the target sample size at 331, representing approximately 13.8% of the total population. This also meets the sample size requirements for Structural Equation Modelling (SEM), as Roscoe (1969) suggested. Sample distribution aligns with the population size in each region: 190 participants from Selangor, 127 from Kuala Lumpur, and 14 from Putrajaya. Within each region, additional stratification by seniority guaranteed representation at all job grades (KB19 to KB38 and Auxiliary Fire Officers). This is due to risk perception and accountability differ at the seniority level, with senior ranks usually taking part in strategic decisions and junior ranks involved in operational tasks.
Participants were randomly selected and invited to complete the survey. All 331 invited officers participated in the survey and 100% response rate was achieved because the survey was administered in person during mandatory training sessions. All invited participants completed the instrument on-site before leaving the session, eliminating the possibility of non-response. These high participation numbers were due to department support, ease in taking part through an online-based survey, and participants’ concern for safety issues. The final sample distribution was 57% from Selangor, 38% from Kuala Lumpur, and 5% from Putrajaya. All respondents were male, reflecting the composition of the operational firefighter workforce at the time. The mean age was 35.4 years (SD = 7.8), and average length of service was 12.3 years (SD = 8.1). Table 2 presents the breakdown of the population and sample by rank and region.
Stratified Random Sampling.
Note. K.Lumpur = Kuala Lumpur. Sample numbers were chosen to be approximately proportional (~13%–14%) of each region’s population. This approach yielded an overall sample of 331, which is about 13.8% of the total 2405 firefighters.
Procedures and Ethics
Data were collected using an online questionnaire administered via Google Forms between 5 April and 5 August 2021. Official approval was obtained from relevant authorities, granting permission to approach fire stations. The survey began with an introductory page outlining the study’s objectives, assuring participants of confidentiality and anonymity, and requiring informed consent through a mandatory checkbox. Only consenting individuals could proceed. Participation was voluntary, with respondents informed that there were no right or wrong answers. All data were reported in aggregate, with no individual or station-specific identifiers. The questionnaire was delivered in Malay, the official language used in Malaysian public service, and one in which all participants were proficient. Before full deployment, the survey was pilot tested with 30 firefighters from a station outside the Klang Valley to evaluate item clarity and relevance. Based on feedback, minor adjustments were made to improve understanding, including simplifying wording and clarifying specific terms such as “near-miss.” The pilot also confirmed acceptable internal consistency, with all Cronbach’s alpha values exceeding .70. To ensure complete datasets, all items in the questionnaire were set as mandatory. Consequently, there were no missing responses. The average completion time was approximately 15 to 20 min.
Survey Instrument
A structured questionnaire was developed based on existing literature and risk management frameworks, tailored to the firefighting context. Experts reviewed the survey and pilot-tested it to ensure clarity and reliability. These items captured how firefighters felt about safety at the scene, reflecting the effectiveness of risk management in high-risk environments. Higher scores indicated greater perceived safety. The questionnaire used clear, simple language and was pilot tested with 30 firefighters to ensure understanding. Responses were recorded using a five-point Likert scale (1 = Strongly Disagree to 5 = Strongly Agree), following the approach of DeArmond et al. (2011), Guo et al. (2016) and Vinodkumar & Bhasi (2010). The self-report format allowed firefighters to share their direct experiences with safety and risk practices during operations. It was divided into three main sections:
Section A: Demographics: This section collected background information including age, gender, rank, years of service, and accident history. These details were used to describe the sample but were not part of the main model.
Section B: Risk Management Practices: This section measured the eight independent risk management elements:
Context Setting: Assessed the presence of safety policies, leadership support, and pre-incident planning. Example: “When firefighters feel safe to speak up, risks are managed better.”
Risk Perception: Measured awareness and understanding of risks at the scene. Example: “Firefighters who can recognise risk make better safety decisions.”
Risk Identification: Focused on how hazards are identified early and systematically. Example: “Identifying hazards early helps prevent accidents during firefighting.”
Risk Analysis: Looked at evaluating risk severity and likelihood. Example: “Good risk judgement helps avoid unsafe choices.”
Risk Assessment: Examined decisions about whether a risk is acceptable. Example: “Good risk checks help avoid very dangerous situations.”
Risk Anticipation: Measured the ability to predict problems using experience and fire dynamics. Example: “Using past experience helps firefighters make better decisions on the fireground.”
Risk Treatment: Assessed the implementation of control measures. Example: “Fixing unsafe conditions quickly can prevent injuries or accidents.”
Monitoring and Communication: Evaluated ongoing observation and sharing of safety information. Example: “Firefighters should warn the team right away when they notice danger.”
Each construct included five items, most adapted from past research in safety climate and risk management and reworded to fit firefighting. Three experts (two from fire safety and one academic) confirmed content validity.
