Prioritizing Worker-Related Factors of Safety Climate Using Fuzzy DEMATEL Analysis

Abstract

The term “safety climate” describes how workers perceive and observe safety within an organization. Workers are typically on the front lines, where they are immediately exposed to safety procedures and working hazards. Their thoughts offer a practical perspective on how safety is applied on a daily basis. Therefore, this study employed the Fuzzy DEMATEL methodology to investigate the critical factors influencing safety climate from workers’ perspective. The research involved nine experts evaluating eight worker-related factors that affect safety climate. The incorporation of fuzzy logic improved the accommodation of the ambiguity and subjectivity inherent in expert judgments, particularly when examining employee viewpoints on safety. The study revealed that the following factors were identified as primary drivers (causal factors) of safety climate: Workers’ safety competence, Workers’ freedom speech about safety matters, and Worker’s ability to perceive hazards. From the perspective of workers, these causal factors have a considerable impact on the other dimensions of safety climate, implying that focused changes in these areas could deliver substantial advantages throughout the full safety spectrum. This distinction provides valuable information for firms to prioritize their safety improvement initiatives and resource allocation. By identifying important cause elements and their relationships, the study offers organizations with a strategic path for improving their safety climate.

1. Introduction

The safety climate refers to the shared thoughts, attitudes, and beliefs of employees about the importance of safety inside an organization [1]. It reflects how safety is prioritized, managed, and integrated into daily operations and organizational culture [2]. Safety climate is based on how employees feel about safety, not necessarily how safe the workplace actually is. It’s about their interpretations and experiences [3]. Moreover, It’s a collective phenomenon. While individuals have their own views, safety climate reflects a general consensus among employees. Safety climate indicates how much emphasis the organization and its members place on safety compared to other priorities. A strong safety climate reduces the likelihood of workplace accidents and injuries [4,5]. Moreover, it fosters a healthier and more secure work environment, improving morale and productivity [6]. Furthermore, a high degree of safety climate ensures compliance with safety requirements and standards, decreasing legal and financial risks for the business [7]. Additionally, it promotes a proactive approach to detecting and addressing safety concerns [8]. Also, employees feel valued and protected when safety is prioritized, leading to increased job satisfaction [9]. Organizations with strong safety climate are perceived as responsible and attract higher personnel [1]. Management changes, incidents, or new safety regulations can all have an impact on the safety climate over time. Consequently, it is crucial to measure the level of safety climate on a regular basis. Measuring safety climate helps pinpoint areas where safety perceptions are weak. This could be due to a lack of managerial commitment, insufficient training, or ineffective communication [10]. By addressing these deficiencies, organizations can proactively lower the likelihood of risky incidents. A high level of safety climate serves as a buffer against safety incidents. Furthermore, comparing safety climate scores over time or against industry benchmarks provides useful insights into performance and helps set realistic targets. Additionally, when employees participate in safety climate surveys, they feel valued, which can boost their commitment in safety measures [11].

Multi-Criteria Decision Making (MCDM) is a structured technique for making decisions, especially when many criteria or factors are handled simultaneously [12,13,14,15,16]. The criteria or factors may be contradictory, making the decision making process harder. The application of MCDM in offers several advantages. For example, MCDM enables a full assessment of several factors impacting safety climate, providing a more nuanced understanding than single-factor analyses. Furthermore, MCDM methodologies assist the prioritization of essential aspects, which guide resource allocation and intervention strategies. Furthermore, MCDM techniques give a structured and objective manner to evaluate options or analyze distinct groups in terms of their safety climate performance. Fuzzy extensions of MCDM approaches can successfully manage the ambiguity and vagueness inherent in safety climate data [17]. The Decision Making and Trial Evaluation Laboratory (DEMATEL) approach is considered as one of the most popular MCDM tools [18,19,20,21,22,23]. The method’s research output reflect its continuous significance and applicability in addressing advanced decision-making challenges. The DEMATEL approach is a structured technique for identifying and studying the intricate relationships of different aspects within a complicated system [24,25,26]. It assists in visualizing and understanding the cause-and-effect relationships between multiple factors [27,28,29,30]. Providing a clear and visual picture of the interrelationships between components makes complex systems easier to understand [31,32]. DEMATEL can also determine the most crucial factors that affect the system. This allows for specific strategies and efficient allocation of resources [33,34]. Fuzzy DEMATEL enhances the traditional DEMATEL approach by introducing fuzzy set theory to deal with the inherent vagueness and uncertainty in human judgments [35,36]. This approach enables decision-makers to express their thoughts using linguistic phrases, which are then translated into fuzzy numbers, usually triangular or trapezoidal [22,37]. This fuzzification enables a more realistic depiction of subjective assessments on the causal linkages between complicated components [38]. Fuzzy arithmetic procedures are then used to build the direct-relation matrix and compute the total-relation matrix, finally determining the effect and relation degrees of each factor. Using fuzzy logic, fuzzy DEMATEL improves the resilience and reliability of the analysis, particularly in cases when precise quantitative data is rare or incorrect.

Safety perception might vary among different organizational levels. For instance, Zohar and Luria [39] suggested a model that differentiates safety climate between organization-level and group-level. The significance of examining these different levels has been stressed in research in different sectors. For example, Lingard et al. [40] studied safety levels in multi-level organizational structure in a construction firm, and found considerable variances within workgroups in the same firm. Also, Sexton et al. [41] developed the Safety Attitudes Questionnaire, which assesses safety climate at both the unit and hospital levels. These multilevel methods have shown to be a more nuanced understanding of how safety perceptions are impacted within large organizations, emphasizing the necessity for specific interventions for each organizational levels to enhance the safety climate [42]. Measuring the safety climate is extremely essential from the standpoint of workers [43]. Workers are most directly exposed to occupational risks. Their perceptions of safety climate reflect the reality of their daily work experience. Their insights provide a realistic and practical view of how safety policies are implemented and enforced on the ground [39]. Workers are often the first to notice potential safety issues or gaps in safety procedures. Their feedback can provide an early warning system for preventing accidents [44]. It empowers them to actively contribute to improving safety in their workplace, rather than just being passive recipients of safety rules. When organizations genuinely seek worker input on safety climate, it fosters trust and open communication between workers and management. A positive safety climate makes workers feel psychologically safe to report safety concerns or near misses without fear of reprisal [45]. When workers are involved in safety climate assessments, they are more likely to support and adopt changes to safety practices.

