{"paper_id":"05608dd7-e492-474d-9f4f-bb0e7c4cedc4","body_text":"Developing the Contractor Health, Safety, and Risk Performance Index: A Hybrid ANP–Machine Learning Approach | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Developing the Contractor Health, Safety, and Risk Performance Index: A Hybrid ANP–Machine Learning Approach Payam Khordoustan, Omid Akbarzadeh, Neda Gilani, Parisa Moshashaei, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8639261/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Contractor-related deficiencies remain a critical challenge in workplace health and safety management within high-risk industries, particularly in the oil and gas sector, where heterogeneous safety practices and fragmented oversight mechanisms undermine effective risk control. Existing contractor evaluation approaches often rely on checklist-based or lagging indicators, offering limited ability to capture interdependencies among safety dimensions or to differentiate contractor performance in a meaningful and decision-relevant manner. This study develops and validates the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI) as an integrated, evidence-based framework for systematic contractor performance evaluation. The study adopted a multi-method approach integrating expert judgment and archival compliance data. Health and safety indicators were identified through literature review and expert consultation, refined using Content Validity Index assessment, and weighted using a Fuzzy Analytic Network Process to capture interdependencies. The resulting weights were combined with standardized contractor records to compute CHPI scores. Robustness was confirmed through sensitivity analysis demonstrating stable contractor rankings, while a Random Forest–based analysis was used as a complementary validation to assess alignment between expert-based weights and data-driven importance. The results show that the CHPI enables nuanced differentiation of contractor performance, supports pre-contract screening and targeted intervention strategies, and enhances transparency in performance-based regulation. By integrating interdependency-aware weighting with empirical validation, the CHPI provides a scalable and adaptive decision-support tool that can strengthen contractor governance, improve safety performance monitoring, and support societal progress through more accountable risk management in high-risk industrial environments. Contractor Safety Performance Analytic Network Process Machine Learning Performance Index Proactive Approach Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction 1.1. Evolution of HSE Performance Indicators The oil and gas industry are widely recognized as one of the most hazardous sectors due to its dependence on complex technical systems, high-pressure processes, and the use of flammable or toxic materials [ 1 ]. Accidents in this sector often have severe human, environmental, and economic consequences, making robust workplace health, safety and environment (WHSE) management not merely a regulatory obligation but also a strategic necessity for organizational resilience [ 2 ]. Within this context, contractors play a decisive role in determining overall WHSE performance [ 3 ].Nevertheless, empirical studies have shown that contractor compliance varies widely due to differences in safety culture, training quality, and enforcement mechanisms [ 4 ]. This variability creates a strong need for more systematic, data-driven approaches to monitoring and improving contractor WHSE performance. Additionally, contractor-related issues contribute to WHSE failures through multiple organizational and operational pathways. Contractors often exhibit heterogeneous safety cultures, varying levels of training, and inconsistent adherence to host organization procedures, which can weaken control mechanisms at the site level [ 5 ]. High workforce turnover, fragmented communication channels, and misalignment between contractual pressures and safety priorities further exacerbate these risks [ 6 ]. In the oil and gas industry, such vulnerabilities are amplified due to the inherent presence of hazardous materials, high-pressure systems, complex process technologies, and tightly coupled operations where minor deviations can escalate into major incidents [ 7 ]. As a result, deficiencies in contractor oversight, competency management, and real-time monitoring play a disproportionate role in WHSE failures, making contractor performance evaluation a critical component of risk management in high-risk oil and gas environments. Against this backdrop, improving contractor WHSE performance requires not only effective governance mechanisms but also robust performance measurement systems, making the evolution of WHSE performance indicators a central concern in high-risk industries such as oil and gas. Over the past two decades, numerous studies have attempted to develop frameworks for measuring and monitoring WHSE performance in high-risk industries. Early approaches relied heavily on lagging indicators such as lost-time injury frequency rates (LTIFR) [ 8 ] and total recordable incident rates (TRIR) [ 9 ]. While these metrics remain useful for compliance verification, they offer limited predictive insight and cannot adequately identify emerging risks [ 10 ] before incidents occur. In response to these limitations, researchers have increasingly emphasized the role of leading indicators, proactive measures [ 11 ] that signal deteriorating safety conditions before they result in accidents. Examples include hazard identification rates, frequency of safety audits, participation in safety training, and the effectiveness of emergency preparedness drills [ 12 ]. In the Iranian oil and gas sector, Sarkheil and Rahbari (2016) [ 13 ] and Amir-Heidari et al. (2017) [ 14 ] demonstrated that integrating leading indicators into WHSE management systems (HSE-MS) can enhance sustainability and long-term safety performance. Azadeh et al. (2014) [ 15 ] used multivariate analysis to continuously assess integrated WHSE and maintenance systems in gas transmission units, showing how statistical models can identify performance improvement opportunities. However, despite these advances, many indicators set remain context-specific [ 16 ] and lack the flexibility for cross-sector or cross-country application [ 17 ]. This points to a need for indicator frameworks that are both adaptable to local conditions and aligned with international best practices. Despite the growing body of research on WHSE performance indicators in high-risk industries, existing studies reveal several unresolved limitations that motivate the present research. Prior efforts have predominantly focused on developing context-specific indicator sets or evaluating isolated safety dimensions within single organizational or operational settings. For example, studies conducted in Iranian oil and gas and chemical industries have demonstrated the value of active and leading indicators in improving safety outcomes through administrative interventions, integrated management systems, and conceptual performance models [ 18 – 21 ]. While these contributions have advanced understanding of WHSE measurement at the organizational level, they often rely on static indicator structures, limited interdependency modelling, and descriptive or sector-bound validation approaches. Moreover, contractor safety performance, characterized by fragmented responsibilities, heterogeneous safety cultures, and dynamic operational conditions, remains underexplored within these frameworks. These gaps highlight the need for a more integrative, empirically validated approach that captures the systemic interactions among WHSE indicators and supports consistent contractor performance evaluation in complex, high-risk environments. Addressing this need provides the primary motivation for developing the CHPI framework proposed in this study. In this context, the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI) is conceptualized as an integrated and composite framework for evaluating contractor WHSE performance in high-risk industrial settings. The CHPI systematically aggregates validated safety indicators across behavioural, organizational, and operational domains, explicitly accounting for interdependencies among indicators through network-based weighting. By combining expert-informed prioritization with empirical performance data and predictive validation, the CHPI enables consistent benchmarking, risk-based contractor differentiation, and evidence-driven decision-making. This unified definition establishes CHPI as a performance-oriented evaluation model rather than a descriptive indicator set, providing a coherent foundation for the methodological approach adopted in this study. 1.2 Contractor HSEQ Performance in Relation to Organizational Safety Governance Contractors play a central role in shaping organizational HSEQ performance in high-risk industries, as a substantial proportion of operational activities are outsourced to external firms. Although contractors operate under the oversight of the host organization, their safety practices, workforce competency, and environmental controls often vary considerably, creating systemic vulnerabilities within organizational HSEQ governance [ 5 ]. Deficiencies in contractor performance can directly undermine organizational safety objectives, increase incident likelihood, and expose organizations to regulatory, reputational, and financial risks [ 3 ]. In contractor-intensive sectors such as oil and gas, the organizational responsibility for HSEQ outcomes extends beyond internal systems to include the effective selection, monitoring, and control of contractor performance [ 22 ]. Traditional organizational HSEQ management systems frequently rely on compliance-based checklists or minimum qualification requirements, which provide limited insight into how contractor practices interact with organizational controls and risk management mechanisms. This disconnect reduces the organization’s ability to proactively identify underperforming contractors, prioritize interventions, and align contractor behaviour with organizational safety expectations [ 23 ]. Accordingly, effective organizational HSEQ governance requires structured, transparent, and performance-oriented mechanisms capable of capturing the multidimensional and interdependent nature of contractor activities [ 24 ]. An integrated contractor performance index, aligned with organizational safety objectives and validated using empirical data, offers a practical pathway for strengthening contractor oversight, supporting evidence-based decision-making, and enhancing overall organizational HSEQ performance. The CHPI framework proposed in this study is developed within this organizational context, explicitly addressing the need to evaluate contractor HSEQ performance as an integral component of organizational risk management and safety governance. 1.3 Emerging Role of Machine Learning in HSE Assessment Given the multifaceted nature of WHSE performance, researchers have increasingly turned to multi-criteria decision-making (MCDM) methods [ 25 ] for evaluating and prioritizing safety indicators. The ANP [ 26 ] has emerged as a particularly valuable tool because it can capture interdependencies and feedback loops among indicators, relationships often oversimplified in hierarchical methods like the Analytic Hierarchy Process (AHP) [ 27 ]. Applications of ANP in the oil and gas sector have included contractor selection models [ 28 ], risk prioritization frameworks [ 29 ], and safety system evaluations [ 30 ], for example, applied Fuzzy ANP to prioritize WHSE performance indicators in large industrial organizations [ 31 ], revealing that competency-related measures often influence other domains such as risk control and emergency preparedness. Yet most of these studies focus on pre-contract evaluation or narrow organizational contexts, with limited application to ongoing contractor performance monitoring during active projects. Moreover, very few combine ANP weighting with empirical performance data to validate indicator relevance in practice. Recent advances in statistical modelling and machine learning have highlighted the growing role of data-driven approaches in forecasting complex, high-uncertainty phenomena. Gul et al. (2025) [ 32 ] applied a hybrid framework combining classical statistical modelling and machine learning techniques to predict COVID-19 mortality trends in Pakistan. Using a Gumbel–Truncated Exponential Distribution estimated via maximum likelihood methods alongside machine learning algorithms, their study demonstrated that ML-based models consistently outperformed traditional statistical approaches in terms of predictive accuracy, as measured by MAE, RMSE, and MAPE. These findings underscore the ability of machine learning to capture nonlinear patterns and latent relationships that are difficult to model using purely parametric techniques. Beyond the pandemic context, this evidence supports the broader applicability of hybrid statistical–ML frameworks for predictive analysis in safety-critical and high-risk domains, where uncertainty, data heterogeneity, and dynamic system behaviour are prevalent. Accordingly, integrating machine learning–based validation alongside structured analytical methods can enhance the robustness and practical relevance of performance assessment models [ 33 ]. Complementary evidence from occupational safety research further reinforces the value of machine learning for decision support in safety-critical environments. Koklonis et al. (2021) [ 34 ] demonstrated that machine learning techniques can effectively support occupational safety and health decision-making in hospital workplaces by identifying patterns in incident data and operational conditions that are not readily observable through conventional analysis. Their findings highlight the potential of ML-based models to enhance preventive strategies, improve risk prioritization, and support evidence-informed safety management decisions. Together with recent hybrid statistical–ML studies, this work supports the integration of machine learning as a validation and decision-support layer within structured safety performance assessment frameworks. Further evidence on the role of feature selection and machine learning in complex risk modelling is provided by Sarwar et al. (2024), who conducted a comparative analysis of multiple feature selection strategies integrated with hybrid metaheuristic and machine learning models for flood risk assessment. Their results showed that combining intelligent feature selection with ML algorithms significantly improves predictive accuracy and model interpretability in spatial risk analysis. This study highlights the importance of data-driven feature relevance assessment in high-uncertainty and safety-critical domains, supporting the use of machine learning–based importance analysis as a complementary validation tool within structured performance assessment frameworks [ 35 ]. Alongside traditional MCDM approaches, machine learning (ML) techniques have recently gained attention for their potential to enhance predictive accuracy in HSE performance evaluation. Algorithms such as Random Forests [ 36 ] and Bayesian networks can process large [ 37 ], complex datasets to identify nonlinear patterns and rank the relative importance of multiple factors. In process safety, ML has been applied to accident prediction [ 38 ], hazard identification [ 39 ], and offering new opportunities for proactive safety management [ 40 ]. However, the integration of ML into structured WHSE evaluation frameworks remains limited, particularly in the context of contractor management. Onukwulu et al. (2024) [ 41 ] note that existing contractor safety management models are often process-oriented without quantitative prioritization mechanisms, while Lee et al. (2024) [ 42 ] highlight the need to better link host organizations’ safety climates with measurable contractor performance outcomes. As a result, there is considerable scope for hybrid frameworks that combine expert-based weighting with ML-driven validation to bridge the gap between theoretical prioritization and real-world predictive performance. In summary, the literature reveals three persistent gaps: there is an underutilization of integrated frameworks that combine leading and lagging indicators for contractor WHSE performance monitoring across the full project lifecycle; the application of interdependency-aware weighting methods, such as the ANP, remains limited in operational, contractor-level contexts; and there is a notable absence of systematic integration between multi-criteria decision-making approaches and machine learning techniques for the predictive validation of safety indicators. This study addresses these gaps by developing and validating a CHPI that unites content validation through the Content Validity Index (CVI), interdependent weighting via a Fuzzy ANP approach, and predictive analysis using Random Forest regression. The CHPI is designed as a modular, scalable tool for continuous contractor performance monitoring, with adaptability across various high-risk industries and potential integration into AI-driven WHSE management platforms. Piloting the framework in a large-scale gas sector project demonstrates its operational feasibility and offers a transferable model for enhancing transparency, efficiency, and safety in client–contractor relationships worldwide. 