Section C: Workplace Safety: This section measured the dependent variable, firefighter workplace safety, based on firefighters’ perceptions of safety during operations. It focused on three key areas:
Physical Safety of the Environment: Example: “The operational area is generally secured and safe for us to work in.”
Organisational Safety Climate: Example: “Safety is not compromised for speed—our leaders value safety in every operation.”
Procedural Compliance: Example: “Standard operating procedures are followed closely during incidents.”
Reliability and Validity of Measures
The study calculated Cronbach’s alpha for each construct to assess internal consistency. All constructs showed acceptable reliability in the pilot test with alpha values above .70. In the final dataset, Cronbach’s alpha values ranged from .78 to .95, indicating strong internal consistency. Composite Reliability (CR) was also calculated, with values between .80 and .93, all exceeding the recommended threshold of .70. Content validity was ensured by grounding all items in the literature and having them reviewed by subject matter experts, which was in line with the guidelines of Hair et al. (2010) and Collier (2020). Construct validity was examined using Confirmatory Factor Analysis (CFA), as detailed in the Results section. This included checks for convergent validity (to confirm that items accurately reflect their intended construct) and discriminant validity (to confirm that constructs are distinct from one another).
Data Analysis
Data were analysed using IBM SPSS 26 for initial checks and descriptive statistics, and AMOS 26 for Confirmatory Factor Analysis (CFA) and Structural Equation Modelling (SEM). SEM was appropriate as it allows simultaneous testing of multiple relationships while accounting for measurement error. It is well-suited for confirming whether the data support the theoretical model (Schreiber et al., 2006). All anonymised survey data are provided in the supplementary material to support transparency and replication. The analysis followed several steps:
Descriptive Statistics and Normality: Means and standard deviations were calculated for all items. Skewness and kurtosis values were used to check normality. Most items had skewness between −0.645 and 0.593 and kurtosis between −1.737 and 2.828, indicating acceptable normal distribution (Chua, 2020; Mackey & Gass, 2021; Pallant, 2020).
Exploratory Factor Analysis (EFA): The Kaiser-Meyer-Olkin (KMO) test and Bartlett’s Test of Sphericity confirmed the suitability of the data for factor analysis. The KMO value was above 0.80, which is considered meritorious (Nkansah, 2018). Nine items were removed during EFA due to low factor loadings or high cross-loadings. The remaining items were evaluated using Average Variance Extracted (AVE > .50), Composite Reliability (CR > .70), and Cronbach’s Alpha (α ≥ .70) to ensure convergent validity and internal consistency (Collier, 2020; Hair et al., 2010).
Confirmatory Factor Analysis (CFA): CFA was conducted to test the measurement model of the eight risk management constructs and workplace safety. Model fit was assessed using several indices: χ2/df below 3; TLI, NFI, GFI and AGFI at or above .90; and RMSEA below .08 for the full measurement model (Hu & Bentler, 1999; Kline, 2023). For single-factor CFA models, a threshold of RMSEA < .10 was adopted following recommendations for models with small degrees of freedom, limited item counts, or high model parsimony (Marsh et al., 2004). This distinction avoids unnecessarily rejecting theoretically sound models with slight RMSEA elevations due to structural constraints. These values showed that the model had an acceptable fit. Convergent validity was supported as AVE values exceeded 0.50 for all constructs. Discriminant validity was confirmed by ensuring that each construct’s AVE was greater than the squared correlations between constructs, and all correlations were well below .95 (Fornell & Larcker, 1981; Hair et al., 2010; Kline, 2023).
Structural Equation Modelling (SEM): SEM was used to test the structural model shown in Figure 1, where the eight risk management practices predict workplace safety. Maximum Likelihood (ML) estimation was applied, which was appropriate due to the normality of the data. Model fit indices again met recommended thresholds. Each hypothesis (H1–H8) was tested by examining the standardised path coefficients. A path was considered significant if p < .05. The R2 value for the Workplace Safety Factors construct was also reported to show how much variance was explained by the predictors.
Hypothesis Testing: Each hypothesis represented a direct relationship between one risk management factor and workplace safety. A hypothesis was supported if the corresponding path was positive and statistically significant (p < .05). Significance at the .01 or .001 level was also noted.