Safety climate is considered a complex system that is affected by multiple factors. These factors might be related to each other. For example, workers’ compliance with safety procedures might be affected by workers’ competence in safety. Therefore, the ultimate goal of the current research is to use the fuzzy DEMATEL approach to find the interrelations among these factors from workers’ perspectives, because workers are regarded the first people to experience dangers during their everyday operations. DEMATEL technique assists in visualizing and understanding the cause-and-effect relationships between worker-related factors that affect safety climate. Moreover, DEMATEL can also determine the most crucial factors that affect the system. Moreover, to enhance the accuracy of the analysis, fuzzy DEMATEL will be used to enable a more realistic depiction of subjective assessments on the causal linkages between worker-related factors affecting safety climate. So, the questions of the current study are:

• From the workers’ perspective, what are the primary factors, i.e., causes, that influence the safety climate?
• From the workers’ perspective, what are the derived factors, or impacts, of the key sources affecting the safety climate?

Based on the literature, previous research has not explored how worker-related factors impact occupational safety climate. As a result, the primary contribution of this study is to use the fuzzy DEMATEL method to worker-related elements that influence safety climate to discover essential characteristics, causal links, and highlight actions to improve the level of safety climate.

In this paper, Section 1 serves as the introduction. Section 2 presents a comprehensive literature review on safety climate and MCDM methods. The methodology is described in Section 3, with particular attention on the fuzzy DEMATEL technique and its mathematical formulations. The findings are shown in Section 4, along with illustrations and interpretations of the fuzzy DEMATEL results. The primary cause and effect factors that have been found are examined in Section 5, along with their implications for safety climate management. Section 6 of the report wraps up by highlighting the key conclusions, discussing the limits, and suggesting future lines of inquiry for safety climate studies utilizing DEMATEL and possibly other complimentary methodologies.

2. Literature Review

Safety climate, a multidimensional construct representing employees’ shared perceptions of safety policies, procedures, and practices, has gained significant attention in contemporary occupational safety research. Recent studies continue to refine the understanding of the key dimensions that constitute effective safety climates across various industries and organizational context. This review synthesizes current literature on these dimensions, highlighting their significance and relationships with safety outcomes.

2.1. Key Dimensions of Safety Climate

The safety climate in workplaces is influenced by several key dimensions related to workers. These dimensions are crucial in shaping the entire safety environment and have a substantial impact on safety performance and accident occurrence. Workers’ self-perception of safety, participation in safety practices, and interactions with coworkers are critical. These dimensions contribute to the development of a strong safety climate by promoting proactive safety behaviors [46]. These dimensions ensure workers are well-informed and equipped to undertake safety procedures, therefore improving the safety climate [47].

The physical safety environment and availability of resources play a significant role in shaping the safety climate. Ensuring a safe working environment and providing necessary resources can improve safety perceptions among workers [48].

Low job satisfaction and poor social support are strongly correlated with a negative safety climate and higher accident occurrence. Improving these factors can enhance the safety climate and reduce accidents [49]. High pressure for production can negatively impact safety climate by increasing risk behaviors. Balancing production demands with safety priorities is essential to maintain a positive safety climate [50].

2.2. MCDM Tools in Safety Climate

To create a cause-and-effect diagram of interdependent factors, the DEMATEL approach is employed. Because this method reveals the linkages between criteria, ranks the criteria according to the type of relationships, and reveals the severity of their impacts on each criterion, it is superior to conventional techniques. An integrated approach is required to address the topic under consideration because a single method is insufficient to uncover connected factors under ambiguity and vagueness. For this reason, flexible information is handled and represented via fuzzy language modeling. Several research have used fuzzy MCDM models in occupational safety and health research, as summarized in

Table 1. Previous important studies that used MCDM tools in occupational safety and health field.

Author/s	Year	Goal	MCDM Method	Main Outcomes
Heravi et al. [51]	2022	Determine the crucial elements for maximizing safety features in large-scale building projects.	Fuzzy Delphi and Fuzzy AHP	The attitudes of workers toward safety and equipment have been identified as important determinants of safety in large-scale projects.
Alshehri et al. [52]	2022	Analyze the main risk factors and potential legal remedies for a sustainable supply chain.	Fuzzy AHP and Fuzzy WASPAS	In order to reduce the risks associated with a sustainable supply chain, management commitment at all levels is essential.
Marhavilas et al. [53]	2020	Prioritizes hazards in the sour crude oil processing business.	AHP and Fuzzy AHP	Government regulation, leadership attention, and dynamic supervision all influence safety levels.
Khan et al. [54]	2019	To identify significant hazards, their causes, and their consequences in building activities.	Fuzzy-TOPSIS	The primary causes of accidents are fire and chemical hazards.
Ilbahar et al. [55]	2018	Develop a novel approach to risk assessment for workplace safety and health.	Fuzzy AHP	Offer an effective, sophisticated approach to risk assessment.
Efe and KURT [56]	2017	Performing risk assessment for cutting techniques.	Fuzzy VIKOR and linear programing	work equipment and electric shock were the main causes of Dangers during the cutting procedure.
Basahel and Taylan [57]	2016	A strategy for identifying and analyzing the factors that affect safety in construction sectors.	Fuzzy AHP and TOPSIS	Safety performance is affected by the level of cooperation, safety norms, safe behavior, training, and safety leadership.
Ilangkumaran et al. [58]	2015	Evaluate worker safety in warm environments.	ANP and fuzzy approaches	Environment and job criteria play important roles in affecting safety of hot settings.
Fu and Chan [59]	2014	Assess the safety climate at the international airport.	Fuzzy Delphi method and fuzzy AHP	Safety supervision and an incentive-penalty system affect safety climate.
Zheng et al. [60]	2012	Develop a system for evaluating workers’ safety in hot and humid environments.	Fuzzy AHP	Provide a dependable early warning system for hot and humid weather.

. However, none of the previous studies used examined the interrelationships between the workers-related factors that affect safety climate.