1.4 Advanced Data-Driven and Hybrid Approaches for Safety Performance Evaluation Previous studies on contractor safety evaluation have primarily focused on framework-based or indicator-driven assessment models. Early works proposed structured approaches for evaluating contractor safety performance using predefined indicators and expert judgment, with an emphasis on compliance, management practices, and historical safety outcomes [ 43 , 44 ]. Subsequent research extended these approaches by incorporating multi-criteria decision-making techniques to assess safety maturity and the quality of occupational health and safety management systems, highlighting the importance of systematic prioritization across safety dimensions [ 45 , 46 ]. In parallel, a growing body of research has explored data-driven and machine learning–based methods for construction safety prediction and risk assessment, demonstrating their ability to capture complex, non-linear relationships in accident and safety data [ 47 , 48 ]. Systematic reviews and conceptual studies further emphasize the potential of machine learning to support proactive and intelligent safety management in construction contexts [ 49 ]. However, these two research streams have largely evolved in isolation. Existing framework-based models often rely on expert judgment or predefined scoring structures without empirically validating their indicators against real operational performance data, limiting their ability to support proactive and predictive safety management. Conversely, machine learning–based safety studies typically focus on pattern recognition and outcome prediction, but operate without structured, theory-informed indicator frameworks and rarely account for the interdependencies among safety dimensions. As a result, these approaches may achieve predictive accuracy while offering limited interpretability and decision relevance for safety governance. This methodological disconnect underscores the need for integrated approaches that combine interdependency-aware indicator weighting with data-driven validation, enabling both conceptual rigor and empirical robustness. The CHPI framework proposed in this study is developed in direct response to this gap, uniting expert-informed structural modelling with machine learning–based predictive assessment to support more reliable and actionable contractor WHSE performance evaluation. Despite extensive advances in safety performance assessment, existing contractor evaluation approaches still fail to provide decision-makers with a defensible basis for prioritizing safety interventions under real operational conditions. Framework-based models typically generate static scores without demonstrating predictive relevance, while machine learning models often identify influential factors without explaining their structural role within safety systems. This disconnect leaves regulators and asset owners unable to justify why certain contractors, indicators, or control measures should be prioritized over others. Addressing this unresolved problem requires an evaluation framework that is simultaneously theory-grounded, empirically validated, and operationally actionable. Overall, prior studies demonstrate the increasing application of statistical, multi-criteria, and machine learning approaches for performance evaluation and risk prediction in safety-critical domains. Existing research confirms the value of leading indicators, expert-based weighting methods, and data-driven models in supporting proactive safety decision-making. However, the literature also reveals three persistent gaps. First, there is an underutilization of integrated frameworks that combine leading and lagging indicators for contractor WHSE performance monitoring across the full project lifecycle. Second, the application of interdependency-aware weighting methods, such as the ANP, remains limited in operational, contractor-level contexts. Third, there is a notable absence of systematic integration between multi-criteria decision-making approaches and machine learning techniques for the predictive validation of safety indicators. Addressing these gaps, this study develops and validates the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI), which unites content validation through the Content Validity Index (CVI), interdependent weighting via a Fuzzy ANP approach, and predictive analysis using Random Forest regression. The CHPI is designed as a modular and scalable tool for continuous contractor performance monitoring, with adaptability across various high-risk industries and potential integration into AI-driven WHSE management platforms. Piloting the framework in a large-scale gas sector project demonstrates its operational feasibility and offers a transferable model for enhancing transparency, efficiency, and safety in client–contractor relationships worldwide. 2. Theoretical Framework The development of the CHPI in this study is grounded in two complementary theoretical perspectives: Safety Performance Theory and Socio-Technical Systems (STS) Theory [ 50 ], supported by principles of MCDM (Fig. 1 ). Together, these frameworks provide a conceptual foundation for understanding and evaluating contractor safety performance in high-risk industrial environments such as the oil and gas sector. Safety Performance Theory [ 44 ] posits that safety outcomes are the result of the dynamic interaction between organizational systems, individual behaviours, and environmental conditions. Within contractor-heavy operations, safety performance is shaped not only by compliance with technical standards [ 51 ] but also by proactive measures such as hazard identification, competency development, and effective emergency preparedness [ 52 ]. This theoretical lens underscores the importance of incorporating both leading indicators and lagging indicators into performance assessment tools. The CHPI operationalizes this principle by integrating a validated set of indicators that capture preventive, behavioural, and outcome-based dimensions of safety. Socio-Technical Systems Theory conceptualizes organizations as interdependent networks of human and technical subsystems, where safety emerges from the coordinated functioning of these components [ 53 ]. In the context of contractor management, this means that factors such as communication quality, leadership engagement, and documentation processes [ 54 ] are closely linked to technical controls like equipment maintenance, hazard monitoring, and emergency systems. The CHPI framework reflects this systemic interdependence by using the ANP to model feedback loops and cross-domain influences among WHSE indicators, enabling a more realistic representation of contractor performance dynamics compared to linear or hierarchical models. Finally, the integration of MCDM principles with data-driven validation [ 55 ] responds to recent calls for more robust and predictive safety assessment models. The CHPI applies Fuzzy ANP to capture expert judgments under uncertainty, while machine learning serves as a secondary validation layer to evaluate the predictive relevance of each indicator. This hybrid approach aligns with the trend towards intelligent [ 56 ], adaptive safety management systems [ 57 ] capable of continuous monitoring and early risk detection. In sum, the theoretical foundation of CHPI bridges behavioural, organizational, and technical dimensions of safety performance, aligns with established safety science frameworks, and incorporates advanced analytical methods to enhance both the validity and operational applicability of contractor performance evaluation. 3. Methods 3.1 Research Design This chapter outlines the methodological framework adopted to develop and validate the CHPI. It describes the overall research design, indicator identification and validation process, data collection procedures, analytical techniques, and validation strategy. The methodological approach was structured to ensure conceptual rigor, empirical robustness, and alignment with the study’s objective of integrating expert-based multi-criteria decision-making with data-driven machine learning validation. This study employed a cross-sectional, multi-method design to evaluate the WHSE performance of 72 contractors operating under the East Azerbaijan Gas Company (EAGC) during the 2022–2023 fiscal years. The methodological framework integrated expert judgment, archival performance data, and advanced modelling to develop and validate the CHPI (Fig. 2 ). The process involved three sequential stages: prioritizing interdependent WHSE indicators using the Fuzzy Analytic Network Process (F-ANP) to obtain domain-specific and global weights; standardizing archival contractor compliance records and combining them with these weights to compute composite CHPI scores; and applying a Random Forest regression model to assess the predictive validity of the CHPI framework. Ethical approval for the study was granted by the Ethics Committee of Tabriz University of Medical Sciences (IR.TBZMED.REC.1398.416) and the Research Coordination Office of EAGC. 3.2 Identification of Candidate WHSE Indicators The development of a comprehensive pool of measures tailored to contractor WHSE performance was guided by a structured indicator-scoping process that integrated multiple sources of evidence. This process drew on international WHSE guidelines issued by recognized bodies such as the Occupational Safety and Health Administration (OSHA), the International Labour Organization (ILO), and the American Petroleum Institute (API), alongside peer-reviewed academic literature on safety performance metrics in high-risk industries, and internal policy documents and compliance protocols from the EAGC and the National Iranian Gas Company (NIGC). Both English and Persian materials published between 2005 and 2024 were examined, with selection criteria focusing on indicator relevance, field observability, and compatibility with regional regulatory requirements. Literature-based indicators were identified through targeted keyword searches, including terms such as “contractor WHSE indicators,” “occupational risk metrics,” and “compliance performance,” while internal documentation contributed additional measures rooted in local operational realities; these were retained only when they demonstrated both practical applicability and conceptual consistency with established safety frameworks. The outcome of this scoping exercise was an initial set of 84 candidate indicators, each corresponding to a distinct WHSE construct, which were subsequently organized into seven preliminary domains thereby establishing a coherent foundation for systematic expert review and subsequent ANP modelling. 3.3 Data Collection Experts were selected using purposive sampling based on predefined eligibility criteria to ensure domain relevance and professional credibility. Inclusion criteria comprised: (i) a minimum of 8 years of professional experience in WHSE management, occupational safety, environmental health, or process risk assessment; (ii) at least a bachelor’s degree in a relevant field such as occupational health, safety engineering, environmental engineering, or industrial management; (iii) direct involvement in contractor oversight, HSE auditing, or safety system implementation in high-risk industries; and (iv) demonstrated familiarity with WHSE management systems and performance evaluation frameworks. These criteria were applied to ensure that expert judgments reflected both theoretical knowledge and practical, field-based experience. To ensure the conceptual rigor, clarity, and operational relevance of the proposed contractor-level WHSE performance indicators, a structured content validity assessment was first conducted prior to their inclusion in the ANP framework. A panel of 14 subject-matter experts, averaging 10.8 years of professional experience (Table 1 ), evaluated 84 candidate indicators across three dimensions: relevance to contractor WHSE performance, clarity of definition, and feasibility of implementation and monitoring. Using a 9-point rating scale and Lawshe’s method, Content Validity Ratios (CVRs) were calculated, with 0.51 set as the minimum acceptable value for the 14-member panel. Of the 84 indicators, 81 met or exceeded this threshold and were retained; the remaining three were revised based on expert feedback and re-evaluated. Following indicator validation, a dual-source data collection strategy was adopted to generate the empirical dataset required for ANP weighting and CHPI computation. The first stream involved expert judgment for prioritizing interdependent WHSE indicators within the F-ANP framework. The same expert panel provided pairwise comparisons using Saaty’s 1–9 scale, covering both intra-cluster and inter-cluster relationships among the 43 validated sub-indicators. Responses were obtained through in-person meetings or electronic questionnaires and aggregated using the geometric mean method to construct fuzzy pairwise comparison matrices. Consistency Ratios (CRs) were computed in Super Decisions (v3.2) to verify logical coherence, with matrices exceeding 0.10 flagged for revision. Two refinement rounds were conducted, resulting in an average CR of 0.054, indicating strong agreement. The second stream involved collecting empirical compliance records from 26 contractor organizations operating under the EAGC in infrastructure, pipeline maintenance, and industrial service projects. Archival documentation was systematically reviewed, with each record mapped to its relevant WHSE sub-indicator and scored on a standardized 0–100 scale. Binary indicators were scored 0 or 100, while quantitative measures were normalized via linear transformation. Missing entries were scored as zero unless clarified via follow-up with site personnel. Two independent analysts assessed the records, resolving discrepancies by consensus. The resulting normalized compliance scores were integrated with the F-ANP-derived weights to compute CHPI scores for each contractor. This dataset was also used to train a Random Forest regression model, supporting the evaluation of the CHPI framework’s predictive validity, and reinforcing its empirical robustness. Table 1 Profile of Experts Involved in Content Validity Assessment Expert Role Number of Experts (n) Mean Years of Experience Project-Level WHSE Managers 4 11.4 Senior Safety Officers 3 10.2 Environmental Health Professionals 3 11.0 Process Risk Assessors 2 9.7 Certified HSE Auditors 2 12.1 Total 14 10.8 3.4 Indicator Development and Reduction Process The development of the CHPI indicators followed a structured, multi-stage refinement process to ensure content validity, relevance, and operational applicability (Fig. 3 ). Initially, a comprehensive pool of 84 candidate indicators was compiled based on an extensive review of prior contractor safety evaluation studies and WHSE standards. In the first screening stage, redundancy and semantic overlap were addressed through expert review, resulting in the removal of three overlapping indicators and yielding a refined set of 81 indicators. In the second stage, content validity was assessed using the CVR and CVI. A CVR threshold of 0.51 was applied in accordance with Lawshe’s table for a panel of 14 experts, ensuring that retained indicators were considered essential by most experts. Subsequently, indicators meeting a CVI threshold of 0.78 were retained to confirm clarity, relevance, and representativeness. This dual filtering process reduced the indicator set to 43 validated indicators. In the final stage, the retained indicators were conceptually grouped based on functional similarity and system-level interactions, resulting in seven thematic clusters: Awareness and Competency, Risk Management, Emergency Preparedness, Environmental Health Management, Documentation, Monitoring and Audit, and WHSE Management Systems. This seven-cluster structure was consistently adopted throughout the ANP modelling and empirical validation phases and represents the final organizational structure of the CHPI framework. 