Results
Demographic Analysis
The demographic breakdown of 331 respondents indicates an obvious gender imbalance with 322 men (97.3%) and just nine women (2.7%), indicating the male-preferred nature of FRDM Officer positions in Klang Valley. The age group is as follows: almost half (46.5%) were 18 to 29. This is followed by 32.0% for 30 to 39 years, 14.8% for 40 to 49 years, and 6.7% for those 50 years old or older. The majority (77.9%) held the KB19 grade level. This is followed by KB22 (15.1%), and the rest were distributed in all the other grades, including AFO, KB24, KB26, KB29, KB32, and KB38. Service duration ranges from 57.1% of more than 6 years’ service to others with less than 1 year (18.1%), 2 to 3 years (14.2%), and 4 to 5 years (10.6%). Surprisingly enough, many with over 6 years’ experience belonged to the 18 to 29 years age group, indicating early recruitment and persistent employment. On frequency on site at fire scenes, 52.9% attended 2 to 3 times per week and 21.5% attended every day. Other frequencies included 2 to 3 times a month (14.8%), once every week (6.6%), once every month (3.9%), and one person (0.3%) never attended. Regarding duration spent at fire scenes, 51.4% spent 3 to 4 hr per case and 35.0% spent 1 to 2 hr. Smaller proportions spent 5 to 6 hr (8.8%) and more than 7 hr (4.8%). This indicates that most participants spend 1 to 4 hr per case performing firefighting work.
Descriptive Statistics and Preliminary Analysis
All 331 responses were valid and included in the analysis. Table 2 presents the descriptive statistics for each latent construct, including mean scores, standard deviations, and reliability coefficients. Respondents generally agreed that risk management practices were present in their units, with mean scores ranging from 3.51 to 4.02 on a 5-point scale. For example, Context Setting had a mean of 3.51 (SD = 0.916), while Monitoring and Communication had a mean of 4.01 (SD = 0.642). The mean for the Workplace Safety outcome was 4.21 (SD = 0.643), indicating a high level of agreement that safety is maintained during operations. All scales demonstrated strong internal consistency, with Cronbach’s alpha values between .864 and .897. These results exceed the recommended threshold of .70 and confirm high reliability across constructs. The normality of the data was also assessed. Table 3 shows skewness values ranged from −.645 to .666, and kurtosis values ranged from 1.737 to 2.828. These results fall within acceptable limits, confirming that the data are typically distributed and appropriate for parametric analysis, including Structural Equation Modelling (SEM).
Mean Score, Cronbach Alpha and Normality.
Correlation is significant at the .01 level (p < .01).
In addition, Pearson correlation coefficients were calculated to examine the relationships between the eight risk management elements and perceived workplace safety. All correlations were positive and statistically significant (p < .01), supporting the hypothesis that stronger risk management practices are associated with better safety outcomes. The strength of the correlations varied. Monitoring and Communication had the strongest correlation with workplace safety factors (r = .571, p < .001), followed by Risk Anticipation (r = .559, p < .001) and Risk Assessment (r = .526, p < .001). Moderate correlations were found for Risk Identification (r = .469, p < .001) and Risk Analysis (r = .445, p < .001). Risk Treatment (r = .382, p < .001) and Risk Perception (r = .413, p < .001) also showed positive associations. The weakest correlation was for Context Setting (r = .156, p = .005), although it remained statistically significant. These results suggest that firefighters who reported strong practices in areas such as continuous monitoring, Communication, and anticipation of hazards were also more likely to report feeling safe during operations. In contrast, a general safety context was less directly linked to immediate perceptions of workplace safety. This point is explored further in the discussion.
Exploratory Factor Analysis Results
The suitability of the data for factor analysis was assessed using the Kaiser-Meyer-Olkin (KMO) measure and Bartlett’s Test of Sphericity, as shown in Table 4. The results indicated that all constructs met the minimum threshold for factor analysis. KMO values ranged from 0.651 to 0.896, with Context Setting and Workplace exhibiting the highest adequacy (KMO = 0.896), indicating excellent sampling adequacy. Although the KMO values for Risk Identification (0.651) and Risk Treatment (0.678) fall into the borderline ‘mediocre’ range according to some guidelines, both constructs were retained due to their theoretical relevance, comprehensive content coverage, and acceptable CR and AVE values. Additionally, Bartlett’s Test of Sphericity was significant (p < .001) for all constructs, suggesting that the correlation matrices were not identity matrices and that there were sufficient inter-item correlations to justify the application of factor analysis. These results confirm the overall adequacy of the dataset for exploratory factor analysis.
Kaiser-Meyer-Olkin Test and Bartlett’s Test of Sphericity.
Table 5 presents the results of the Exploratory Factor Analysis, including the items retained and removed, as well as the Average Variance Extracted (AVE), Composite Reliability (CR), and Cronbach's alpha values for each construct. Forty-seven items were retained across nine constructs, while eight were removed due to low or cross-loading factor loadings. All constructs achieved AVE values above 0.50, confirming acceptable convergent validity. Context Setting had the highest AVE at .731, while Risk Anticipation had the lowest acceptable AVE at .507. These results indicate that each construct accounted for a sufficient portion of item variance. Composite Reliability (CR) values ranged from .780 to .930, showing strong internal consistency. Similarly, Cronbach’s alpha coefficients ranged from .765 to .932, confirming the reliability of each construct. The Workplace Safety construct, which included more items, also showed excellent reliability (CR = .861, α = .932) despite removing two items. Overall, these findings confirm the instrument's structural validity and show that the retained items are valid and reliable for assessing workplace accident risk factors in the firefighting context.