3. Methodology

This paper employs the Fuzzy DEMATEL methodology to analyze safety climate factors. The process, illustrated in Figure 1, comprises several key steps. The initial step involved identifying relevant safety climate factors. These factors were determined through a comprehensive review of the existing literature on safety climate. Following the identification of factors, an expert panel was selected. This panel comprised individuals with extensive knowledge and experience in occupational safety. A meeting was conducted with the expert panel to provide a thorough explanation of the research topic and the objectives of the study. This meeting ensured that all experts had a clear understanding of the safety climate factors under consideration and the overall research goals. A linguistic assessment scale was developed to facilitate the evaluation of the relationships between the identified safety climate factors. This scale consisted of five linguistic terms, such as ‘No influence’, ‘Very Low influence’, ‘Low influence’, ‘High influence’, ‘Very high influence’. Each linguistic term was associated with a corresponding triangular fuzzy number (TFN) to capture the inherent uncertainty in expert judgments as presented in

Table 2. Fuzzy linguistic term scale.

	Linguistic Terms	Fuzzy Triangular Number
		L	M	U
1	No influence	0	0	0.25
2	Very low influence	0	0.25	0.5
3	Low influence	0.25	0.5	0.75
4	High influence	0.5	0.75	1
5	Very high influence	0.75	1	1

. Prior to the evaluation process, the experts received detailed training on how to use the evaluation form. This training included guidance on interpreting the linguistic terms, applying the fuzzy linguistic scale, and completing the evaluation form accurately. The expert panel then proceeded to evaluate the pairwise relationships between the safety climate factors using the developed linguistic assessment scale. Each expert provided their judgment on the direct influence of each factor on every other factor. Following the expert evaluations, a consistency check was performed to assess the reliability and agreement of the experts’ judgments. The Fuzzy DEMATEL method was then applied to the aggregated expert evaluations to determine causal and effect factors and construct the diagram.

3.1. Key Worker-Related Factors Affecting Safety Climate

The worker-related factors that affect safety climate were selected drawing on previous research that documented important factors or criteria linked to improving the safety climate level from an employee points of view. The first factor was “Workers’ compliance with safety procedures”, which mean how workers follow safety regulations. Multiple studies demonstrate that a positive safety climate significantly predicts and promotes workers’ compliance with safety procedures [43,61]. The second factor was “Workers promptly report accident”, which means that workers directly inform the supervisor or top managers about accidents even it is a near-miss incidents. Workplaces with a poor safety climate or inconsistent supervisor enforcement have significantly higher rates of accident underreporting [62,63]. Moreover, A positive safety climate, characterized by open communication and prompt reporting, is associated with fewer workplace accidents. Conversely, more safety climate problems (such as lack of reporting) are linked to higher accident rates [63,64]. The third factor was, “Workers’ safety competence”, which indicates how workers are competent with respect to occupational safety and health. Multiple studies identify worker competence as a critical factor in enhancing the safety climate. Additionally, Effective training and procedures, which build worker competence, are shown to improve perceptions of safety climate and the effectiveness of safety systems [65]. The fourth factor was “Workers promptly find a solution to improve safety”, which indicates that workers once they face a hazard they immediately try to search for solutions based on their safety competence. For superintendents and skilled laborers, increasing worker competence is among the most significant ways to improve the recognizing hazards and counter them [66,67]. The fifth factor was “Peer influence” which means how co-workers attitude and behavior affect their peers in terms of safety. Co-worker support for safety acts as a powerful intermediary, often exerting a greater impact on safety compliance and participation than supervisory support. This effect is consistent across both individual and team levels, highlighting the unique and potent role of peer influence in shaping safety outcomes [68,69]. The presence of a strong co-worker safety compliance—where peers actively value and prioritize safety—significantly boosts safety behaviors and participation, sometimes more so than the influence of supervisors [70]. The sixth factor was “Workers’ discussion after a risk occurred”, which indicates the extent to which workers engage in conversations and communication with each other following an incident, near miss, or the identification of a potential hazard. Studies show that interventions aimed at improving safety climate almost always involve enhancing communication within organizations. These interventions, which include open discussions about safety and health, lead to significant improvements in safety climate and related outcomes such as teamwork, safety behavior, and safety performance A previous study showed that interventions aimed at improving safety climate almost always involve enhancing communication within organizations. These interventions, which include open discussions about safety and health, lead to significant improvements in safety climate and related outcomes such as teamwork, safety behavior, and safety performance [71]. The seventh factor was “Workers’ freedom speech about safety matters”, which indicate extent to which employees feel comfortable, safe, and empowered to openly communicate their concerns, ideas, suggestions, and opinions related to workplace safety and health without fear of reprisal, negative consequences, or being ignored. Improving communication through daily feedback sessions significantly improves safety climate, and workers’ behavior [72]. The eighth factor was “Worker’s ability to perceive hazards”, which indicates individual capacity of employees to identify, recognize, and understand potential sources of harm or danger present in their work environment. This ability is crucial for preventing accidents, injuries, and illnesses. When workers perceive risks as serious, they are more likely to engage in proactive safety behaviors which enhance the level of safety climate [73]. Moreover, to enhance the validity of the selected factors, the selection of these specific factors was guided by the Nordic Occupational Safety Climate Questionnaire (NOSACQ-50), a well-validated and widely used instrument designed to measure safety climate [64]. The NOSACQ-50 comprises 50 items across seven dimensions of safety climate. The factors that directly corresponded to key aspects of worker behavior, attitudes, and perceptions relevant to safety were selected. As mentioned earlier, none of the previous research studies have studied the relationship between these factors. Drawing on previous research that documented important factors or criteria linked to improving the safety climate level from an employee points of view, these factors are summarized in

Table 3. Factors affecting the level of safety climates from an employee perspective.

Code	Criteria
W1	Workers’ compliance with safety procedures
W2	Workers promptly report accident
W3	Workers’ safety competence
W4	Workers promptly find a solution to improve safety
W5	Peer influence
W6	Workers’ discussion after a risk occurred
W7	Workers’ freedom speech about safety matters
W8	Worker’s ability to perceive hazards

.

3.2. Experts Selection

The methodology for data collecting in DEMATEL analysis used a systematic approach to get expert assessments on the interrelationship among components. A carefully selected panel of nine subject matter experts, each with sufficient level of experience in occupational safety and health domains, was used in this analysis, as delineated in

Table 4. Profiles of the Occupational safety and health (OSH) experts recruited in the study.

Expert #	Qualification	Field	Experience (Yrs)
1	Ph.D.	OSH Training Center	19
2	Ph.D.	Industrial Safety Engineering	7
3	M.Sc.	OSH Engineer	22
4	B.Sc.	OSH supervisor	15
5	B.Sc.	OSH supervisor	19
6	B.Sc.	OSH supervisor	8
7	B.Sc.	Safety specialist	13
8	Diploma	Safety assistant	20
9	Diploma	Safety assistant	18

below. These specialists were presented with a comprehensive inventory of 8 worker-related factors deemed pivotal to the safety climate within industrial environments.