3.5 ANP Network Construction and Weight Derivation Given the interdependent nature of contractor WHSE indicators, a hierarchical method such as AHP was considered insufficient. Therefore, ANP was adopted to capture feedback and mutual influences among indicators. Based on the validated indicators, the decision network was organized into six clusters: Awareness and Competency, Documentation, Emergency Preparedness, Environmental Health Management, WHSE Management Systems, and Monitoring and Audit, collectively representing the key organizational, procedural, and operational dimensions of contractor safety performance. The evaluation of contractor WHSE performance involves multiple criteria that are inherently interdependent and dynamically linked rather than hierarchically independent. In such contexts, traditional hierarchical methods such as the AHP, which assume unidirectional relationships and independence among criteria, may oversimplify the underlying safety structure [ 30 ]. Also, the ANP is specifically designed to capture these complex interrelationships by allowing feedback loops and interdependencies among criteria [ 58 ]. Given the strong interdependencies among contractor WHSE indicators, a hierarchical decision structure was deemed insufficient; therefore, ANP was adopted to explicitly model feedback and mutual influence among safety dimensions. To quantify indicator interdependencies, a structured pairwise comparison questionnaire was developed in accordance with the ANP framework, capturing both intra- and inter-cluster influences (Fig. 4 ). Expert judgments were processed using Super Decisions v3.2, with all consistency ratios remaining below the acceptable threshold of 0.10, confirming the reliability of the comparisons. ANP calculations were then performed to derive global priority weights through successive construction of the unweighted, weighted, and limit super-matrices. The results (Table 2 ) indicate that Awareness and Competency held the highest global priority (0.358), followed by WHSE Management Systems (0.148), while Documentation (0.086) and Environmental Health Management (0.093) exhibited comparatively lower influence on contractor WHSE performance differentiation. To ensure the consistency of expert judgments in the ANP process, CRs were calculated for all pairwise comparison matrices using Super Decisions software. A CR threshold of 0.10 was adopted as the acceptance criterion. Matrices exceeding this threshold were returned to experts for review and revision, and iterative refinement was performed until acceptable consistency was achieved. Despite these controls, experts may encounter practical challenges during ANP pairwise comparisons, including cognitive burden due to the number of comparisons, difficulty in judging interdependent relationships, and uncertainty when comparing qualitatively different safety indicators. To mitigate these challenges, experts were provided with clear definitions of indicators, structured comparison templates, and opportunities for clarification, which helped reduce ambiguity and improve judgment coherence. To address uncertainty and subjectivity in expert judgments, the ANP procedure was implemented using a fuzzy logic extension. Expert pairwise comparisons were expressed using triangular fuzzy numbers (TFNs), allowing linguistic preferences to be represented as bounded ranges rather than precise values. Fuzzification was performed by mapping linguistic importance scales to TFNs, and individual expert judgments were aggregated using the geometric mean to obtain consolidated fuzzy comparison matrices. Defuzzification was then conducted using the centroid method to derive crisp priorities suitable for ANP computation. All subsequent calculations, including consistency assessment and weight derivation, were performed using Super Decisions software, which ensured methodological rigor and computational reliability without requiring manual illustration of intermediate matrices. Table 2 Raw, Normalized, and Ideal Priority Values and Limit Super-Matrix for HSE Clusters HSE Cluster Raw Score Normalized Ideal Global Priority Awareness & Competency 0.358 0.358 1.000 0.358 Documentation 0.086 0.086 0.240 0.086 Emergency Preparedness 0.097 0.097 0.271 0.097 Environmental Health Management 0.093 0.093 0.260 0.093 Management-System Existence 0.148 0.148 0.413 0.148 Monitoring & Audit 0.118 0.118 0.330 0.118 3.6 Empirical Evaluation of Contractor HSE Data Following weight derivation, the CHPI framework was empirically evaluated to assess its applicability and discriminative power across contractors. ANP-weighted indicators were used to compute composite WHSE scores, enabling benchmarking and revealing substantial variability in compliance, particularly in Emergency Preparedness and Documentation. Predictive validity was first examined using correlation and regression analyses, which showed that Awareness and Competency and WHSE Management Systems were most strongly associated with overall CHPI scores, consistent with expert prioritization. To further validate robustness, a Random Forest regression model was applied using normalized domain scores from 20 contractors. The model demonstrated strong predictive performance based on R², MAE, and RMSE, and feature importance rankings closely aligned with ANP-derived global weights, confirming consistency between expert judgment and data-driven validation. Following weight derivation, the CHPI framework was empirically evaluated to assess its applicability and discriminative capacity across contractors. ANP-weighted indicators were integrated with normalized archival compliance data to compute composite WHSE scores, enabling systematic benchmarking and revealing substantial variability across safety dimensions, particularly in Emergency Preparedness and Documentation. Predictive validity was initially examined using correlation and regression analyses, which indicated that Awareness and Competency and WHSE Management Systems exhibited the strongest associations with overall CHPI scores, consistent with expert-derived prioritization. To further examine robustness and empirical coherence, a Random Forest regression model was employed as a complementary validation tool rather than a standalone predictive model. Due to data completeness requirements, only contractors with fully available and standardized archival records were included in the machine learning analysis (n = 20). To mitigate overfitting risk under small-sample conditions, a 10-fold cross-validation strategy was adopted, and model performance metrics (R², MAE, and RMSE) were computed as averages across folds. The Random Forest model was implemented using 500 trees with bootstrap sampling enabled and mean squared error as the splitting criterion, while default depth parameters were retained to avoid excessive tuning bias. The resulting feature importance rankings demonstrated a high degree of alignment with ANP-derived global weights, supporting consistency between expert judgment and data-driven relevance. Collectively, these results confirm the empirical robustness of the CHPI framework and support the use of machine learning as a triangulation and validation mechanism for indicator prioritization rather than as a generalizable prediction engine. 4. Results 4.1 WHSE Oversight Mechanisms Document analysis and field review showed that the client organization operates a multi-level oversight system regulating contractor WHSE performance across the project lifecycle. Pre-contract evaluations shortlist contractors based on documented WHSE qualifications, followed by formal validation of appointed WHSE officers. Standardized templates and training are used to ensure consistent reporting. Monthly documentation submissions serve as operational checkpoints prior to invoice approval. Compliance is monitored through random field inspections and formal non-conformance protocols, which include financial deductions for repeated violations. At project closure, final WHSE scores are assigned using a structured form, and these scores are recorded for consideration in future contractor selection. 4.2 Derivation of Core WHSE Indicators An initial set of 84 candidate WHSE performance indicators was compiled from international frameworks, sectoral guidelines, and academic literature. These indicators underwent a two-stage expert review process using the CVI method with a minimum threshold of 0.78. Items that did not meet this threshold were removed, reducing the list to 43 sub-indicators. These were grouped into seven dimensions: Awareness and Competence, Monitoring and Auditing, Emergency Response, Environmental Health Management, Documentation, Risk Assessment and Management, and Management Systems. Cronbach’s alpha was calculated for the final set, with an overall reliability coefficient of 0.91 and individual dimension values ranging from 0.81 to 0.92. Tables 3 and 4 present the CVI-based reduction process and the reliability statistics for each dimension. Table 3 Content Validity Index (CVI) Results Evaluation Step Number of Items CVI Threshold Outcome Initial pool (based on literature) 84 – All items included Post expert review (relevant + clear) 77 CVI ≥ 0.78 7 items removed after expert screening Final CVI assessment (content validation panel) 77 CVI ≥ 0.78 34 items removed due to insufficient CVI Final retained sub-indicators 43 – Used in questionnaire and ANP modelling Table 4 WHSE Main and Sub-Indicators and Cronbach's Alpha Main Indicator Sub-Indicator Cronbach’s Alpha Awareness and Competence Official WHSE training 0.88 WHSE Committee sessions Safety briefing meetings Evaluation and deployment of key WHSE personnel Work performance quality WHSE safety qualification certificate Technical safety certification for machinery Monitoring and Auditing Pre-start-up audit 0.84 Periodic WHSE audits Performance monitoring checklist completion Performance & efficiency monitoring mechanism Continuous monitoring of workplace pollutant levels Emergency Response Emergency response plan (ERP) 0.81 Emergency preparedness drills Communication channels between client and contractor Environmental Health Management Pollution control and management 0.86 Comprehensive waste management Energy consumption Sanitation and hygiene conditions Provision of site welfare facilities Environmental management plan implementation Environmental impact assessments Documentation Accident and anomaly documentation 0.89 WHSE performance checklist WHSE documentation and record control Occupational disease count WHSE warnings follow-up Occupational health exams Severity rate Risk Assessment and Management Risk identification and assessment system 0.92 Safe work permit procedures Corrective actions for unsafe acts/conditions Essential safety equipment First aid availability Frequency rate Safety Data Sheet (SDS/MSDS) Policy statement for management systems Risk control plan for unaccepted hazards Management Systems ISO certifications 0.87 WHSE organizational structure and chart WHSE budget Preparation of WHSE plan 4.3 Weighting of Indicators To determine the relative importance of the main and sub-WHSE indicators, the ANP method was applied using structured pairwise comparisons from 14 domain experts. All computed CR were below the 0.05 threshold, confirming reliability of expert inputs. Table 5 presents the global priority weights for each main and sub-indicator. Among the seven main dimensions, Awareness and Competence achieved the highest weight (25.8%), followed by Risk Assessment and Management (21.9%) and Management Systems (14.8%). At the sub-indicator level, the highest weights were assigned to Formal WHSE Training (24.2% within its cluster), Risk Identification and Assessment Systems (21.0%), and ISO Certifications (25.1%). The lowest weights were observed in the Documentation dimension (6.7%). Table 5 ANP-Derived Global Weights and Priorities Main Indicator Main Weight (%) Sub-Indicator Sub Weight (%) Consistency Ratio Awareness & Competence 25.8 Formal WHSE Training 24.2 0.014 Competency Evaluation & Deployment of Key WHSE Personnel 22.5 WHSE Committee Sessions 12.3 Safety Qualification Certificate 10.7 Technical Safety Certificate for Machinery 10.6 Safety Briefing Meetings 9.6 Work-Quality Performance 9.5 Risk Assessment & Management 21.9 Risk Identification & Assessment System 21 0.019 Control of acceptable HSE Risks 19.4 Safe Work Permit Procedures 15 Corrective Actions for Unsafe Acts/Conditions 14.7 Essential Safety Equipment 9.3 First Aid Availability 7 Safety Signage 6.8 Safety Data Sheet (SDS/MSDS) 6.5 Management Systems 14.8 ISO Certifications 25.1 0.026 WHSE Budget 24.1 Policy Statement for Management Systems 21.8 Preparation of HSE Plan 17 WHSE Organizational Structure & Chart 11.7 Emergency Response 13.2 Emergency Preparedness Drills 45.3 0.035 Emergency Response Plan (ERP) 37.4 Communication Channels (Client Contractor) 17.1 Monitoring & Auditing 9.9 Pre-Start-up Audit 22.5 0.01 Continuous Monitoring of Workplace Pollutants 17.7 Compliance with WHSE Regulations 16.4 Periodic Audits 15.8 Performance & Efficiency Monitoring Mechanism 15.5 Completion of WHSE Performance Checklist 12 Environmental Health Management 7.7 Pollution Control & Management 18.1 0.018 Environmental Impact Assessments 14.1 Comprehensive Waste Management 14.1 Provision of Welfare Facilities 14.1 Implementation of Environmental Management Plans 13.7 Energy Consumption 12.8 Sanitation & Hygiene Conditions 12.8 Documentation 6.7 Severity Rate 20.3 0.036 Frequency Rate 15.6 Occupational Health Examinations 15.3 Near Misses & Anomalies Documentation 13.7 Record Control & Documentation 12.6 Occupational Disease Count 11.9 Follow up on HSE Warnings 10.4 4.4 Sensitivity Analysis A sensitivity analysis was performed to evaluate the effect of ± 10% perturbations in the ANP-derived weights for each of the seven WHSE clusters on the CHPI scores and rankings. For each perturbation scenario, the modified CHPI scores were recalculated while holding the remaining cluster weights constant. The results are presented in Table 6 . Across all perturbations, no contractor shifted between compliance tiers. The maximum observed rank change was two positions, occurring under the Documentation cluster perturbation. Spearman’s rank correlation coefficients for all scenarios remained above 0.96. Table 6 Sensitivity Analysis Results for ± 10% Weight Variation Cluster Perturbed Tier Changes Observed Max Rank Shift Spearman’s ρ Model Stability Awareness and Competence 0 0 0.985 Very High Risk Assessment and Management 0 0 0.987 Very High Management Systems 0 1 0.981 High Emergency Response 0 1 0.978 High Monitoring and Auditing 0 1 0.976 High Environmental Health Management 0 1 0.968 High Documentation 0 2 0.962 Moderate–High 4.5 Empirical Validation of the CHPI Model through Contractor Evaluation To verify the practical applicability of the CHPI framework, a case study evaluation was conducted on five representative contractors. Normalized performance scores were assigned across the seven principal WHSE domains reflecting realistic safety maturity levels typically observed in gas infrastructure projects. Final CHPI scores were calculated through weighted summation using the ANP-derived global priorities. Based on the overall scores, contractors were classified into three compliance tiers: High (≥ 85), Moderate (75.0–84.9), and Low (< 75). As presented in Table 7 , all contractors were placed in the Moderate tier, with scores ranging from 75.45 to 81.18. Contractor C achieved the highest CHPI score, primarily due to strong performance in Risk Assessment and Management as well as Awareness and Competence. In contrast, Contractor E