Total Items Used with Reliable Values.
Confirmatory Factor Analysis Results
The Confirmatory Factor Analysis (CFA) of the nine-factor model, comprising eight risk management factors and one workplace safety factor, showed that the model fit the data well. The key fit indices were: χ2/df = 2.569, TLI = 0.934, NFI = 0.973, GFI = 0.965, AGFI = 0.935, RMSEA = 0.069 and p = .000. These values meet or closely approach the commonly recommended thresholds (TLI and NFI ≥ 0.90, RMSEA < 0.08), as shown in Table 6. The statistically significant chi-square is expected due to the large sample size and model complexity. Therefore, the study relies more on relative fit indices to evaluate model adequacy. All items loaded significantly on their respective constructs (p < 0.001), with standardised loadings ranging from 0.60 to 0.88. This confirms convergent validity, as each set of items shares substantial variance with its underlying factor. All constructs' Average Variance Extracted (AVE) values exceeded 0.50, typically ranging from 0.55 to 0.70.
Goodness of the Model Fit.
Discriminant validity was assessed for the eight elements of risk management to determine the extent to which each construct is distinct from the others. Following (Kline, 2023), correlations between variables in the model estimation did not exceed 0.95. The assessment was conducted using the Fornell–Larcker criterion, which compares the square root of the average variance extracted (√AVE) for each construct with its correlations with other constructs (Fornell & Larcker, 1981). Table 7 presents the discriminant validity results, with diagonal entries showing √AVE values and off-diagonal entries representing inter-construct correlations. The highest correlation was observed between Context Setting and Risk Treatment (0.826), followed by Context Setting and Risk Identification (0.774). Because the √AVE for Risk Treatment is 0.804, the Fornell–Larcker criterion is not met for the Context Setting–Risk Treatment pair, although the √AVE for Context Setting (0.855) exceeds both correlations. To address this, the Heterotrait–Monotrait (HTMT) ratio was calculated as a more stringent test (Henseler et al., 2015). All HTMT values were below the conservative threshold of 0.85, and their 95% bootstrap confidence intervals excluded 1.00, supporting discriminant validity despite the Fornell–Larcker exception.
Correlations Among Latent Constructs and √AVE.
To assess common method bias, the study conducted Harman’s single-factor test, which revealed that the first factor accounted for less than 50% of the variance, suggesting that common method bias was not a major concern (Kock, 2021). In addition, the unmeasured latent method construct (ULMC) approach indicated no substantial changes to the factor loadings, providing further evidence that common method bias was minimal.
Structural Model and Hypothesis Testing
The structural equation model (SEM), illustrated in Figure 2, was analysed to test the hypothesised relationships. The structural model analysis showed that respondents consider workplace safety factors components a key element in forming the accident risk management model for firefighting operations by the FRDM. The model fit indices were very similar to those from the Confirmatory Factor Analysis (CFA), as expected, since the only change was the addition of directional paths representing hypotheses. The fit statistics were: χ2/df = 2.569, RMSEA = 0.069, GFI = 0.965, AGFI = 0.935, TLI = 0.934, NFI = 0.973, and p value = .000. These values indicate that the model fit remains acceptable and that adding the hypothesised paths did not negatively affect the model’s performance. Overall, 71% of the variance in the workplace safety factors construct could be explained by the eight main elements tested, indicating strong model explanatory power. Six out of the eight elements had a statistically significant influence on the workplace safety factors construct. In the structural model, Monitoring & Communication showed the largest positive effect on workplace safety factors (β = .330, p < .001). In addition, Risk Perception (β = .143, p < .01), Context Setting (β = .116, p < .05) and Risk Analysis (β = .109, p < .01) also made significant positive contributions to the workplace safety factors construct.
Structural Equation Model of firefighter risk management influencing workplace safety factors.