Each expert was then given the responsibility of creating a pairwise comparison matrix in which they assessed how each factor affected the others. As stated in Table 1 of this study, this evaluation used a preset scale that ranged from 1 (showing no influence) to 5 (indicating a very high influence). The experts were given thorough instructions explaining how to interpret and use the rating scale in order to guarantee uniformity and accuracy in the assessment process. When creating these assessments, the experts were urged to draw on their knowledge and professional expertise. Note, employing a purposive sampling strategy, this study deliberately selected experts based on their specialized knowledge and extensive experience within the occupational safety and health (OSH) domain, rather than relying on random selection. It is important to acknowledge that while the number of experts might appear limited, this is a common and accepted practice within multi-criteria decision-making studies that utilize employee expert evaluations in this field. Prior research consistently demonstrates that methodological validity in such contexts is not fundamentally determined by a high volume of participants. Instead, the critical factor lies in the strategic recruitment of individuals who offer significant empirical depth and comprehensive conceptual understanding of the subject matter under investigation [74,75,76,77,78,79,80,81].

3.3. Fuzzy DEMATEL Analysis

In the realm of multi-criteria decision-making (MCDM), particularly in examining safety climate from an employee perspective, several challenges arise. These include effectively weighing various criteria, accurately assessing alternatives against these criteria, and consolidating the subjective ratings provided by decision-makers.

When a safety climate assessment lacks clearly defined parameters, it can lead to uncertainty among those evaluating the workplace environment. This ambiguity, often appearing from incomplete or imprecise information about safety factors, can be effectively addressed using fuzzy set theory. This method is particularly valuable in capturing the nuanced perceptions of employees regarding safety climate. The basis of the DEMATEL method consists of the following steps [37,82,83,84,85]:

Step 1: The structural relationships between a problem’s constituent elements, including their respective degrees of influence, are established according to the linguistic term scale values shown in Table 2. Subsequently, the influential factors within this complex system are identified and defined through a process of information gathering, which may include reviews of existing literature, collaborative brainstorming sessions, and/or consultations with subject matter experts.

Step 2: Construct the fuzzy direct relationship matrix (Z) after collecting the required data from the experts as it is presented in Equation (1).

$$Z=\left[\begin{array}{cccc}0& {\tilde{z}}_{21}& \dots & {\tilde{z}}_{n1}\\ {\tilde{z}}_{12}& 0& \dots & {\tilde{z}}_{n2}\\ ⋮& ⋮& ⋮& ⋮\\ {\tilde{z}}_{1n}& {\tilde{z}}_{2n}& \dots & 0\end{array}\right]$$

(1)

Step 3: The direct influence matrix is then normalized using Equation (2), which results in a standardized matrix of relationships.

$${\tilde{x}}_{ij}={\displaystyle \frac{{\tilde{z}}_{ij}}{r}}=({\displaystyle \frac{l_{ij}}{r}}, {\displaystyle \frac{m_{ij}}{r}}, {\displaystyle \frac{u_{ij}}{r}})$$

(2)

where

$$r={max}_{i,j}\left\{{max}_i{\sum }_{j=1}^n{u}_{ij},{max}_j{\sum }_{i=1}^n{u}_{ij}\right\} i, j\in \left\{1, 2, 3, \cdots , n\right\}$$

Step 4: The fuzzy total relational matrix, denoted as $\tilde{T}$, is derived using the computational method outlined in Formula (3).

$$\tilde{T}=\underset{k\to +\infty }{\mathit{lim}}({\tilde{x}}^1 ⨁ {\tilde{x}}^2⨁ \dots ⨁ {\tilde{x}}^k)$$

(3)

If each element of the fuzzy total relation matrix is depicted as ${\tilde{t}}_{ij}=\left(l_{ij}^{″},m_{ij}^{″},u_{ij}^{″}\right)$ then it can be calculated as expressed in Equations (4)–(6).

$$\left[l_{ij}^{″}\right]=x_l\times {(l-x_l)}^{-1}$$

(4)

$$\left[m_{ij}^{″}\right]=x_m\times {(l-x_m)}^{-1}$$

(5)

$$\left[u_{ij}^{″}\right]=x_u\times {(l-u)}^{-1}$$

(6)

Step 5: Defuzzify the values into crisp values to obtain crisp value of total relation matrix (T) using Equation (7). The upper and lower bounds of normalized values is calculated using Equations (8) and (9).

$$x_{ij}={\displaystyle \frac{[l_{ij}^s\left(1-l_{ij}^s\right)+u_{ij}^s\times u_{ij}^s]}{[1-l_{ij}^s+u_{ij}^s]}}$$

(7)

where

$$l_{ij}^n={\displaystyle \frac{\left(l_{ij}^t-\min l_{ij}^t\right)}{{\Delta }_{min}^{max}}}$$

$$m_{ij}^n={\displaystyle \frac{(m_{ij}^t-min{l}_{ij}^t)}{{\Delta }_{min}^{max}}}$$

$$u_{ij}^n={\displaystyle \frac{(u_{ij}^t-min{l}_{ij}^t)}{{\Delta }_{min}^{max}}}$$

$${\Delta }_{min}^{max}=\max u_{ij}^t-\min l_{ij}^t$$

$$l_{ij}^s={\displaystyle \raisebox{1ex}{$m_{ij}^n$}\!\left/ \!\raisebox{-1ex}{$(1+m_{ij}^n-l_{ij}^n)$}\right.}$$

(8)

$$u_{ij}^s={\displaystyle \raisebox{1ex}{$u_{ij}^n$}\!\left/ \!\raisebox{-1ex}{$(1+u_{ij}^n-l_{ij}^n)$}\right.}$$

(9)

Step 6: Set the value of threshold which is obtained by finding the average values of the matrix T using Equation (10) where n is the total number of factors included in matrix T.

$$\theta ={\displaystyle \frac{{\sum }_{i=1}^n{\sum }_{j=1}^n\left[t_{ij}\right]}{n^2}}$$

(10)

where n is the number of criteria, and $t_{ij}$ represents elements of the crisp total relation matrix. The threshold value in DEMATEL helps filter out weak relationships, highlighting only the most significant influences among factors. This simplifies complex interactions, making causal analysis clearer and more actionable. By setting a threshold, decision-makers can focus on key drivers, visualize dominant connections, and improve strategic insights without unnecessary noise.