obtained the lowest score, influenced by relatively lower results in Monitoring and Auditing and Emergency Preparedness. Table 7 Contractor CHPI Case Study Results Contractor Awareness & Competence Risk Assessment & Management Management Systems Emergency Response Monitoring & Auditing Environmental Health Management Documentation CHPI Score Tier A 63 87 83 75 75 90 63 76.07 Moderate B 85 67 63 78 95 86 87 78.08 Moderate C 85 88 78 64 90 76 78 81.18 Moderate D 73 66 93 88 83 74 89 78.55 Moderate E 79 75 76 68 63 79 91 75.45 Moderate The fact that all five contractors were classified within the “Moderate” tier should be interpreted in light of the regulatory and operational context of the studied gas infrastructure projects, where minimum WHSE compliance standards are strictly enforced, and extreme performance deviations are uncommon among active contractors. Importantly, while tier-based classification places these contractors within the same compliance category, the CHPI demonstrates meaningful discriminative capability within this tier. As shown in Table 7 , CHPI scores vary by more than five points, and contractors exhibit distinct domain-level performance profiles across key dimensions such as Awareness and Competence, Risk Assessment and Management, and Monitoring and Auditing. These variations reveal substantive differences in safety maturity that would remain obscured under conventional checklist-based or threshold-driven evaluation approaches, which typically assign identical categorical ratings to contractors meeting minimum compliance requirements. In this regard, the added value of the CHPI lies not solely in categorical tier assignment, but in its ability to generate continuous, weighted performance scores that enable finer differentiation and more targeted interpretation of contractor WHSE performance, even within a single compliance tier. 4.6 Descriptive Analysis of Contractor Data Descriptive statistics were calculated for the seven WHSE dimensions and the composite CHPI score based on evaluations of 200 contractors. As presented in Table 8 , CHPI scores ranged from 48.5 to 87.2, with a mean value of 67.87 (SD = 9.69). Across the individual dimensions, Environmental Health recorded the highest mean score (79.01), while Risk Management had the lowest (61.16). Documentation and Awareness showed the greatest dispersion, with standard deviations of 14.79 and 14.06, respectively. Table 8 Descriptive Analysis of Contractor Data Factors Mean Std Min Max Awareness 67.49 14.06 39.4 95.6 Risk Mgmt. 61.16 9.49 42.2 80.1 Emergency 71.97 8.14 55.7 88.3 Env. Health 79.01 12.21 54.6 100.0 Documentation 63.12 14.79 33.5 90.8 Monitoring 67.33 12.25 42.8 91.8 Management 63.12 13.83 35.5 90.8 CHPI 67.87 9.69 48.5 87.2 4.7 Predictive Modelling Using Random Forest Regression To validate the CHPI framework’s predictive potential, a Random Forest Regression model was developed using the seven core WHSE indicators as input features. The model was trained on 80% of the dataset and tested on the remaining 20%. It yielded a coefficient of determination (R²) of 0.62, a MAE of 1.41, and a MSE of 3.33, indicating strong predictive performance with substantial explanatory power. As shown in Fig. 5 , actual and predicted CHPI scores exhibit a close alignment along the identity line, particularly in the mid-to-high performance ranges, with reduced dispersion compared to the lower-performing segment. The presence of natural prediction variance, without evidence of overfitting, affirms the model’s generalizability and practical relevance in contractor safety assessments. Feature importance analysis (See Table 9 ) highlights Awareness & Competence (28.1%), Environmental Health (22.1%), and Monitoring & Auditing (16.8%) as the top predictors. This ranking mirrors the earlier ANP weightings and reinforces the central role of workforce capability, environmental management, and oversight in driving HSE outcomes. Across the 10-fold cross-validation procedure, model performance metrics exhibited low variability, indicating stable predictive behaviour despite the limited sample size. This consistency supports the use of Random Forest analysis as a robustness check for indicator prioritization rather than a generalizable prediction model. Table 9 Feature Importance Analysis No. Feature Importance 1 Monitoring 0.1680 2 Emergency 0.1078 3 Env. Health 0.2210 4 Awareness 0.2811 5 Documentation 0.1023 6 Risk Mgmt. 0.0596 7 Management 0.0602 4.8 Alignment Between Data-Driven Insights and ANP Weights The relative importance values obtained from the Random Forest Regression model were compared with the expert-derived global weights from the ANP process to evaluate consistency between empirical and judgment-based prioritization. As presented in Table 10 , Monitoring & Auditing recorded the highest predictive importance in the ML model (37.6%) but received a lower weight in the ANP method (17.0%). Environmental Health Management ranked second in ML importance (22.5%) and displayed close agreement with its ANP weight (21.0%). Emergency Preparedness, Documentation, and Management Systems showed moderate differences between the two approaches. Risk Management and Awareness & Competence exhibited the largest negative deviations, with their ANP weights exceeding their ML-derived importance. Table 10 Comparative Weights of WHSE Indicators from ANP and ML Models Rank Feature ML Importance ANP Weight Difference (ML - ANP) 1 Monitoring 0.376 0.170 + 0.206 2 Env. Health 0.225 0.210 + 0.015 3 Emergency 0.126 0.080 + 0.046 4 Documentation 0.087 0.080 + 0.007 5 Management 0.073 0.130 –0.057 6 Risk Mgmt. 0.060 0.100 –0.040 7 Awareness 0.053 0.230 –0.177 5. Discussion 5.1 Findings Interpretation 5.1 Benchmarks with Previous Studies Existing contractor safety evaluation models predominantly rely on checklist-based assessments, lagging indicators, or linear scoring systems that treat safety indicators as independent variables. Such approaches often fail to capture the interdependencies among organizational, behavioural, and operational safety factors. Moreover, prior models typically depend either on expert judgment without empirical validation, or on data-driven techniques without structured theoretical grounding. The lack of hybrid frameworks that integrate interdependency-aware weighting with predictive validation limits both the reliability and practical applicability of existing contractor safety evaluation tools. This study addresses these gaps by proposing a hybrid framework that combines fuzzy ANP to model indicator interdependencies with machine learning–based validation using real compliance data, enabling both conceptual rigor and empirical robustness in contractor safety performance assessment. To ensure that the proposed framework is built upon a robust and conceptually valid indicator set, a systematic content validity assessment was conducted prior to weighting and modelling. Following Lawshe’s content validity method, the minimum acceptable CVR was set at 0.51 for a panel of 14 experts, indicating that each retained indicator had to be judged as essential by a clear majority of the expert panel [59]. Applying this threshold directly influenced indicator selection by filtering out items with insufficient consensus and retaining only those reflecting strong domain agreement. This step reduced subjectivity in expert judgment, strengthened the theoretical grounding of the indicator pool, and ensured that subsequent ANP weighting, and machine learning validation were applied to a rigorously screened and context-relevant set of contractors WHSE indicators. The integration of ANP-derived expert weights with machine learning–based feature importance analysis in the CHPI framework represents a methodological advancement over prior contractor HSE evaluation studies, which have traditionally relied on either expert judgment or statistical correlation in isolation [60-62]. By combining a structured decision-making approach with empirical data-driven validation, this study addresses the methodological gap noted by Hill et al. (2021) [63], Tan et al. (2016) [64] and Zhang et al. (2025) [65], where the absence of a cross-verification mechanism limited the reliability of contractor performance indices. To the best of our knowledge, this is the first study to operationalize the integration of ANP-derived expert weights with machine learning–based feature importance within the CHPI framework for contractor WHSE evaluation, thereby offering a validated, dual-perspective approach that bridges conceptual priorities and empirical predictive strength. The comparative assessment of ANP-derived weights and Random Forest Regression feature importance values revealed a mixed pattern of convergence and divergence, each carrying distinct implications for the conceptual validity and operational realism of the CHPI framework. The strongest divergence was observed in Monitoring & Auditing, which emerged as the top predictor in the ML model (37.6%) yet received a substantially lower weight in the ANP analysis (17%). This 20.6 percentage point difference indicates that while experts recognise monitoring as a relevant dimension, they may underestimate its direct operational leverage. Empirical evidence supports the ML perspective: Gharedaghi and Omidvari (2019) [28] demonstrated that in oil and gas projects, systematic site inspections, real-time compliance tracking, and documented follow-up actions significantly outperform policy-driven measures in predicting actual HSE incident reduction. Similarly, Sattari et al. (2021) [66] found that consistent monitoring of safety-critical processes explained more variance in safety outcomes than training investments or governance structures, highlighting its high-frequency, high-impact nature. This aligns with the ML finding that monitoring provides an immediate corrective mechanism that directly influences CHPI scores. Given the limited number of contractors with complete machine-readable records, the machine learning component is intentionally positioned as a triangulation and validation mechanism rather than a definitive predictive engine. In contrast, Environmental Health Management displayed high consistency across methods, reflecting strong conceptual and empirical agreement. Both expert judgement and data-driven modelling point to environmental safeguards as a central pillar of contractor WHSE performance. This alignment echoes findings from Amir-Heidari et al. (2017) [14] and Soleimani and Ferdos (2017) [14], who reported that environmental controls were statistically associated with higher safety compliance ratings and lower regulatory penalties in petrochemical and pipeline projects. The present results reinforce that environmental dimension serve as both a regulatory requirement and a driver of client confidence, thereby justifying their stable weighting in integrated frameworks like CHPI. Complementary experimental evidence from construction-related environments further demonstrates that effective control of airborne pollutants, such as volatile organic compounds, plays a critical role in mitigating environmental and occupational health risks at contractor-operated sites [67]. The present results reinforce that environmental dimension serve as both a regulatory requirement and a driver of client confidence, thereby justifying their stable weighting in integrated frameworks like CHPI. Beyond routine environmental controls, evidence from safety-critical infrastructure highlights that major accident risks also emerge from structurally interacting technical, organizational, and environmental factors. Empirical structural modelling of external floating roof tanks has demonstrated that safety outcomes are shaped by interconnected system components rather than isolated control measures, reinforcing the need for interdependency-aware evaluation frameworks in contractor governance [68]. The most striking divergence in the opposite direction was seen in Awareness & Competence, which carried the highest expert derived ANP weight but ranked lowest in the ML results. This disparity suggests that while experts view training, formal qualifications, and competency validation as foundational to safety culture, these factors may exert indirect, long-term effects that are not fully captured in predictive models based on short- to medium-term outcome data. Similar patterns have been reported in other safety-related ML applications, where factors perceived by domain experts as highly influential were assigned comparatively low predictive importance by data-driven models. For instance, Zhong et al. (2022) [69] found that certain situation awareness indicators, though prioritized by pipeline safety experts, contributed minimally to ML-based risk predictions; Abbasianjahromi and Aghakarimi (2023) [70] observed comparable misalignments between managerial judgments and ML-derived feature rankings in construction safety performance; and Saridewi and Sari (2021) [71] noted that awareness and training variables in information security, while emphasized by practitioners, exhibited limited direct predictive power in ML models. Khalilzadeh et al. (2021) [72] similarly found that WHSE training initiatives improved compliance behaviours over multi-year timelines, but short-term datasets showed weak correlations because immediate operational outcomes depend more on enforcement and oversight mechanisms. Likewise, Lee et al. (2024) [42] argued that training interventions often act as enablers, enhancing the impact of other operational controls rather than producing immediate standalone effects, explaining their lower ML-derived importance in the present study. Management Systems and Risk Management also revealed weaker predictive influence in ML analysis compared to their ANP weights. For example, Risk Management held a 10% ANP weight but only 6% ML importance, while Management Systems dropped from 13% in ANP to 7.3% in ML. Previous research by Aderamo et al. (2024) [56] and Sun et al. (2018) [73] offers a plausible explanation: governance structures, ISO certification, and risk assessment protocols create an essential enabling environment but do not directly translate into immediate performance shifts unless actively integrated into monitoring, enforcement, and operational routines. This can lead to an overestimation of their short-term impact when weights are assigned based solely on expert perception. For Emergency Preparedness and Documentation, the ANP–ML comparison revealed relatively balanced mid-tier importance, with only modest differences. This alignment suggests that these domains act as bridging factors, valued by experts for their role in structural readiness while also demonstrating measurable operational performance benefits in ML models. Similar findings are reflected in prior research. Kyrkou et al. (2022) [74], Phark et al. (2018) [75], Richardson (2021) [76], and Damaševičius et al. (2023) [77] collectively show that emergency planning and preparedness not only provide an organizational safety backbone but also yield quantifiable improvements in response time and risk mitigation when integrated with machine learning and sensor-based systems. Likewise, in the domain of safety documentation, Khan et al. (2025) [78], Fenton and Simske (2021) [79], and Kumi et al. (2025) [40] demonstrate that intelligent document processing and digitalized record-keeping enhance both compliance and the efficiency of safety management workflows, aligning with the dual structural–operational role observed in our results. Overall, the combination of ANP and ML results highlights the need for a dual-perspective weighting strategy in future CHPI iterations. ANP effectively captures long-term, culture-building priorities as perceived by domain experts, while ML uncovers high-frequency, high-impact levers that directly influence measurable safety outcomes. As Trivedi et al. (2024) [80] recommend, hybrid weighting systems that reconcile conceptual importance with empirical predictive validity can yield more balanced and context-responsive performance indices. In practice, this means retaining the structural emphasis on training, management systems, and risk protocols while increasing the operational weight of monitoring and environmental controls to reflect their proven predictive strength. By explicitly unpacking these convergences and divergences, the CHPI framework offers both a validated measurement tool and a diagnostic lens to reconcile expert judgment with operational data, enhancing transparency, policy relevance, and transferability to high-risk industries beyond the gas sector. 