The structural model analysis revealed that six of the eight hypothesised relationships were statistically supported. Monitoring and Communication (β = .330, p < .001) was the strongest predictor of workplace safety factors, highlighting the importance of continuous information sharing in dynamic firefighting contexts (Table 8). Significant positive effects were also observed for Context Setting (β = .116, p = .045), Risk Perception (β = .107, p = 0.030), Risk Identification (β = .129, p = .002), Risk Analysis (β = .153, p = .004), and Risk Anticipation (β = .119, p = .022). These findings indicate that proactive and ongoing situational strategies, including planning, perception, hazard identification, analytical judgment, and anticipation, contribute significantly to firefighter safety during operations. In contrast, Risk Assessment (β = .055, p = .183) and Risk Treatment (β = .088, p = .057) were not statistically significant. This suggests that formal evaluation of risk tolerance thresholds and subsequent treatment actions may be less influential under the rapid and high-pressure conditions of fireground operations. Instead, early-stage planning, situational awareness, and effective communication are the dominant drivers of workplace safety within FRDM’s risk management framework.
Results of Structural Model—Path Coefficients for Hypothesized Effects.
Note. WSF = Workplace safety factors; S.E. = Standard Error of coefficient; CR = Critical Ratio (t-value). “Supported” indicates whether the hypothesis is confirmed at p < .05 level.
p < .05, **p < .01, ***p < .001.
Discussion
This study examined how a comprehensive set of risk management elements contributes to safety in firefighting operations. Overall, the findings confirm that certain risk management practices significantly positively impact firefighters’ operational safety, while others appear less directly effective in the heat of an operation. In this discussion, the study interprets each significant result, connects it with the context of firefighting and prior research, and draws theoretical and practical implications.
The fact that the model accounts for 71% of the variance indicates that risk management practice is strongly associated with perceived safety in high-risk activities. While the primary model adhered to RMSEA < .08, single-factor models were evaluated using RMSEA < .10 in accordance with established guidelines for small-df models, ensuring theoretical coherence without overfitting. This is noteworthy because it affirms a core principle within the science of safety that good management is less about individual actions and more about integrated and systemic management. It is about all the steps from planning and preparation through decision-making and action in real time during high-risk events. This confirms and supports an argument by science writers such as F. Khan et al. (2015), who argue that safety is not an inherent result but an emergent property of an integrated system of practice and behaviour. This finding is particularly noteworthy due to its demonstration within a context for fighting fires. In this context, hasty decisions and unpredictable events occur daily, and the applicability of earlier models is increased primarily by testing within an industry setting.
Standardised coefficients show that monitoring and communication (H8) is the strongest predictor of workplace safety (β ≈ .33), followed by risk analysis (H4) at about 0.15 and risk identification (H3) at about 0.13. Risk anticipation (H6), context setting (H1) and risk perception (H2) had smaller but significant effects of 0.10 to 0.12. These findings confirm that many risk management practices influence safety, but timely communication and awareness are most critical. The model’s R2 of 0.71 indicates strong predictive power. Modification indices suggested minor covariances between similar items, but no new paths were added, preserving the original hypotheses. Of eight hypotheses, six were supported: H1, H2, H3, H4, H6 and H8. Risk assessment (H5) and risk treatment (H7) were not significant, suggesting that proactive awareness, early hazard detection, thorough analysis, forward-looking anticipation, and effective communication have greater impact on safety performance than formal assessment or treatment steps.
The study found that context setting (H1) positively impacts workplace safety, although its effect is negligible compared to other factors. This supports the idea from organisational behaviour theory that a strong safety climate encourages safer work practices (Ullah et al., 2021). In firefighting, context setting includes clear standard operating procedures, strong leadership support, and a culture that values safety. Respondents who felt their organisation had a solid safety foundation also felt safer during operations. This is in line with findings by Gao et al. (2019), who showed that organisations with a strong focus on safety and good communication tend to have fewer accidents. While context alone cannot prevent accidents like a building collapse, it helps create conditions where firefighters are more likely to make safe choices and respond to danger appropriately. For example, a clear safety policy can give a firefighter the confidence to call “Mayday” or back out of a risky situation without fear of getting in trouble. That kind of decision can prevent serious harm. Fire departments should continue strengthening safety policies, training programs, and culture. A strong context supports better decisions and improves safety in high-risk environments.
Risk perception (H2) and risk identification (H3) contributed notably to safety outcomes and illustrated cognitive awareness's role in hazardous circumstances. When firefighters can perceive and identify risks and signs of a backdraft through cognitive awareness, they can better act appropriately and effectively communicate. This is echoed by the work of Fialho et al. (2024), who noted that training and experience in increasing risk perception led to safer behaviour. These findings build upon theirs by demonstrating that risk identification is an essential and anticipatory aspect of preparedness, perceived as often being a procedural and team activity. Research conducted by Cvetkovska and Mitrevska (2024) indicates that actively identifying hazards enhances preparedness. The results support this and show that those who regularly identify risks report a greater preparedness and act more safely in the field. These conclusions accord with sensemaking theory in emergencies. For (Weick, 1993), working with chaotic circumstances begins with noticing and classifying possible dangers and ascribing a label to them. That perception is needed to make quick, informed decisions under pressure. Fire departments might, therefore, invest in experiential training aids that build perception and identification. Virtual reality simulation training, ranging from simple structure as presented through the 5W1H method (Who, What, Where, When, Why, How), and tactical worksheets, allows for valuable aids that some already apply therein. Because the greater role played by risk identification concerning safety is clearly illustrated in practice and context, developing specific positions, as with a dedicated safety officer tasked with identifying hazards, could yield significant improvement for operational safety.