Step 7: Within the aggregate relational matrix (T), the values within each row and column are summed. These summations are represented by $D_i$ for the ith row and $R_j$ for the jth column. $D_i$ and $R_j$ serve as indicators of the total influence—both direct and indirect—exerted and received by each factor, respectively. $D_i$ and $R_j$ equations are presented in Equation (11) and Equation (12), respectively.

$$D_i={\sum }_{j=1}^n{t}_{ij} (i=1, 2, \dots , n)$$

(11)

$$R_j={\sum }_{i=1}^n{t}_{ij} (j=1, 2, \dots , n)$$

(12)

Step 8: Ordering the relative importance of each factor was determined by calculating the weight $w_j$ using its prominence and relation values and then normalize and rank the factors as shown below in Equations (13) and (14). Subsequently, a causal figure is constructed. The horizontal axis, representing “prominence”, is derived by summing R and D (D + R), thus indicating the relative importance of each criterion. Moreover, the vertical axis, termed “relation”, is generated by subtracting R from D (D − R), illustrating the net influence. Criteria exhibiting a negative (D − R) value are classified as belonging to the “effect” group, signifying their susceptibility to influence from other criteria. Conversely, a positive (D − R) value denotes a substantial impact exerted by the criterion, suggesting prioritization in improvement efforts.

$$w_j =\sqrt{{(R_i+D_j)}^2+{(R_i-D_j)}^2}$$

(13)

$$w_j^{\prime }={\displaystyle \frac{w_j}{{\sum }_{j=1}^n{w}_j}}$$

(14)

3.4. Reliability Evaluation

Internal consistency is one of the most often used methods to measure the reliability of a questionnaire [86]. One of the methods used to assess internal consistency is the item-total correlation method, which is the correlation between an individual item and the total score without that particular item. The corrected item-total correlation test is performed to check if any item in the set of the questionnaire is inconsistent with the averaged behavior of the others. Formula XX was used to find the corrected item-total correlation [87]. For evaluating internal consistency of the questionnaire, corrected item-total correlation is used with the criterion of 0.30 for being considered as an acceptable corrected item-total correlation [88].

$$Corrected item-total correlation=Corr (x_i^{\left(a\right)} , {\sum }_{i=1}^N{X}_j-x_i^{\left(a\right)})$$

(15)

where,

${\sum }_{i=1}^N{X}_j:$ Sum of experts’ responses in question i.

$x_i^{\left(a\right)}$: Responses of expert a.

Moreover, Chen et al. [89] used the consistency matrix to measure the agreement between experts in the DEMATEL evaluation. Equation (16) was used to find the consistency matrix regarding experts’ evaluation. When rating the influence of one factor on another, if two experts give equal ratings, then the two experts’ agreement degree reached the maximum value, i.e., equals to 1. On the other hand, when the difference between the ratings is the largest, the two experts’ agreement degree reach the minimum value, i.e., equal to 0.

$${\sigma }_{ab}=1-{\displaystyle \frac{|x_{ij}^{\left(a\right)}-x_{ij}^{\left(b\right)}|}{T_N}}$$

(16)

where,

${\sigma }_{ab}$: consistency between expert a, and b.

$x_{ij}^{\left(a\right)}$, $x_{ij}^{\left(b\right)}$: Evaluation of influence of factor i on factor j by expert a, and b, respectively.

$T_N$: difference between the maximum and minimum values in the DEMATEL evaluation scale, and in scales {1,2,3,4,5} of this paper, $T_N$= 4.

4. Analysis and Results

Upon completion of individual evaluations by all experts, the discrete matrices were aggregated using the Fuzzy Geometric Mean to generate a single fuzzy direct-relation matrix, ensuring a more precise representation of the experts’ collective judgments. The consolidation method serves to mitigate individual predispositions and construct a balanced representation of the perceived interconnections among safety climate variables from the employee standpoint. The resultant evaluation matrix is shown in

Table 5. Aggregated fuzzy direct-relation matrix using the fuzzy geometric mean approach.

	W1	W2	W3	W4	W5	W6	W7	W8
W1	(0.000, 0.000, 0.000)	(0.639, 0.889, 1.000)	(0.139, 0.389, 0.639)	(0.611, 0.861, 1.000)	(0.667, 0.917, 1.000)	(0.694, 0.944, 1.000)	(0.028, 0.278, 0.528)	(0.667, 0.917, 1.000)
W2	(0.667, 0.917, 1.000)	(0.000, 0.000, 0.000)	(0.583, 0.833, 1.000)	(0.611, 0.861, 1.000)	(0.639, 0.889, 1.000)	(0.611, 0.861, 1.000)	(0.111, 0.361, 0.611)	(0.694, 0.944, 1.000)
W3	(0.667, 0.917, 1.000)	(0.139, 0.389, 0.639)	(0.000, 0.000, 0.000)	(0.694, 0.944, 1.000)	(0.611, 0.861, 1.000)	(0.639, 0.889, 1.000)	(0.111, 0.361, 0.611)	(0.611, 0.861, 1.000)
W4	(0.611, 0.861, 1.000)	(0.639, 0.889, 1.000)	(0.667, 0.917, 1.000)	(0.000, 0.000, 0.000)	(0.639, 0.889, 1.000)	(0.167, 0.389, 0.639)	(0.056, 0.222, 0.472)	(0.583, 0.833, 1.000)
W5	(0.694, 0.944, 1.000)	(0.639, 0.889, 1.000)	(0.028, 0.083, 0.333)	(0.556, 0.806, 1.000)	(0.000, 0.000, 0.000)	(0.611, 0.861, 1.000)	(0.611, 0.861, 1.000)	(0.111, 0.361, 0.611)
W6	(0.111, 0.361, 0.611)	(0.583, 0.833, 1.000)	(0.556, 0.806, 1.000)	(0.639, 0.889, 1.000)	(0.611, 0.861, 1.000)	(0.000, 0.000, 0.000)	(0.583, 0.833, 1.000)	(0.528, 0.778, 1.000)
W7	(0.111, 0.361, 0.611)	(0.667, 0.917, 1.000)	(0.000, 0.111, 0.361)	(0.139, 0.389, 0.639)	(0.583, 0.833, 1.000)	(0.639, 0.889, 1.000)	(0.000, 0.000, 0.000)	(0.083, 0.306, 0.556)
W8	(0.639, 0.889, 1.000)	(0.694, 0.944, 1.000)	(0.583, 0.833, 1.000)	(0.556, 0.806, 1.000)	(0.583, 0.833, 1.000)	(0.556, 0.806, 1.000)	(0.056, 0.278, 0.528)	(0.000, 0.000, 0.000)

.