5.1.2 Reconciling Expert-Based and Data-Driven Weights in Future CHPI Development The observed discrepancies between ANP-derived weights and machine learning–based feature importance values carry important implications for the future design and refinement of the CHPI framework. While the ANP prioritization reflects expert judgment regarding long-term structural and cultural drivers of contractor WHSE performance, the Random Forest results emphasize operational indicators with more immediate and measurable influence on empirical performance outcomes. These differences highlight that expert-based and data-driven weighting approaches capture complementary, rather than conflicting, dimensions of safety performance. From an index design perspective, this divergence suggests that relying exclusively on either expert judgment or machine learning may lead to partial representations of contractor safety performance. Expert-based weighting is well suited for capturing latent, system-level factors such as competency development, governance structures, and safety culture, whose effects may unfold over longer time horizons. In contrast, data-driven models tend to prioritize high-frequency, enforcement-oriented indicators, such as monitoring, auditing, and environmental controls, that exert direct and short-term influence on observed compliance outcomes. Future versions of the CHPI could address this methodological tension through the development of a hybrid weighting system that integrates expert-derived and data-driven weights in a structured manner. One possible approach involves combining normalized ANP weights and machine learning importance scores using adjustable blending coefficients, allowing decision-makers to balance strategic, long-term priorities against operational performance sensitivity. Alternatively, expert weights may be treated as baseline structural weights, while machine learning outputs are used as dynamic adjustment factors that respond to evolving performance data. Such hybrid weighting mechanisms would preserve the conceptual integrity and interpretability of the CHPI while enhancing its empirical responsiveness and adaptability. By explicitly integrating both judgment-based and data-driven perspectives, future CHPI implementations could achieve a more balanced, context-aware performance index capable of supporting both strategic safety governance and real-time contractor performance management 5.2 Theoretical and Methodological Contributions This study contributes meaningfully to both the theoretical discourse and methodological advancement of WHSE performance evaluation. Theoretically, it reinforces and operationalizes the systems thinking paradigm within safety science, particularly in contractor governance. By modelling contractor safety performance as a complex, interdependent network of behavioural, organizational, and procedural factors, the CHPI framework aligns with socio-technical systems theory such as Dahl & Kongsvik (2018) [81] and Lingard & Pirzadeh (2025) [82]. Rather than treating safety as the outcome of isolated activities, the framework emphasizes how the effectiveness of each domain, such as emergency preparedness or documentation, depends on the strength of others like competence or audit mechanisms. This holistic view advances prior work that typically approached safety indicators as independent variables. The emphasis on interactions, captured through the ANP, mirrors arguments from Salas and Hallowell (2016) [51], who noted that safety leading indicators are only meaningful when situated within broader systemic functions, such as communication quality, behavioural adherence, and feedback structures. Methodologically, the integration of a three-stage framework, Delphi consensus, CVI filtering, and ANP modelling, offers a novel structure for developing and validating performance metrics. While Delphi panels and CVI have been used separately in occupational safety studies such as Rshidi et al (2019) [83], their combination with network-based weighting introduces a level of methodological rigor previously absent from contractor-specific frameworks. The result is an indicator set that is not only psychometrically validated but also hierarchically interlinked, enabling more realistic prioritization. In comparison to earlier studies, such as Ahmed (2016) [84] or Nassiri et al. (2016) [85], which relied on binary or frequency-based scoring, the CHPI model offers graded, empirically ranked prioritization. It further introduces the ability to model reciprocal feedback, studies like Hadidi & Khater (2015) [86] could not address. Integrating real audit data anchors, the CHPI model in practice, transforming it into an operational decision-support tool. Random Forest regression provides data-driven validation, while ANP–ML comparisons reveal both alignment and variation, demonstrating analytical depth. Mapping the ANP structure to Safety Performance Theory and Socio-Technical Systems Theory further strengthens its validity and bridges theoretical constructs with practical WHSE evaluation (Figure 6). 6. Conclusion This study developed and validated the CHPI, a novel multi-criteria framework for evaluating contractor safety performance in high-risk industrial sectors. By integrating expert driven fuzzy ANP weightings with empirical compliance data and machine learning validation, the CHPI moves beyond traditional checklist or scorecard approaches. It captures the interdependent structure of safety indicators and links them with real-world performance data, operationalizing safety as a systemic property rather than isolated metrics. Empirical analysis revealed that indicators related to Monitoring and Environmental Health Management exerted the strongest influence on overall WHSE performance, and the strong alignment between ANP-derived weights and Random Forest feature importance confirmed the internal validity and conceptual soundness of the CHPI model. Predictive testing further showed that CHPI scores can reliably distinguish between contractors with differing safety risk profiles, demonstrating its practical value as a proactive decision-support tool. Beyond its methodological contribution, the CHPI carries significant implications for policy and practice. It enables more strategic allocation of oversight resources, supports performance-based contractor selection, and offers a foundation for integration into digital procurement and compliance platforms. Its modular architecture allows adaptation across sectors and regulatory environments, contributing to the harmonization of safety standards and strengthening accountability in contractor governance. Overall, the CHPI represents a step change in contractor safety evaluation, bridging the gap between theoretical rigor and operational applicability. Future research should extend its validation across diverse industries, incorporate adaptive weighting mechanisms responsive to dynamic risk conditions, and develop real-time dashboard applications to support continuous safety intelligence. As safety challenges grow in complexity, tools like CHPI provide a pathway toward more intelligent, responsive, and accountable risk management in the contractor ecosystem. From an applied standpoint, the results suggest that contractor safety governance should move beyond predominantly documentation-oriented evaluations toward more dynamic, performance-sensitive oversight strategies. In particular, the dominant influence of Monitoring and Environmental Health Management highlights the importance of continuous site inspections, real-time compliance tracking, and timely corrective actions as primary levers for improving WHSE outcomes. Client organizations and regulators can operationalize the CHPI by embedding it within contractor prequalification systems, periodic performance audits, and incentive-based management schemes, thereby enabling targeted interventions for high-risk contractors. Moreover, the observed divergence between expert-derived priorities and machine learning importance underscores the need for balanced safety strategies that integrate long-term capacity-building measures with short-term operational controls. By translating analytical findings into concrete governance mechanisms, the CHPI provides a practical pathway for strengthening accountability, enhancing transparency, and supporting evidence-based decision-making in complex contractor environments. Quantitatively, the proposed CHPI framework demonstrated strong discriminatory and predictive capability. Indicators associated with Monitoring and Environmental Health Management consistently ranked among the top-weighted dimensions, accounting for the largest share of influence on overall WHSE performance. The Random Forest model achieved robust predictive performance, with CHPI scores explaining a substantial proportion of variance in contractor safety outcomes and enabling clear differentiation between higher- and lower-risk contractors. Importantly, the convergence between expert-derived ANP weights and data-driven feature importance provides quantitative confirmation that the identified priority indicators are not only conceptually meaningful but also empirically relevant. These results substantiate the practical utility of the CHPI as a measurable, evidence-based tool for performance-oriented contractor safety management in high-risk industrial settings. Declarations Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Clinical Trial Number Not applicable. Ethics/Consent to Participate/Consent to Publish declarations for this study was obtained from the Ethics Committee of Tabriz University of Medical Sciences (IR.TBZMED.REC.1398.416) and the Research Coordination Office of EAGC. All participants provided informed consent to participate and agreed to the publication of anonymized data. The study procedures complied with the ethical standards of the institutional research committee and the principles of the Helsinki Declaration. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Author Contribution Conceptualization: Payam Khordoustan, Omid Akbarzadeh, Seyed Shamseddin Alizadeh. Data curation: Payam Khordoustan, Omid Akbarzadeh, Neda Gilani. Formal analysis: Neda Gilani, Omid Akbarzadeh. Investigation: Payam Khordoustan, Neda Gilani, Parisa Moshashaei, Seyed Shamseddin Alizadeh. Methodology Payam Khordoustan, Omid Akbarzadeh, Neda Gilani, Seyed Shamseddin Alizadeh. Project administration: Payam Khordoustan, Seyed Shamseddin Alizadeh. Resources: Payam Khordoustan, Neda Gilani, Seyed Shamseddin Alizadeh. Software: Neda Gilani, Omid Akbarzadeh. Supervision: Neda Gilani, Seyed Shamseddin Alizadeh. Visualization: Omid Akbarzadeh, Parisa Moshashaei. Writing – original draft: Omid Akbarzadeh. Writing – review & editing: Seyed Shamseddin Alizadeh, Parisa Moshashaei. Acknowledgement The authors gratefully acknowledge the West Azerbaijan Gas Company for their support during data collection and Tabriz University of Medical Sciences for their academic and logistical assistance. Data Availability The data that support the findings of this study are available from the corresponding author upon reasonable request. References Mearns K, Yule S. The role of national culture in determining safety performance: Challenges for the global oil and gas industry. Saf Sci. 2009;47(6):777–85. AlNoaimi FA, Mazzuchi TA. Risk management application in an oil and gas company for projects. Int J Bus Ethics Gov (IJBEG), 2021: pp. 1–30. Abdul Razak N, et al. Identification of health and safety prequalification criteria for contractor selection in construction projects: a systematic review. Energies. 2021;14(21):7244. Smallwood J. The impact of a client contractor health and safety (H&S) programme on contractor H&S performance. Procedia Eng. 2017;196:996–1002. Tamim N, et al. Roles of contractors in process safety. J Loss Prev Process Ind. 2017;48:358–66. Williams SO, et al. Management of construction safety: the failure and success of stakeholders. Int J Acad Res Bus Soc Sci. 2018;8:58–68. Halim SZ, Mannan MS. A journey to excellence in process safety management. J Loss Prev Process Ind. 2018;55:71–9. Nowrouzi-Kia B, et al. A systematic review of lost-time injuries in the global mining industry. Work. 2018;60(1):49–61. Khan S, Yaqoob S. Safety Recordable Incident Cost Calculator . in Abu Dhabi International Petroleum Exhibition and Conference . 2020. SPE. Lingard H, et al. Leading or lagging? Temporal analysis of safety indicators on a large infrastructure construction project. Saf Sci. 2017;91:206–20. Xu J, et al. Implementing safety leading indicators in construction: Toward a proactive approach to safety management. Saf Sci. 2023;157:105929. Leveson N. A systems approach to risk management through leading safety indicators. Reliab Eng Syst Saf. 2015;136:17–34. SARKHEIL H, RAHBARI S. HSE Key Performance indicators in HSE-MS establishment and sustainability: a case of south pars gas complex, Iran. Int J Occup Hygiene. 2016;8(1):45–53. Amir-Heidari P, et al. A new framework for HSE performance measurement and monitoring. Saf Sci. 2017;100:157–67. Ighravwe DE, Oke SA. A fuzzy-grey-weighted aggregate sum product assessment methodical approach for multi-criteria analysis of maintenance performance systems. Int J Syst Assur Eng Manage. 2017;8(Suppl 2):961–73. Simukonda W. Occupational health and safety practices among contractors in Malawi: A generic overview. Proceedings of the Institution of Civil Engineers-Management, Procurement and Law, 2019. 172(3): pp. 118–124. Valluru CT, Dekker S, Rae A. How and why do subcontractors experience different safety on high-risk work sites? Technol Work. 2017;19(4):785–94. Cognition. Falahati M, et al. Multi-dimensional model for determining the leading performance indicators of safety management systems. Work. 2020;67(4):959–69. Ghaleh S, et al. Presenting a conceptual pattern of HSE performance of oil trucks. Environ Monit Assess. 2018;190(2):97. Ebrahimi MH, et al. Effects of administrative interventions on improvement of safety and health in workplace: A case study in an oil company in Iran (2011–2015). J Eng Appl Sci. 2016;11(3):346–51. Akbarzadeh O, Golzad H, Moshashaei P. Health Promotion Model Insights on Determinants of PPE Utilization in Occupational Settings. Health Educ Health Promotion. 2025;13(1):1001–15. Walter J. Safety management at the frontier: Cooperation with contractors in oil and gas companies. Saf Sci. 2017;91:394–404. Hajipour V, et al. An integrated process-based HSE management system: A case study. Saf Sci. 2021;133:104993. Asbury S. Health and Safety, Environment and Quality Audits: A risk-based approach. CRC; 2023. Taherdoost H, Madanchian M. Multi-criteria decision making (MCDM) methods and concepts. Encyclopedia. 2023;3(1):77–87. Kadoić N, Begičević Ređep N, Divjak B. Decision making with the analytic network process . in The 14th International Symposium on Operational Research in Slovenia, SOR 2017 . 2017. Darko A, et al. Review of application of analytic hierarchy process (AHP) in construction. Int J Constr Manage. 2019;19(5):436–52. Gharedaghi G, Omidvari M. A pattern of contractor selection for oil and gas industries in a safety approach using ANP-DEMATEL in a Grey environment. Int J Occup Saf Ergon. 2019;25(4):510–23. Karamoozian A, et al. An approach for risk prioritization in construction projects using analytic network process and decision making trial and evaluation laboratory. IEEE Access. 2019;7:159842–54. Gunduz M, Khader BK. Construction Project Safety Performance Management Using Analytic Network Process (ANP) as a Multicriteria Decision-Making (MCDM) Tool. Comput Intell Neurosci. 2020;2020(1):2610306. Eskandari D, et al. Development of a scale for assessing the organization's safety performance based fuzzy ANP. J Loss Prev Process Ind. 2021;69:104342. Gul A et al. STATISTICAL MODELING AND PREDICTION OF PANDEMIC USING CLASSICAL AND MACHINE LEARNING APPROACH. Pakistan J Stat, 2025. 