Risk analysis (H4) was found to have a clear and positive impact on safety outcomes, suggesting that even a brief evaluation of hazard likelihood and consequences by firefighters or commanders can lead to better tactical decisions and safer, more effective operations. This finding reinforces the recommendation made by (Mohd Zahari et al., 2024), who emphasised that comprehensive risk analysis factoring in the environment, crew condition, and situational variables leads to better decision-making in the field. This can be as straightforward in practice as an experienced officer making a rapid, intuitive judgment at the scene. The data indicates that teams who engage in risk analysis, even briefly, tend to work more safely. This supports well-established decision-making theories under pressure, such as Klein’s Recognition-Primed Decision (RPD) model (1993), which explains how experienced individuals make quick, sound decisions by mentally simulating likely outcomes (Klein, 1993). One key implication is that firefighter training should go beyond identifying hazards by including realistic exercises that develop the ability to assess and weigh risks against potential benefits, while also providing quick-reference tools to support less experienced commanders in applying risk analysis effectively under pressure. Overall, the findings highlight the value of analytical thinking during emergency operations. Even brief, structured risk assessments can lead to smarter decisions and enhanced safety in the field.
Risk Anticipation (H6) showed a statistically significant positive relationship with the ability to anticipate safety factors in the workplace (β = .119, p = .022). Although an earlier draft contained inconsistent reporting, the final analysis confirms a positive and significant relationship between risk anticipation and workplace safety. This finding aligns with theoretical expectations, emphasising the proactive value of anticipating hazards. This proves that firefighters who can effectively predict and mentally simulate future scenarios tend to minimise operational risks and avoid unexpected events. They identify potential threats and make decisions before the situation becomes critical, rather than waiting for danger to occur before acting. This type of anticipatory behaviour has proven effective in high-risk and time-sensitive environments, where the room for error is small and immediate action is often required. By developing forward thinking, individuals are better prepared to face various possibilities and are not trapped in slow and inefficient reactive responses. This approach aligns with the more comprehensive anticipatory action framework, which emphasises the importance of early intervention based on early warnings, model predictions or data-driven risk analysis. As Uwantege Hart et al. (2024) emphasised, this strategy is now increasingly recognised as more effective, economically sustainable, and operationally robust than traditional reactive approaches to risk and disaster management. These findings reinforce the importance of using future-oriented cognitive strategies in enhancing safety, especially in dynamic and complex operational environments such as emergency response and disaster preparedness. Operations teams can act earlier by practising forward-thinking strategies more accurately and efficiently, significantly reducing risk while ensuring operational continuity in critical situations.
Monitoring and Communication (H8) is the strongest factor associated with safety. This finding supports previous research showing that poor communication often results in firefighter fatalities (Gabor, 2015). This finding confirms that keeping everyone alert and informed during operations is key to managing risk. Firefighters feel safer when their teams are alert and communicate well. This finding is consistent with previous research, such as in the field of Crew Resource Management (CRM) in the aviation industry, which has shown that strong communication leads to higher levels of safety. Many fire services are now also using this approach (Perkins et al., 2022). The data also highlights that it can save lives when a firefighter detects a hazard, such as a collapsing roof, and immediately alerts others. However, the outcome can be tragic if such warnings are not communicated or are ignored. This underscores the critical need for real-time risk management that involves continuous monitoring and feedback, rather than relying on one-time reviews. This also reflects the High-Reliability Organisation (HRO) theory, which states that constant awareness and effective communication help teams stay safe (Ford, 2018). Firefighters who practice HRO principles, such as being sensitive to failure, paying attention to detail, and adapting under pressure, are better prepared to handle the unexpected and prevent accidents from occurring.
Risk Assessment (H5) and Risk Treatment (H7) did not show a direct effect on safety outcomes. This does not diminish their importance but suggests they are harder to apply or less effective in high-tempo firefighting. In such conditions, assessing and acting often occur simultaneously with limited information, making formal step-by-step approaches impractical (Teixeira et al., 2024). Firefighters may instead rely on quick mental judgments shaped by perception, experience, and communication. These rapid decisions often merge into processes like situational awareness or teamwork, making them less visible as distinct steps. These findings are consistent with earlier research, suggesting that traditional sequential risk management approaches, where risks are formally assessed and treated, provide valuable structure. However, they may be less adaptable to the rapid hazard evolution and time compression inherent in firefighting (Ahmadi et al., 2020). Their staged processes often assume the availability of complete information and time for deliberation, conditions rarely present on the fireground. Instead, more flexible and adaptive approaches are recommended. The study supports this view, indicating that conventional assess-and-treat models maybe less adaptable for the realities of frontline emergency response.