This matrix underwent a normalization process utilizing the equations outlined in step 3. The resulting normalized fuzzy relationship matrix is shown in

Table 6. Normalized fuzzy relationship matrix.

	W1	W2	W3	W4	W5	W6	W7	W8
W1	(0.000, 0.000, 0.000)	(0.091, 0.127, 0.143)	(0.020, 0.056, 0.091)	(0.087, 0.123, 0.143)	(0.095, 0.131, 0.143)	(0.099, 0.135, 0.143)	(0.004, 0.040, 0.075)	(0.095, 0.131, 0.143)
W2	(0.095, 0.131, 0.143)	(0.000, 0.000, 0.000)	(0.083, 0.119, 0.143)	(0.087, 0.123, 0.143)	(0.091, 0.127, 0.143)	(0.087, 0.123, 0.143)	(0.016, 0.052, 0.087)	(0.099, 0.135, 0.143)
W3	(0.095, 0.131, 0.143)	(0.020, 0.056, 0.091)	(0.000, 0.000, 0.000)	(0.099, 0.135, 0.143)	(0.087, 0.123, 0.143)	(0.091, 0.127, 0.143)	(0.016, 0.052, 0.087)	(0.087, 0.123, 0.143)
W4	(0.087, 0.123, 0.143)	(0.091, 0.127, 0.143)	(0.095, 0.131, 0.143)	(0.000, 0.000, 0.000)	(0.091, 0.127, 0.143)	(0.024, 0.056, 0.091)	(0.008, 0.032, 0.067)	(0.083, 0.119, 0.143)
W5	(0.099, 0.135, 0.143)	(0.091, 0.127, 0.143)	(0.004, 0.012, 0.048)	(0.079, 0.115, 0.143)	(0.000, 0.000, 0.000)	(0.087, 0.123, 0.143)	(0.087, 0.123, 0.143)	(0.016, 0.052, 0.087)
W6	(0.016, 0.052, 0.087)	(0.083, 0.119, 0.143)	(0.079, 0.115, 0.143)	(0.091, 0.127, 0.143)	(0.087, 0.123, 0.143)	(0.000, 0.000, 0.000)	(0.083, 0.119, 0.143)	(0.075, 0.111, 0.143)
W7	(0.016, 0.052, 0.087)	(0.095, 0.131, 0.143)	(0.000, 0.016, 0.052)	(0.020, 0.056, 0.091)	(0.083, 0.119, 0.143)	(0.091, 0.127, 0.143)	(0.000, 0.000, 0.000)	(0.012, 0.044, 0.079)
W8	(0.091, 0.127, 0.143)	(0.099, 0.135, 0.143)	(0.083, 0.119, 0.143)	(0.079, 0.115, 0.143)	(0.083, 0.119, 0.143)	(0.079, 0.115, 0.143)	(0.008, 0.040, 0.075)	(0.000, 0.000, 0.000)

. The fuzzy total relation matrix, derived from the normalized fuzzy matrix, captures indirect influences and provides a comprehensive view of factor interactions. It reveals hidden relationships and the full impact of each factor on others the fuzzy total relation matrix is illustrated in

Table 7. The fuzzy total-relation matrix.

	W1	W2	W3	W4	W5	W6	W7	W8
W1	(0.070, 0.275, 0.883)	(0.161, 0.408, 1.057)	(0.076, 0.271, 0.858)	(0.156, 0.399, 1.061)	(0.167, 0.423, 1.096)	(0.161, 0.401, 1.048)	(0.039, 0.213, 0.762)	(0.153, 0.379, 1.000)
W2	(0.166, 0.413, 1.062)	(0.083, 0.316, 0.987)	(0.136, 0.339, 0.946)	(0.165, 0.422, 1.118)	(0.174, 0.443, 1.155)	(0.160, 0.415, 1.105)	(0.051, 0.235, 0.814)	(0.164, 0.403, 1.054)
W3	(0.156, 0.389, 1.014)	(0.097, 0.347, 1.023)	(0.052, 0.214, 0.778)	(0.166, 0.407, 1.068)	(0.160, 0.415, 1.103)	(0.154, 0.394, 1.055)	(0.048, 0.222, 0.777)	(0.145, 0.370, 1.006)
W4	(0.153, 0.380, 1.002)	(0.155, 0.395, 1.047)	(0.138, 0.325, 0.891)	(0.074, 0.281, 0.928)	(0.162, 0.409, 1.087)	(0.096, 0.329, 1.000)	(0.037, 0.197, 0.748)	(0.142, 0.362, 0.993)
W5	(0.149, 0.364, 0.966)	(0.156, 0.384, 1.018)	(0.051, 0.209, 0.786)	(0.138, 0.362, 1.018)	(0.074, 0.280, 0.929)	(0.146, 0.368, 1.008)	(0.111, 0.269, 0.788)	(0.075, 0.287, 0.916)
W6	(0.087, 0.327, 1.010)	(0.152, 0.402, 1.106)	(0.126, 0.319, 0.938)	(0.157, 0.401, 1.109)	(0.161, 0.418, 1.148)	(0.072, 0.285, 0.973)	(0.110, 0.282, 0.854)	(0.132, 0.361, 1.045)
W7	(0.058, 0.249, 0.826)	(0.138, 0.336, 0.915)	(0.034, 0.176, 0.704)	(0.067, 0.265, 0.875)	(0.128, 0.333, 0.947)	(0.132, 0.324, 0.908)	(0.026, 0.133, 0.589)	(0.052, 0.234, 0.813)
W8	(0.159, 0.399, 1.053)	(0.169, 0.421, 1.103)	(0.134, 0.331, 0.939)	(0.154, 0.403, 1.109)	(0.162, 0.424, 1.145)	(0.149, 0.396, 1.095)	(0.042, 0.217, 0.797)	(0.071, 0.273, 0.920)

.

Converting to crisp values as presented in

Table 8. The crisp total-relation matrix.