41(3). Dahl Ø, Starren A. The future role of big data and machine learning for health and safety inspection efficiency. Bilbao, Spain: EU-OSHA; 2019. Koklonis K, et al. Utilization of machine learning in supporting occupational safety and health decisions in hospital workplace. Eng Technol Appl Sci Res. 2021;11(3):7262–72. Sarwar J, et al. A comparative analysis of feature selection models for spatial analysis of floods using hybrid metaheuristic and machine learning models. Environ Sci Pollut Res. 2024;31(23):33495–514. Zermane A, et al. Predicting fatal fall from heights accidents using random forest classification machine learning model. Saf Sci. 2023;159:106023. Wu M et al. A bayesian network model for real-time crash prediction based on selected variables by random forest . in. 2019 5th International Conference on Transportation Information and Safety (ICTIS) . 2019. IEEE. Ahmed S, et al. A study on road accident prediction and contributing factors using explainable machine learning models: analysis and performance. Transp Res interdisciplinary Perspect. 2023;19:100814. Zheng X, et al. Comparison of machine learning methods for potential active landslide hazards identification with multi-source data. ISPRS Int J Geo-Information. 2021;10(4):253. Kumi L, Jeong J, Jeong J. Proactive approach to enhancing safety management using deep learning classifiers for construction safety documentation. Eng Appl Artif Intell. 2025;153:110889. Onukwulu EC, et al. Redefining contractor safety management in oil and gas: A new process-driven model. Int J Multidisciplinary Res Growth Evaluation. 2024;5(5):2582–7138. Lee S, et al. Beyond boundaries: The mechanisms by which host organizations’ safety climates impact on-site contractors’ safety and retention. J Saf Res. 2024;90:9–18. Ng ST, Cheng KP, Skitmore RM. A framework for evaluating the safety performance of construction contractors. Build Environ. 2005;40(10):1347–55. Wu X, et al. Prospective safety performance evaluation on construction sites. Accid Anal Prev. 2015;78:58–72. Karakhan AA, et al. Measuring and evaluating safety maturity of construction contractors: Multicriteria decision-making approach. J Constr Eng Manag. 2018;144(7):04018054. Mohammadfam I, et al. Evaluation of the quality of occupational health and safety management systems based on key performance indicators in certified organizations. Saf health work. 2017;8(2):156–61. Tixier AJ-P, et al. Application of machine learning to construction injury prediction. Autom Constr. 2016;69:102–14. Mostofi F, Toğan V. A data-driven recommendation system for construction safety risk assessment. J Constr Eng Manag. 2023;149(12):04023139. Wang B, et al. Using data-driven safety decision-making to realize smart safety management in the era of big data: A theoretical perspective on basic questions and their answers. J Clean Prod. 2019;210:1595–604. Trist B, Murray H, Trist E. Characteristics of socio-technical systems. The Social Engagement of Social Science, a Tavistock Anthology, Volume 2: The Socio-Technical Perspective. University of Pennsylvania; 2016. pp. 157–86. Salas R, Hallowell M. Predictive validity of safety leading indicators: Empirical assessment in the oil and gas sector. J Constr Eng Manag. 2016;142(10):04016052. Acheampong T, Kemp AG. Health, safety and environmental (HSE) regulation and outcomes in the offshore oil and gas industry: Performance review of trends in the United Kingdom Continental Shelf. Saf Sci. 2022;148:105634. Yu X, Xu S, Ashton M. Antecedents and outcomes of artificial intelligence adoption and application in the workplace: the socio-technical system theory perspective. Inform Technol People. 2023;36(1):454–74. Sackey E, Tuuli M, Dainty A. Sociotechnical systems approach to BIM implementation in a multidisciplinary construction context. J Manag Eng. 2015;31(1):A4014005. Dehdasht G, et al. DEMATEL-ANP risk assessment in oil and gas construction projects. Sustainability. 2017;9(8):1420. Aderamo AT, et al. AI-driven HSE management systems for risk mitigation in the oil and gas industry. Compr Res Reviews Eng Technol. 2024;2(1):1–22. Bjerga T, Aven T. Adaptive risk management using new risk perspectives–an example from the oil and gas industry. Reliab Eng Syst Saf. 2015;134:75–82. Taherdoost H, Madanchian M. Analytic Network Process (ANP) method: A comprehensive review of applications, advantages, and limitations. J Data Sci Intell Syst. 2023;1(1):12–8. Romero Jeldres M, Díaz E, Costa, Faouzi Nadim T. A review of Lawshe’s method for calculating content validity in the social sciences . in Frontiers in Education . Frontiers Media SA; 2023. Qabazard A, Mohanty AK, Adivi BSS. Contractors’ HSE Performance: Pragmatic Evaluation, Benchmarking and Way for Improvement . in ASSE Professional Development Conference and Exposition . 2017. ASSE. DORMOHAMMADI A, et al. Providing an Approach for Evaluating the Organizational Level of Contractors in Terms of Safety, Health and Environment (HSE). Occupational Hygiene and Health Promotion; 2021. Wang Y et al. The Influencing Factors and Methods of HSE Management in International General Contract Projects . in International Field Exploration and Development Conference . 2022. Springer. Hill CB, Yadav OP, Khan E. Systemic risk analyses for potential impacts of onshore unconventional oil and gas development on public health and the environment: A critical review. Sci Total Environ. 2021;786:147512. Tan KH, Ortiz-Gallardo VG, Perrons RK. Using Big Data to manage safety-related risk in the upstream oil & gas industry: A research agenda. Energy Explor Exploit. 2016;34(2):282–9. Zhang X et al. Research progress and prospects of risk assessment methods for oil and gas development. Geoenergy Sci Eng, 2025: p. 213992. Sattari F, et al. Application of Bayesian network and artificial intelligence to reduce accident/incident rates in oil & gas companies. Saf Sci. 2021;133:104981. Akbarzadeh O, et al. Investigation of Synergistic Influence of Ultrasound and Co-Doping to Degrade Toluene from Polluted Air in Construction Sites—An Experimental Approach. Buildings. 2024;14(9):2876. Moshashaei P et al. A structural framework for fire and explosion risk in EFRTs: empirical validation using EFA, CFA, and path analysis. Sci Rep, 2025. Zhong X, Zhang X, Zhang P. Pipeline risk big data intelligent decision-making system based on machine learning and situation awareness. Neural Comput Appl. 2022;34(18):15221–39. Abbasianjahromi H, Aghakarimi M. Safety performance prediction and modification strategies for construction projects via machine learning techniques. Eng Constr architectural Manage. 2023;30(3):1146–64. Saridewi VS, Sari RF. Implementation of machine learning for human aspect in information security awareness. J Appl Eng Sci. 2021;19(4):1126–42. Khalilzadeh M, Shakeri H, Zohrehvandi S. Risk identification and assessment with the fuzzy DEMATEL-ANP method in oil and gas projects under uncertainty. Procedia Comput Sci. 2021;181:277–84. Sun Z-Y, Zhou J-L, Gan L-F. Safety assessment in oil drilling work system based on empirical study and Analytic Network Process. Saf Sci. 2018;105:86–97. Kyrkou C et al. Machine learning for emergency management: A survey and future outlook. Proceedings of the IEEE, 2022. 111(1): pp. 19–41. Phark C, et al. Prediction of issuance of emergency evacuation orders for chemical accidents using machine learning algorithm. J Loss Prev Process Ind. 2018;56:162–9. Richardson N. Emergency Response Planning: Leveraging Machine Learning for Real-Time Decision-Making. Emergency. 2021;4:14. Damaševičius R, Bacanin N, Misra S. From sensors to safety: Internet of Emergency Services (IoES) for emergency response and disaster management. J Sens actuator networks. 2023;12(3):41. Khan M et al. A machine learning driven automated system to extract multiple information fields from safety data sheet documents. Heliyon, 2025. 11(4). Fenton K, Simske S. Engineering of an artificial intelligence safety data sheet document processing system for environmental, health, and safety compliance . in Proceedings of the 21st ACM Symposium on Document Engineering . 2021. Trivedi P, et al. A hybrid best-worst method (BWM)—technique for order of preference by similarity to ideal solution (TOPSIS) approach for prioritizing road safety improvements. IEEe Access. 2024;12:30054–65. Dahl Ø, Kongsvik T. Safety climate and mindful safety practices in the oil and gas industry. J Saf Res. 2018;64:29–36. Lingard H, Pirzadeh P. Workplace health and safety performance at the client-contractor interface: measurement, management and behaviour. Saf Sci. 2025;184:106753. Rshidi S, et al. Ranking Key Performance Indicators of Health, Safety, Environment, and Energy Education Using Multi-criteria Decision-making Techniques. J Occup Hygiene Eng. 2019;6(1):26–34. Ahmed GA. Development of a health safety and environment (HSE) performance review. Methodology for the oil and gas industry in Libya . 2016. Nassiri P, et al. Health, safety, and environmental management system operation in contracting companies: A case study. Arch Environ Occup Health. 2016;71(3):178–85. Hadidi LA, Khater MA. Loss prevention in turnaround maintenance projects by selecting contractors based on safety criteria using the analytic hierarchy process (AHP). J Loss Prev Process Ind. 2015;34:115–26. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 09 Feb, 2026 Reviews received at journal 01 Feb, 2026 Reviews received at journal 28 Jan, 2026 Reviewers agreed at journal 27 Jan, 2026 Reviewers agreed at journal 27 Jan, 2026 Reviewers invited by journal 27 Jan, 2026 Editor assigned by journal 21 Jan, 2026 Submission checks completed at journal 21 Jan, 2026 First submitted to journal 19 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-8639261\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":581725081,\"identity\":\"e3ca880b-2291-4f13-88e0-9a9d5bb71c4c\",\"order_by\":0,\"name\":\"Payam Khordoustan\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Tabriz University of Medical Sciences\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Payam\",\"middleName\":\"\",\"lastName\":\"Khordoustan\",\"suffix\":\"\"},{\"id\":581725082,\"identity\":\"1d47eff6-76b6-4f43-8493-bd897902cf9d\",\"order_by\":1,\"name\":\"Omid Akbarzadeh\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Tabriz University of Medical Sciences\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Omid\",\"middleName\":\"\",\"lastName\":\"Akbarzadeh\",\"suffix\":\"\"},{\"id\":581725083,\"identity\":\"07fa90cb-aae5-4188-99ae-d22236ffb24e\",\"order_by\":2,\"name\":\"Neda Gilani\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Tabriz University of Medical Sciences\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Neda\",\"middleName\":\"\",\"lastName\":\"Gilani\",\"suffix\":\"\"},{\"id\":581725084,\"identity\":\"110cb2ba-023a-4599-beee-27cc64d1cd11\",\"order_by\":3,\"name\":\"Parisa Moshashaei\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"University of Canberra\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Parisa\",\"middleName\":\"\",\"lastName\":\"Moshashaei\",\"suffix\":\"\"},{\"id\":581725085,\"identity\":\"95c72c4a-8cc7-4ffc-83a6-53db4ff026d9\",\"order_by\":4,\"name\":\"Seyed Shamseddin Alizadeh\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBElEQVRIiWNgGAWjYDACdsYGBgY2hgQGCeYDzGCRAyDCAI8WZrgWtsRmIrWACLAWHkMkLXgAfzNz84cfZTZ58rN7vj8u+LNNju8A88MPDAX3cGqROMzYYNhzLq3Y4M7Zjc0zeG4bSx5gM5ZgMCjGbQ1QSwJv2+HEDRK5G5t5JG4nbjjAYAb0SwJOHfJALQf/tv1PnD8j52Ezj8Ht+g0H2L/h1WJwmLGxmbftQGLDjRzGZp6E2wkGB3jw22J4mLGZWeZccuKGG2mGs2ccuG048zBPsUQCHi1yx9sff3xTZgd0WPKDzwV/bsvzHW/f+OHDH9xasABQTJGkYRSMglEwCkYBBgAA6lZamBSfluoAAAAASUVORK5CYII=\",\"orcid\":\"\",\"institution\":\"Tabriz University of Medical Sciences\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Seyed\",\"middleName\":\"Shamseddin\",\"lastName\":\"Alizadeh\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2026-01-19 12:08:42\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-8639261/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-8639261/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":101433874,\"identity\":\"a967a502-75fe-4707-a290-98b8d41917c8\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":97958,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eTheoretical and Methodological Foundations of CHPI Development\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage1.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/67bce987d1554c8436a89968.png\"},{\"id\":101433875,\"identity\":\"19b0fde3-be68-4fd6-8a76-7ba6f19156a5\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"png\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":120765,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eFlowchart of the CHPI development and validation process\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage2.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/dcf202b90e051c11fed72a4a.png\"},{\"id\":101433877,\"identity\":\"7c152cc9-2826-4e72-b208-a870b8d88915\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"png\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":70358,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eIndicator development and reduction process for the CHPI framework\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage3.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/4e3e7c100092e549be1d974c.png\"},{\"id\":101433878,\"identity\":\"aa90d81b-84bc-4165-908f-d815c2cf108f\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"jpeg\",\"order_by\":4,\"title\":\"Figure 4\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":849154,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003ePairwise Comparison by Super Decision Software\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage4.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/62d18e2fef569d1f6240d95c.jpeg\"},{\"id\":101433879,\"identity\":\"6fe7a1f1-eb5e-4928-8137-8f0831cb50fd\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"jpeg\",\"order_by\":5,\"title\":\"Figure 5\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":256541,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003erelationship between actual and predicted CHPI values\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage5.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/ee16342d4205014dc9e32100.jpeg\"},{\"id\":101433876,\"identity\":\"e75f78aa-de8d-497e-8f41-209245568ef8\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:39\",\"extension\":\"jpeg\",\"order_by\":6,\"title\":\"Figure 6\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":339401,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eTheoretical-to-Operational Mapping of the CHPI Framework\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage6.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/b8ff20ae3f306b93cb63d94a.jpeg\"},{\"id\":101433896,\"identity\":\"dc4ca758-bd59-4f32-8929-c5145da25ac4\",\"added_by\":\"auto\",\"created_at\":\"2026-01-29 16:01:53\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":3315634,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-8639261/v1/d9018f1a-3564-4df0-ba1f-185ec88aff23.pdf\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Developing the Contractor Health, Safety, and Risk Performance Index: A Hybrid ANP–Machine Learning Approach\",\"fulltext\":[{\"header\":\"1. Introduction\",\"content\":\"\\u003cdiv id=\\\"Sec2\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e1.1. Evolution of HSE Performance Indicators\\u003c/h2\\u003e \\u003cp\\u003eThe oil and gas industry are widely recognized as one of the most hazardous sectors due to its dependence on complex technical systems, high-pressure processes, and the use of flammable or toxic materials [\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e]. Accidents in this sector often have severe human, environmental, and economic consequences, making robust workplace health, safety and environment (WHSE) management not merely a regulatory obligation but also a strategic necessity for organizational resilience [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. Within this context, contractors play a decisive role in determining overall WHSE performance [\\u003cspan citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e].Nevertheless, empirical studies have shown that contractor compliance varies widely due to differences in safety culture, training quality, and enforcement mechanisms [\\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e]. This variability creates a strong need for more systematic, data-driven approaches to monitoring and improving contractor WHSE performance. Additionally, contractor-related issues contribute to WHSE failures through multiple organizational and operational pathways. Contractors often exhibit heterogeneous safety cultures, varying levels of training, and inconsistent adherence to host organization procedures, which can weaken control mechanisms at the site level [\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e]. High workforce turnover, fragmented communication channels, and misalignment between contractual pressures and safety priorities further exacerbate these risks [\\u003cspan citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e]. In the oil and gas industry, such vulnerabilities are amplified due to the inherent presence of hazardous materials, high-pressure systems, complex process technologies, and tightly coupled operations where minor deviations can escalate into major incidents [\\u003cspan citationid=\\\"CR7\\\" class=\\\"CitationRef\\\"\\u003e7\\u003c/span\\u003e]. As a result, deficiencies in contractor oversight, competency management, and real-time monitoring play a disproportionate role in WHSE failures, making contractor performance evaluation a critical component of risk management in high-risk oil and gas environments. Against this backdrop, improving contractor WHSE performance requires not only effective governance mechanisms but also robust performance measurement systems, making the evolution of WHSE performance indicators a central concern in high-risk industries such as oil and gas.