The non-significance for Risk Assessment (H5) and Risk Treatment (H7) may be due to measurement and contextual factors. The nonsignificant direct paths in the model may be partly explained by collinearity or shared variance between predictors. Variance inflation factors (VIFs) for all constructs were below the conservative threshold of 3.3, indicating acceptable independence. These results suggest that the influence of certain constructs may operate indirectly through other variables rather than via direct paths. First, in high tempo fireground operations, structured assessment and treatment often occur simultaneously and implicitly, making them indistinguishable from other processes such as risk perception or hazard identification. Second, there may have been statistical overlap with other constructs due to certain survey items that indirectly caught these quick mental judgements. Third, implementing formal assessment and treatment stages in real time may be restricted by operational impediments, such as physical limitations, communication delays, and incomplete information. These factors may explain why participants did not perceive these elements enhancing workplace safety during active incidents. This pattern of results indicates that in a dynamic fireground environment, the processes with the greatest immediate influence (perception, identification, analysis, and monitoring & communication) are those that operate “in the moment” to support rapid situational adaptation. In contrast, context setting and treatment protocols function as background enabling structures. While they are foundational to overall preparedness, their influence may be largely indirect and already embedded within the real-time processes. This framing suggests that the nonsignificant direct paths from these background structures are not anomalies but rather reflect boundary conditions of the model’s applicability in high-tempo operational settings.
This study shows that Malaysian firefighters’ safety operations reflect key High-Reliability Organisation (HRO) traits, particularly the ability to function effectively in hazardous, rapidly changing conditions. The multifactor risk management model, including risk perception, hazard identification, monitoring & communication, aligns with HRO principles such as preoccupation with failure, sensitivity to operations and deference to expertise (Christianson et al., 2011). Like risk treatment, rigid procedural planning appears less critical, highlighting a shift toward dynamic, real-time coordination. Incorporating the Jahn and Black (2017) model underscores the importance of inclusive communication in hierarchical firefighting organisations, including supervisor openness and cross-level facilitation. This supports Berthod et al. (2015), who emphasise cultural integration, shared situational awareness and flexible coordination for inter-organisational reliability. These findings reframe safety as an emergent, socially embedded process maintained through mindful organising and collaborative sensemaking across roles and ranks.
The results show both convergence and divergence, with comparative research conducted in various contexts in various countries. For instance, our findings closely matched those of (Campbell & Hall, 2024) in the US and (Perkins et al., 2022) in Australia, who similarly found that communication and monitoring were important predictors of safety. However, in contrast to our Malaysian dataset, both studies reported significant effects for formalised risk assessment, which may reflect a stronger institutional embedding of these processes and greater operational time allowances in their respective contexts. The study model's unexpectedly negative beta for risk anticipation also deviates from results from Dynamic Risk Assessment (DRA) experiments conducted in the UK, which showed a positive correlation between anticipation and safety outcomes (Okoli et al., 2016). Variations in crew autonomy levels, incident command systems, and the degree of environmental predictability in various geographical areas could cause this disparity. Dynamic incident conditions, characterised by severe time pressure, incomplete or ambiguous information, and rapidly evolving hazards, reduce the utility of conventional sequential risk assessment procedures. In such environments, a streamlined and communication-centred process is more effective for maintaining situational awareness and ensuring coordinated action. This helps explain the stronger influence of perception, identification, analysis, and communication processes in our model relative to background enabling structures such as context setting and treatment protocols.
Based on these findings, the study propose three targeted actions to strengthen risk management practices in firefighting operations. Pre-incident preparation should include the standardisation of briefing formats and radio communication protocols to ensure that all personnel share a common situational understanding from the outset. During incident response, real-time team monitoring, and the use of closed-loop communication can help maintain coordination and adaptability under dynamic and high-pressure conditions. Post-incident processes should incorporate structured debrief templates that systematically capture lessons learned and feed them directly into ongoing training programmes and operational guidelines.
Theoretical Implications
This study identifies the risk management factors most critical during emergencies, showing that human and organisational elements, especially risk perception and communication, affect safety more than formal evaluation and treatment in fast-moving contexts. The findings suggest that safety models for dynamic environments should emphasise adaptability and communication, while structured assessment and treatment suit more stable settings. This aligns with Sharma and Sharma (2016), who highlight resilience through responding, monitoring, anticipating and learning, underscoring the value of real-time monitoring and communication. Viewed through high-reliability organisation theory, the results reflect principles of preoccupation with failure, commitment to resilience and sensitivity to operations. Perception, identification, analysis and communication enable rapid hazard detection and response. Practically, this translates to structured briefings (shared mental model), real-time monitoring, closed-loop communication (situational awareness) and post-incident debriefs (organisational learning). Integrating theory and practice ensures the model advances academic understanding and operational guidance.