	W1	W2	W3	W4	W5	W6	W7	W8
W1	0.366	0.491	0.355	0.486	0.509	0.485	0.295	0.46
W2	0.492	0.414	0.419	0.509	0.531	0.502	0.318	0.484
W3	0.469	0.441	0.303	0.493	0.505	0.481	0.304	0.455
W4	0.462	0.481	0.402	0.38	0.499	0.425	0.282	0.447
W5	0.446	0.470	0.301	0.454	0.381	0.457	0.343	0.38
W6	0.421	0.494	0.404	0.494	0.513	0.39	0.361	0.452
W7	0.339	0.421	0.262	0.36	0.425	0.411	0.21	0.325
W8	0.481	0.507	0.412	0.495	0.516	0.488	0.303	0.371

Notice: Bold values indicate numbers greater than or equal to the value of threshold $\theta$ of 0.422.

simplifies interpretation and decision-making, allowing for precise ranking and prioritization of factors. This transformation enables clear visualization of cause-effect relationships and facilitates the identification of critical elements within the system under study. The DEMATEL methodology employs crucial calculations to derive prominence, relation, weight, and final criteria weight. These metrics elucidate the complex interplay among factors in a system. Prominence quantifies a criterion’s overall influence, while relation distinguishes between cause-and-effect elements.

Weights reflect the relative importance of each factor. The final criteria weight, synthesizing these measures, provides a comprehensive assessment of each element’s role, and those values are presented in

Table 9. Final results of DEMATEL analysis.

	${\mathit{R}}_{\mathit{i}}$	${\mathit{D}}_{\mathit{i}}$	${\mathit{D}}_{\mathit{i}}$$+{\mathit{R}}_{\mathit{i}}$	${\mathit{D}}_{\mathit{i}}$$-{\mathit{R}}_{\mathit{i}}$	${\mathbf{W}}_{\mathit{j}}^{\mathbf{\prime }}$	Rank	Identity
W1	3.476	3.447	6.924	−0.029	0.130	6	Effect
W2	3.719	3.668	7.387	−0.051	0.140	1	Effect
W3	2.857	3.45	6.307	0.593	0.117	7	Cause
W4	3.671	3.377	7.048	−0.294	0.130	4	Effect
W5	3.879	3.232	7.111	−0.648	0.131	3	Effect
W6	3.638	3.529	7.167	−0.109	0.132	2	Effect
W7	2.416	2.753	5.168	0.337	0.095	8	Cause
W8	3.372	3.573	6.945	0.201	0.128	5	Cause

Note, bolded numbers exceed the threshold.

. This multifaceted analysis facilitates informed decision-making by revealing intricate system dynamics and prioritizing influential factors.

DEMATEL’s visual cause-effect mapping provides a graphical representation of complex system interactions. This diagram illustrates the interdependencies among factors, highlighting influential elements and those most affected. The map’s benefit lies in its ability to simplify intricate relationships, facilitating strategic decision-making by identifying key drivers and potential intervention points within the system. In this study, Figure 2 shows the cause-and-effect factors of safety climate according to an employee point of view.

The value of the corrected item-total correlation between the experts and the behaviors of the others are shown in

Table 10. Corrected item—total correlation between experts’ Reponses.

Between Expert i and the Remaining	Corrected Item—Total Correlation
E1 & other	0.886
E2 & other	0.925
E3 & other	0.889
E4 & other	0.886
E5 & other	0.879
E6 & other	0.877
E7 & other	0.874
E8 & other	0.878
E9 & other	0.844

, which indicates suitable consistency level. Moreover,

Table 11. Average consistency value for all experts in each pairwise comparison.

	W1	W2	W3	W4	W5	W6	W7	W8
W1	1.000	0.836	0.849	0.883	0.857	0.892	0.915	0.842
W2	0.904	1.000	0.861	0.852	0.849	0.836	0.884	0.920
W3	0.893	0.850	1.000	0.937	0.836	0.891	0.881	0.883
W4	0.883	0.842	0.890	1.000	0.877	0.887	0.827	0.861
W5	0.926	0.849	0.776	0.926	1.000	0.883	0.850	0.851
W6	0.833	0.860	0.870	0.883	0.834	1.000	0.844	0.946
W7	0.884	0.842	0.889	0.892	0.848	0.840	1.000	0.826
W8	0.834	0.870	0.896	0.932	0.909	0.920	0.877	1.000

shows the average consistency values for all experts in each pairwise comparison. It is worth noting that the minimum consistency value obtained was 0.776, which indicates that the experts reach sufficient level of consensus in the pairwise comparisons.

5. Discussion

This study provides a comprehensive analysis of workplace safety climate, highlighting the complex interactions among key factors that shape employee behavior and overall safety outcomes. The findings from the DEMATEL analysis reveal distinct causal and effect relationships, offering valuable insights into how different aspects of workplace safety influence one another.

Among the identified factors, Workers’ safety competence (W3) stands out as the most significant causative element, demonstrating strong influence on multiple dimensions of workplace safety. The crisp total relation matrix indicates that W3 primarily impacts Peer influence (W5), followed by Workers promptly finding a solution to improve safety (W4), Workers’ discussion after a risk occurred (W6), Workers’ compliance with safety procedures (W1), Workers promptly reporting accidents (W2) and Workers’ ability to perceive hazards (W8). This strong influence underscores the critical role of safety knowledge and competence in fostering an environment where employees proactively engage in risk mitigation and compliance. This emphasizes how crucial ongoing safety training and skill improvement are. A positive safety climate is greatly influenced by worker safety competency, which includes the knowledge and abilities required for safe work practices [66].

Competent employees know how important it is to report accidents as soon as possible since it helps to prevent such situations in the future. Their knowledge also encourages others to follow safety procedures [90]. This proficiency facilitates knowledgeable safety conversations, allowing employees to offer insightful opinions and solutions [91]. Additionally, when employees get the rationale behind safety protocols, they are more inclined to adhere to them, which improves compliance [92]. In the end, a skilled staff transforms safety from a collection of regulations to a shared duty by fostering a culture where knowledge and experience drive safety.