\\u003c/p\\u003e \\u003cp\\u003eOver the past two decades, numerous studies have attempted to develop frameworks for measuring and monitoring WHSE performance in high-risk industries. Early approaches relied heavily on lagging indicators such as lost-time injury frequency rates (LTIFR) [\\u003cspan citationid=\\\"CR8\\\" class=\\\"CitationRef\\\"\\u003e8\\u003c/span\\u003e] and total recordable incident rates (TRIR) [\\u003cspan citationid=\\\"CR9\\\" class=\\\"CitationRef\\\"\\u003e9\\u003c/span\\u003e]. While these metrics remain useful for compliance verification, they offer limited predictive insight and cannot adequately identify emerging risks [\\u003cspan citationid=\\\"CR10\\\" class=\\\"CitationRef\\\"\\u003e10\\u003c/span\\u003e] before incidents occur. In response to these limitations, researchers have increasingly emphasized the role of leading indicators, proactive measures [\\u003cspan citationid=\\\"CR11\\\" class=\\\"CitationRef\\\"\\u003e11\\u003c/span\\u003e] that signal deteriorating safety conditions before they result in accidents. Examples include hazard identification rates, frequency of safety audits, participation in safety training, and the effectiveness of emergency preparedness drills [\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e]. In the Iranian oil and gas sector, Sarkheil and Rahbari (2016) [\\u003cspan citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e] and Amir-Heidari et al. (2017) [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e] demonstrated that integrating leading indicators into WHSE management systems (HSE-MS) can enhance sustainability and long-term safety performance. Azadeh et al. (2014) [\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e] used multivariate analysis to continuously assess integrated WHSE and maintenance systems in gas transmission units, showing how statistical models can identify performance improvement opportunities. However, despite these advances, many indicators set remain context-specific [\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e] and lack the flexibility for cross-sector or cross-country application [\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e]. This points to a need for indicator frameworks that are both adaptable to local conditions and aligned with international best practices.\\u003c/p\\u003e \\u003cp\\u003eDespite the growing body of research on WHSE performance indicators in high-risk industries, existing studies reveal several unresolved limitations that motivate the present research. Prior efforts have predominantly focused on developing context-specific indicator sets or evaluating isolated safety dimensions within single organizational or operational settings. For example, studies conducted in Iranian oil and gas and chemical industries have demonstrated the value of active and leading indicators in improving safety outcomes through administrative interventions, integrated management systems, and conceptual performance models [\\u003cspan additionalcitationids=\\\"CR19 CR20\\\" citationid=\\\"CR18\\\" class=\\\"CitationRef\\\"\\u003e18\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR21\\\" class=\\\"CitationRef\\\"\\u003e21\\u003c/span\\u003e]. While these contributions have advanced understanding of WHSE measurement at the organizational level, they often rely on static indicator structures, limited interdependency modelling, and descriptive or sector-bound validation approaches. Moreover, contractor safety performance, characterized by fragmented responsibilities, heterogeneous safety cultures, and dynamic operational conditions, remains underexplored within these frameworks. These gaps highlight the need for a more integrative, empirically validated approach that captures the systemic interactions among WHSE indicators and supports consistent contractor performance evaluation in complex, high-risk environments. Addressing this need provides the primary motivation for developing the CHPI framework proposed in this study. In this context, the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI) is conceptualized as an integrated and composite framework for evaluating contractor WHSE performance in high-risk industrial settings. The CHPI systematically aggregates validated safety indicators across behavioural, organizational, and operational domains, explicitly accounting for interdependencies among indicators through network-based weighting. By combining expert-informed prioritization with empirical performance data and predictive validation, the CHPI enables consistent benchmarking, risk-based contractor differentiation, and evidence-driven decision-making. This unified definition establishes CHPI as a performance-oriented evaluation model rather than a descriptive indicator set, providing a coherent foundation for the methodological approach adopted in this study.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec3\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e1.2 Contractor HSEQ Performance in Relation to Organizational Safety Governance\\u003c/h2\\u003e \\u003cp\\u003eContractors play a central role in shaping organizational HSEQ performance in high-risk industries, as a substantial proportion of operational activities are outsourced to external firms. Although contractors operate under the oversight of the host organization, their safety practices, workforce competency, and environmental controls often vary considerably, creating systemic vulnerabilities within organizational HSEQ governance [\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e]. Deficiencies in contractor performance can directly undermine organizational safety objectives, increase incident likelihood, and expose organizations to regulatory, reputational, and financial risks [\\u003cspan citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e]. In contractor-intensive sectors such as oil and gas, the organizational responsibility for HSEQ outcomes extends beyond internal systems to include the effective selection, monitoring, and control of contractor performance [\\u003cspan citationid=\\\"CR22\\\" class=\\\"CitationRef\\\"\\u003e22\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eTraditional organizational HSEQ management systems frequently rely on compliance-based checklists or minimum qualification requirements, which provide limited insight into how contractor practices interact with organizational controls and risk management mechanisms. This disconnect reduces the organization\\u0026rsquo;s ability to proactively identify underperforming contractors, prioritize interventions, and align contractor behaviour with organizational safety expectations [\\u003cspan citationid=\\\"CR23\\\" class=\\\"CitationRef\\\"\\u003e23\\u003c/span\\u003e]. Accordingly, effective organizational HSEQ governance requires structured, transparent, and performance-oriented mechanisms capable of capturing the multidimensional and interdependent nature of contractor activities [\\u003cspan citationid=\\\"CR24\\\" class=\\\"CitationRef\\\"\\u003e24\\u003c/span\\u003e]. An integrated contractor performance index, aligned with organizational safety objectives and validated using empirical data, offers a practical pathway for strengthening contractor oversight, supporting evidence-based decision-making, and enhancing overall organizational HSEQ performance. The CHPI framework proposed in this study is developed within this organizational context, explicitly addressing the need to evaluate contractor HSEQ performance as an integral component of organizational risk management and safety governance.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec4\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e1.3 Emerging Role of Machine Learning in HSE Assessment\\u003c/h2\\u003e \\u003cp\\u003eGiven the multifaceted nature of WHSE performance, researchers have increasingly turned to multi-criteria decision-making (MCDM) methods [\\u003cspan citationid=\\\"CR25\\\" class=\\\"CitationRef\\\"\\u003e25\\u003c/span\\u003e] for evaluating and prioritizing safety indicators. The ANP [\\u003cspan citationid=\\\"CR26\\\" class=\\\"CitationRef\\\"\\u003e26\\u003c/span\\u003e] has emerged as a particularly valuable tool because it can capture interdependencies and feedback loops among indicators, relationships often oversimplified in hierarchical methods like the Analytic Hierarchy Process (AHP) [\\u003cspan citationid=\\\"CR27\\\" class=\\\"CitationRef\\\"\\u003e27\\u003c/span\\u003e]. Applications of ANP in the oil and gas sector have included contractor selection models [\\u003cspan citationid=\\\"CR28\\\" class=\\\"CitationRef\\\"\\u003e28\\u003c/span\\u003e], risk prioritization frameworks [\\u003cspan citationid=\\\"CR29\\\" class=\\\"CitationRef\\\"\\u003e29\\u003c/span\\u003e], and safety system evaluations [\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e], for example, applied Fuzzy ANP to prioritize WHSE performance indicators in large industrial organizations [\\u003cspan citationid=\\\"CR31\\\" class=\\\"CitationRef\\\"\\u003e31\\u003c/span\\u003e], revealing that competency-related measures often influence other domains such as risk control and emergency preparedness. Yet most of these studies focus on pre-contract evaluation or narrow organizational contexts, with limited application to ongoing contractor performance monitoring during active projects. Moreover, very few combine ANP weighting with empirical performance data to validate indicator relevance in practice.\\u003c/p\\u003e \\u003cp\\u003eRecent advances in statistical modelling and machine learning have highlighted the growing role of data-driven approaches in forecasting complex, high-uncertainty phenomena. Gul et al. (2025) [\\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e] applied a hybrid framework combining classical statistical modelling and machine learning techniques to predict COVID-19 mortality trends in Pakistan. Using a Gumbel\\u0026ndash;Truncated Exponential Distribution estimated via maximum likelihood methods alongside machine learning algorithms, their study demonstrated that ML-based models consistently outperformed traditional statistical approaches in terms of predictive accuracy, as measured by MAE, RMSE, and MAPE. These findings underscore the ability of machine learning to capture nonlinear patterns and latent relationships that are difficult to model using purely parametric techniques. Beyond the pandemic context, this evidence supports the broader applicability of hybrid statistical\\u0026ndash;ML frameworks for predictive analysis in safety-critical and high-risk domains, where uncertainty, data heterogeneity, and dynamic system behaviour are prevalent. Accordingly, integrating machine learning\\u0026ndash;based validation alongside structured analytical methods can enhance the robustness and practical relevance of performance assessment models [\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e]. Complementary evidence from occupational safety research further reinforces the value of machine learning for decision support in safety-critical environments. Koklonis et al. (2021) [\\u003cspan citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e] demonstrated that machine learning techniques can effectively support occupational safety and health decision-making in hospital workplaces by identifying patterns in incident data and operational conditions that are not readily observable through conventional analysis. Their findings highlight the potential of ML-based models to enhance preventive strategies, improve risk prioritization, and support evidence-informed safety management decisions. Together with recent hybrid statistical\\u0026ndash;ML studies, this work supports the integration of machine learning as a validation and decision-support layer within structured safety performance assessment frameworks. Further evidence on the role of feature selection and machine learning in complex risk modelling is provided by Sarwar et al. (2024), who conducted a comparative analysis of multiple feature selection strategies integrated with hybrid metaheuristic and machine learning models for flood risk assessment. Their results showed that combining intelligent feature selection with ML algorithms significantly improves predictive accuracy and model interpretability in spatial risk analysis. This study highlights the importance of data-driven feature relevance assessment in high-uncertainty and safety-critical domains, supporting the use of machine learning\\u0026ndash;based importance analysis as a complementary validation tool within structured performance assessment frameworks [\\u003cspan citationid=\\\"CR35\\\" class=\\\"CitationRef\\\"\\u003e35\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eAlongside traditional MCDM approaches, machine learning (ML) techniques have recently gained attention for their potential to enhance predictive accuracy in HSE performance evaluation. Algorithms such as Random Forests [\\u003cspan citationid=\\\"CR36\\\" class=\\\"CitationRef\\\"\\u003e36\\u003c/span\\u003e] and Bayesian networks can process large [\\u003cspan citationid=\\\"CR37\\\" class=\\\"CitationRef\\\"\\u003e37\\u003c/span\\u003e], complex datasets to identify nonlinear patterns and rank the relative importance of multiple factors. In process safety, ML has been applied to accident prediction [\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e], hazard identification [\\u003cspan citationid=\\\"CR39\\\" class=\\\"CitationRef\\\"\\u003e39\\u003c/span\\u003e], and offering new opportunities for proactive safety management [\\u003cspan citationid=\\\"CR40\\\" class=\\\"CitationRef\\\"\\u003e40\\u003c/span\\u003e]. However, the integration of ML into structured WHSE evaluation frameworks remains limited, particularly in the context of contractor management. Onukwulu et al. (2024) [\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e] note that existing contractor safety management models are often process-oriented without quantitative prioritization mechanisms, while Lee et al. (2024) [\\u003cspan citationid=\\\"CR42\\\" class=\\\"CitationRef\\\"\\u003e42\\u003c/span\\u003e] highlight the need to better link host organizations\\u0026rsquo; safety climates with measurable contractor performance outcomes. As a result, there is considerable scope for hybrid frameworks that combine expert-based weighting with ML-driven validation to bridge the gap between theoretical prioritization and real-world predictive performance.