Limitations and Future Research
While the study gives valuable information, there are several limitations to consider in future research. The findings should be interpreted within the scope of the study, which measured perceived workplace safety rather than actual incident rates. The generalizability of these findings may be limited by the sample’s gender composition (97% male) and its focus on a single geographic region.
A key limitation of this study is the reliance on self-reported data, which can lead to bias if firefighters overestimate the safety of their work or their ability to manage risks. While anonymity was intended to reduce this, it cannot remove the risk entirely. Future research should include objective measures such as injury records or near-miss reports to balance perception-based responses. It also helps to explore how safety culture influences decision-making over time and how digital tools support quicker, more accurate decisions during emergencies. The sample was limited to the Klang Valley, so the findings may not apply to regions or countries with different systems and cultures. Replicating the study in different contexts, such as urban and rural comparisons or international studies, would test applicability. Adding environmental factors like weather, building structures, and incident timing could also strengthen the model.
Applying this approach to other high-risk professions, such as paramedics, police, or industrial emergency teams, could test its adaptability. However, the cross-sectional design limits the ability to make cause-and-effect conclusions. Safer operations may improve risk management practices, creating a positive feedback loop. However, this should be tested in longitudinal or intervention studies by training one group in communication strategies and comparing outcomes with a control group. Finally, how “risk assessment” and “risk treatment” were measured may not match how firefighters apply them in emergencies. Overlap with other procedures could explain their limited statistical significance. Qualitative methods like interviews or focus groups could provide insight into real-world use, challenges, and communication gaps. These findings could guide targeted training, policy changes, and future interventions to improve safety performance.
Conclusion
This study developed and tested a multifactorial risk management model for Malaysian firefighting, examining eight elements: context setting, risk perception, hazard identification, risk analysis, risk assessment, risk anticipation, risk treatment, and Monitoring and communication. Using Confirmatory Factor Analysis (CFA) and Structural Equation Modelling (SEM), six elements, namely context setting, risk perception, hazard identification, risk analysis, risk anticipation, and especially Monitoring and communication, showed significant positive effects on perceived workplace safety. Monitoring and communication emerged as the strongest predictors, highlighting the importance of real-time situational awareness and information flow. Risk assessment and treatment did not significantly influence safety in fast-paced fireground scenarios, suggesting that traditional sequential procedures may be less adaptable in high-tempo operations. Instead, flexible, proactive, and communication-focused approaches are more effective. Theoretically, the study advances dynamic risk management by framing frontline safety as compliance with structured procedures and the emergent result of real-time cognition, vigilance, and adaptive coordination. Practically, it calls for prioritising intuitive risk perception, rapid hazard identification and analysis, and strong communication in training, policy, and technology. While rooted in Malaysia, the model applies to emergency response agencies globally. Future research should test it across different contexts and explore how digital tools, environmental factors and long-term interventions can improve operational safety. This provides an evidence-based framework for enhancing firefighter safety and resilience through context-specific, real-time risk management.
Supplemental Material
sj-docx-1-sgo-10.1177_21582440251403831 – Supplemental material for Assessing Risk Management Practices on Workplace Safety Factors Among Malaysian Firefighters
Supplemental material, sj-docx-1-sgo-10.1177_21582440251403831 for Assessing Risk Management Practices on Workplace Safety Factors Among Malaysian Firefighters by Kadir Arifin, Mohamad Xazaquan Mansor Ali, Wan Mohammad Zaidi Wan Isa, Mohammad Lui Juhari and Azlan Abas in SAGE Open
Footnotes
Acknowledgements
This study has been supported by Universiti Kebangsaan Malaysia. The authors would like to thank Universiti Kebangsaan Malaysia for supporting for this study.
Ethical Considerations
The researchers confirm that all research was performed in accordance with relevant guidelines/regulations applicable when human participants are involved (e.g., Declaration of Helsinki or similar). This study was approved by the Fire and Rescue Department of Malaysia (FRDM) (approval no. JBPM:PPKPG:100/15/1/1(10)) on March 31, 2021.
Consent to Participate
Informed consent was obtained from all participants involved in the study.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: from Universiti Kebangsaan Malaysia.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The data that support the findings of this study are available from Fire and Rescue Department of Malaysia (FRDM but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Fire and Rescue Department of Malaysia (FRDM).
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