Interestingly, “Workers’ freedom speech about safety matters” (W7) showed a more focused effect, primarily influencing “Peer influence” (W5), reinforcing the role of open communication in shaping collective workplace behaviors. The crisp total relation matrix indicates that W7 significantly affects W5, suggesting that when employees feel encouraged to voice their concerns without fear of retaliation, they actively contribute to a culture where safety discussions become normalized. This transparent communication environment fosters peer accountability, ensuring that safe behaviors are not only followed individually but are also reinforced through group dynamics. Research supports this notion, emphasizing that workplace transparency enhances peer-driven adherence to safety protocols [93], facilitates positive reinforcement of safe behaviors [94], and increases employee engagement in risk discussions [95]. Furthermore, studies highlight that fostering open dialogue about safety builds trust between workers, enabling stronger peer-to-peer influence [96], and ultimately improves safety climate perceptions, ensuring more effective hazard identification and risk mitigation [97]. These findings underscore the importance of organizational efforts to enhance worker communication, as it directly impacts peer interactions, influencing workplace safety outcomes in a profound and measurable way.

Moreover, “Worker’s ability to perceive hazards” (W8) exhibited diverse effects, most notably on W2, W5, W1, W4, and W6. A favorable safety climate is mostly driven by workers’ perceptions of hazards, which impact many safety-related activities. People with strong hazard perception skills are more likely to report incidents quickly, identifying possible threats and near-misses that could turn into major concerns down the road [98]. This understanding of potential consequences also strengthens workers’ compliance with safety procedures, making adherence a conscious and meaningful choice [99]. Additionally, personnel who perceive hazards well are able to quickly determine the underlying reasons and create safety solutions, which promotes proactive prevention [88]. Finally, It enhances safety conversations by allowing employees to offer pertinent information and identify possible hazards, which makes them more fruitful and efficient [100]. Ultimately, Strong hazard perception encourages employees to take initiative and be involved, creating a culture where everyone prioritizes safety. This highlights how important hazard perception abilities are in influencing different safety attitudes and behaviors. Knowing these complex relationships offers important direction for creating focused interventions and all-encompassing safety plans that can take advantage of the cascading effects of addressing important contributing elements.

Furthermore, the analysis reveals “Peer influence” W5 as a highly dependent effect factor, with an R − D score of −0.648, the most negative value in the analysis. This indicates that W5 is strongly shaped by external influences rather than exerting significant causative impact on other factors. A negative R − D value suggests that Peer Influence does not independently drive workplace safety behaviors but rather responds to the safety climate shaped by other elements. In this case, W5 is highly influenced by variables like workers’ compliance and accident reporting.

Additionally, W4 is ranked fourth and functions primarily as a receiver rather than a driver within the safety ecosystem, as indicated by its negative D-R value of −0.294. This positioning suggests that workers’ ability to develop safety solutions is significantly influenced by various workplace factors rather than being an independent catalyst itself. Interestingly, “Peer influence” (W5) demonstrates the strongest connection with solution-finding behaviors, with a relationship value of 0.499. This highlights how the social fabric of the workplace substantially shapes innovation in safety practices. When colleagues demonstrate safety consciousness, individual workers appear more motivated and empowered to develop safety improvements themselves. The analysis also reveals meaningful connections between solution-finding behaviors and workers’ compliance with safety protocols (0.462), their willingness to report accidents promptly (0.481), engagement in post-incident discussions (0.425), and their fundamental ability to recognize hazards (0.447). These interconnections paint a picture of safety improvement as deeply embedded within a broader cultural and behavioral context. These findings suggest that organizations aiming to enhance workers’ problem-solving capabilities regarding safety should focus on cultivating a supportive social environment rather than merely targeting individual skills. By strengthening reporting systems, encouraging open discussions after incidents, and developing hazard recognition competencies, organizations can create conditions where safety solutions emerge more naturally from the workforce.

6. Conclusions

The Fuzzy DEMATEL methodology was used in this study to examine the key elements affecting safety atmosphere from the viewpoint of the employees. The research involved nine experts evaluating eight critical factors, leveraging the Fuzzy DEMATEL approach to address the inherent uncertainty in expert opinions when dealing with complex, interrelated factors in workplace safety.

Clarifying the complex interactions between safety climate factors was made possible by the use of fuzzy DEMATEL. Fuzzy logic made it easier to accommodate the subjectivity and ambiguity present in expert opinions, especially when evaluating employee viewpoints on safety. This methodological technique improved the validity and relevance of the research findings by producing a more realistic and robust portrayal of the intricate dynamics within workplace safety climate.

The study revealed vital insights into the causal-effect linkages between safety climate elements. Three factors were identified as primary drivers (causal factors): “Workers’ safety competence” (W3), “Workers’ freedom speech about safety matters” (W7), and “Worker’s ability to perceive hazards” (W8). These causal factors have a major influence on other dimensions of safety climate, implying that focused improvements in these areas could deliver substantial advantages throughout the full safety spectrum. Conversely, the factors identified as effect factors—including “Workers’ compliance with safety procedures” (W1), “Workers promptly report accident” (W2), “Workers promptly find a solution to improve safety” (W4), “Workers’ discussion after a risk occurred” (W6) and “Peer influence” (W5)—are more likely to be influenced by changes in the causal factors. This distinction helps firms prioritize their safety improvement initiatives and resource allocation.

The outcomes of this study provide various strategic implications for firms aiming to improve their safety climate:

• Provide thorough safety training to improve workers’ safety competence and hazard perception abilities.
• Foster an open communication atmosphere where staff can raise safety concerns without fear of repercussions.
• Utilize peer influence to shape safety behaviors and attitudes, and promote peer-led initiatives.

While this study provides valuable insights, it is important to acknowledge its limitations. The Fuzzy DEMATEL technique relies on nine experts, which may not reflect the complete spectrum of opinions within varied industrial settings. Future studies could benefit from broadening the expert pool and even including cross-industry comparisons or comparisons between different countries. Moreover, to gain a deeper understanding of the complex factors that shape safety climate and develop more effective strategies, future studies might employ machine learning techniques to predict the level of safety climate based on workers’ behavior in an organization. Prediction of safety climate level is considered as one of the most important proactive approaches that will prevent the occurrence of future incidents of fatalities or injuries. Additionally, while the study focused on employee opinions, future research might address the alignment or discrepancy between employee and management views on safety climate elements. This could lead to a more comprehensive knowledge of firms’ safety climate.

In conclusion, this study’s use of Fuzzy DEMATEL to investigate safety climate parameters from an employee perspective gave useful insights into the complicated dynamics of workplace safety. By identifying key causal factors and their interactions, the research provides a roadmap for organizations to enhance their safety climate strategically. The findings emphasize the importance of a holistic approach to safety management, where compliance, competence, communication, and proactive problem-solving are integrated into a cohesive safety strategy.