\\u003c/p\\u003e \\u003cp\\u003eIn summary, the literature reveals three persistent gaps: there is an underutilization of integrated frameworks that combine leading and lagging indicators for contractor WHSE performance monitoring across the full project lifecycle; the application of interdependency-aware weighting methods, such as the ANP, remains limited in operational, contractor-level contexts; and there is a notable absence of systematic integration between multi-criteria decision-making approaches and machine learning techniques for the predictive validation of safety indicators. This study addresses these gaps by developing and validating a CHPI that unites content validation through the Content Validity Index (CVI), interdependent weighting via a Fuzzy ANP approach, and predictive analysis using Random Forest regression. The CHPI is designed as a modular, scalable tool for continuous contractor performance monitoring, with adaptability across various high-risk industries and potential integration into AI-driven WHSE management platforms. Piloting the framework in a large-scale gas sector project demonstrates its operational feasibility and offers a transferable model for enhancing transparency, efficiency, and safety in client\\u0026ndash;contractor relationships worldwide.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec5\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e1.4 Advanced Data-Driven and Hybrid Approaches for Safety Performance Evaluation\\u003c/h2\\u003e \\u003cp\\u003ePrevious studies on contractor safety evaluation have primarily focused on framework-based or indicator-driven assessment models. Early works proposed structured approaches for evaluating contractor safety performance using predefined indicators and expert judgment, with an emphasis on compliance, management practices, and historical safety outcomes [\\u003cspan citationid=\\\"CR43\\\" class=\\\"CitationRef\\\"\\u003e43\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR44\\\" class=\\\"CitationRef\\\"\\u003e44\\u003c/span\\u003e]. Subsequent research extended these approaches by incorporating multi-criteria decision-making techniques to assess safety maturity and the quality of occupational health and safety management systems, highlighting the importance of systematic prioritization across safety dimensions [\\u003cspan citationid=\\\"CR45\\\" class=\\\"CitationRef\\\"\\u003e45\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR46\\\" class=\\\"CitationRef\\\"\\u003e46\\u003c/span\\u003e]. In parallel, a growing body of research has explored data-driven and machine learning\\u0026ndash;based methods for construction safety prediction and risk assessment, demonstrating their ability to capture complex, non-linear relationships in accident and safety data [\\u003cspan citationid=\\\"CR47\\\" class=\\\"CitationRef\\\"\\u003e47\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR48\\\" class=\\\"CitationRef\\\"\\u003e48\\u003c/span\\u003e]. Systematic reviews and conceptual studies further emphasize the potential of machine learning to support proactive and intelligent safety management in construction contexts [\\u003cspan citationid=\\\"CR49\\\" class=\\\"CitationRef\\\"\\u003e49\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eHowever, these two research streams have largely evolved in isolation. Existing framework-based models often rely on expert judgment or predefined scoring structures without empirically validating their indicators against real operational performance data, limiting their ability to support proactive and predictive safety management. Conversely, machine learning\\u0026ndash;based safety studies typically focus on pattern recognition and outcome prediction, but operate without structured, theory-informed indicator frameworks and rarely account for the interdependencies among safety dimensions. As a result, these approaches may achieve predictive accuracy while offering limited interpretability and decision relevance for safety governance. This methodological disconnect underscores the need for integrated approaches that combine interdependency-aware indicator weighting with data-driven validation, enabling both conceptual rigor and empirical robustness. The CHPI framework proposed in this study is developed in direct response to this gap, uniting expert-informed structural modelling with machine learning\\u0026ndash;based predictive assessment to support more reliable and actionable contractor WHSE performance evaluation.\\u003c/p\\u003e \\u003cp\\u003eDespite extensive advances in safety performance assessment, existing contractor evaluation approaches still fail to provide decision-makers with a defensible basis for prioritizing safety interventions under real operational conditions. Framework-based models typically generate static scores without demonstrating predictive relevance, while machine learning models often identify influential factors without explaining their structural role within safety systems. This disconnect leaves regulators and asset owners unable to justify why certain contractors, indicators, or control measures should be prioritized over others. Addressing this unresolved problem requires an evaluation framework that is simultaneously theory-grounded, empirically validated, and operationally actionable. Overall, prior studies demonstrate the increasing application of statistical, multi-criteria, and machine learning approaches for performance evaluation and risk prediction in safety-critical domains. Existing research confirms the value of leading indicators, expert-based weighting methods, and data-driven models in supporting proactive safety decision-making. However, the literature also reveals three persistent gaps. First, there is an underutilization of integrated frameworks that combine leading and lagging indicators for contractor WHSE performance monitoring across the full project lifecycle. Second, the application of interdependency-aware weighting methods, such as the ANP, remains limited in operational, contractor-level contexts. Third, there is a notable absence of systematic integration between multi-criteria decision-making approaches and machine learning techniques for the predictive validation of safety indicators. Addressing these gaps, this study develops and validates the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI), which unites content validation through the Content Validity Index (CVI), interdependent weighting via a Fuzzy ANP approach, and predictive analysis using Random Forest regression. The CHPI is designed as a modular and scalable tool for continuous contractor performance monitoring, with adaptability across various high-risk industries and potential integration into AI-driven WHSE management platforms. Piloting the framework in a large-scale gas sector project demonstrates its operational feasibility and offers a transferable model for enhancing transparency, efficiency, and safety in client\\u0026ndash;contractor relationships worldwide.\\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"2. Theoretical Framework\",\"content\":\"\\u003cp\\u003eThe development of the CHPI in this study is grounded in two complementary theoretical perspectives: Safety Performance Theory and Socio-Technical Systems (STS) Theory [\\u003cspan citationid=\\\"CR50\\\" class=\\\"CitationRef\\\"\\u003e50\\u003c/span\\u003e], supported by principles of MCDM (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e). Together, these frameworks provide a conceptual foundation for understanding and evaluating contractor safety performance in high-risk industrial environments such as the oil and gas sector.\\u003c/p\\u003e \\u003cp\\u003eSafety Performance Theory [\\u003cspan citationid=\\\"CR44\\\" class=\\\"CitationRef\\\"\\u003e44\\u003c/span\\u003e] posits that safety outcomes are the result of the dynamic interaction between organizational systems, individual behaviours, and environmental conditions. Within contractor-heavy operations, safety performance is shaped not only by compliance with technical standards [\\u003cspan citationid=\\\"CR51\\\" class=\\\"CitationRef\\\"\\u003e51\\u003c/span\\u003e] but also by proactive measures such as hazard identification, competency development, and effective emergency preparedness [\\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e]. This theoretical lens underscores the importance of incorporating both leading indicators and lagging indicators into performance assessment tools. The CHPI operationalizes this principle by integrating a validated set of indicators that capture preventive, behavioural, and outcome-based dimensions of safety.\\u003c/p\\u003e \\u003cp\\u003eSocio-Technical Systems Theory conceptualizes organizations as interdependent networks of human and technical subsystems, where safety emerges from the coordinated functioning of these components [\\u003cspan citationid=\\\"CR53\\\" class=\\\"CitationRef\\\"\\u003e53\\u003c/span\\u003e]. In the context of contractor management, this means that factors such as communication quality, leadership engagement, and documentation processes [\\u003cspan citationid=\\\"CR54\\\" class=\\\"CitationRef\\\"\\u003e54\\u003c/span\\u003e] are closely linked to technical controls like equipment maintenance, hazard monitoring, and emergency systems. The CHPI framework reflects this systemic interdependence by using the ANP to model feedback loops and cross-domain influences among WHSE indicators, enabling a more realistic representation of contractor performance dynamics compared to linear or hierarchical models.\\u003c/p\\u003e \\u003cp\\u003eFinally, the integration of MCDM principles with data-driven validation [\\u003cspan citationid=\\\"CR55\\\" class=\\\"CitationRef\\\"\\u003e55\\u003c/span\\u003e] responds to recent calls for more robust and predictive safety assessment models. The CHPI applies Fuzzy ANP to capture expert judgments under uncertainty, while machine learning serves as a secondary validation layer to evaluate the predictive relevance of each indicator. This hybrid approach aligns with the trend towards intelligent [\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e], adaptive safety management systems [\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e] capable of continuous monitoring and early risk detection. In sum, the theoretical foundation of CHPI bridges behavioural, organizational, and technical dimensions of safety performance, aligns with established safety science frameworks, and incorporates advanced analytical methods to enhance both the validity and operational applicability of contractor performance evaluation.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e\"},{\"header\":\"3. Methods\",\"content\":\"\\u003cdiv id=\\\"Sec8\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.1 Research Design\\u003c/h2\\u003e \\u003cp\\u003eThis chapter outlines the methodological framework adopted to develop and validate the CHPI. It describes the overall research design, indicator identification and validation process, data collection procedures, analytical techniques, and validation strategy. The methodological approach was structured to ensure conceptual rigor, empirical robustness, and alignment with the study\\u0026rsquo;s objective of integrating expert-based multi-criteria decision-making with data-driven machine learning validation.\\u003c/p\\u003e \\u003cp\\u003eThis study employed a cross-sectional, multi-method design to evaluate the WHSE performance of 72 contractors operating under the East Azerbaijan Gas Company (EAGC) during the 2022\\u0026ndash;2023 fiscal years. The methodological framework integrated expert judgment, archival performance data, and advanced modelling to develop and validate the CHPI (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e). The process involved three sequential stages: prioritizing interdependent WHSE indicators using the Fuzzy Analytic Network Process (F-ANP) to obtain domain-specific and global weights; standardizing archival contractor compliance records and combining them with these weights to compute composite CHPI scores; and applying a Random Forest regression model to assess the predictive validity of the CHPI framework. Ethical approval for the study was granted by the Ethics Committee of Tabriz University of Medical Sciences (IR.TBZMED.REC.1398.416) and the Research Coordination Office of EAGC.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec9\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.2 Identification of Candidate WHSE Indicators\\u003c/h2\\u003e \\u003cp\\u003eThe development of a comprehensive pool of measures tailored to contractor WHSE performance was guided by a structured indicator-scoping process that integrated multiple sources of evidence. This process drew on international WHSE guidelines issued by recognized bodies such as the Occupational Safety and Health Administration (OSHA), the International Labour Organization (ILO), and the American Petroleum Institute (API), alongside peer-reviewed academic literature on safety performance metrics in high-risk industries, and internal policy documents and compliance protocols from the EAGC and the National Iranian Gas Company (NIGC). Both English and Persian materials published between 2005 and 2024 were examined, with selection criteria focusing on indicator relevance, field observability, and compatibility with regional regulatory requirements. Literature-based indicators were identified through targeted keyword searches, including terms such as \\u0026ldquo;contractor WHSE indicators,\\u0026rdquo; \\u0026ldquo;occupational risk metrics,\\u0026rdquo; and \\u0026ldquo;compliance performance,\\u0026rdquo; while internal documentation contributed additional measures rooted in local operational realities; these were retained only when they demonstrated both practical applicability and conceptual consistency with established safety frameworks. The outcome of this scoping exercise was an initial set of 84 candidate indicators, each corresponding to a distinct WHSE construct, which were subsequently organized into seven preliminary domains thereby establishing a coherent foundation for systematic expert review and subsequent ANP modelling.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec10\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.3 Data Collection\\u003c/h2\\u003e \\u003cp\\u003eExperts were selected using purposive sampling based on predefined eligibility criteria to ensure domain relevance and professional credibility. Inclusion criteria comprised: (i) a minimum of 8 years of professional experience in WHSE management, occupational safety, environmental health, or process risk assessment; (ii) at least a bachelor\\u0026rsquo;s degree in a relevant field such as occupational health, safety engineering, environmental engineering, or industrial management; (iii) direct involvement in contractor oversight, HSE auditing, or safety system implementation in high-risk industries; and (iv) demonstrated familiarity with WHSE management systems and performance evaluation frameworks. These criteria were applied to ensure that expert judgments reflected both theoretical knowledge and practical, field-based experience.\\u003c/p\\u003e \\u003cp\\u003eTo ensure the conceptual rigor, clarity, and operational relevance of the proposed contractor-level WHSE performance indicators, a structured content validity assessment was first conducted prior to their inclusion in the ANP framework. A panel of 14 subject-matter experts, averaging 10.8 years of professional experience (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e), evaluated 84 candidate indicators across three dimensions: relevance to contractor WHSE performance, clarity of definition, and feasibility of implementation and monitoring. Using a 9-point rating scale and Lawshe\\u0026rsquo;s method, Content Validity Ratios (CVRs) were calculated, with 0.51 set as the minimum acceptable value for the 14-member panel. Of the 84 indicators, 81 met or exceeded this threshold and were retained; the remaining three were revised based on expert feedback and re-evaluated. Following indicator validation, a dual-source data collection strategy was adopted to generate the empirical dataset required for ANP weighting and CHPI computation. The first stream involved expert judgment for prioritizing interdependent WHSE indicators within the F-ANP framework. The same expert panel provided pairwise comparisons using Saaty\\u0026rsquo;s 1\\u0026ndash;9 scale, covering both intra-cluster and inter-cluster relationships among the 43 validated sub-indicators. Responses were obtained through in-person meetings or electronic questionnaires and aggregated using the geometric mean method to construct fuzzy pairwise comparison matrices. Consistency Ratios (CRs) were computed in Super Decisions (v3.2) to verify logical coherence, with matrices exceeding 0.10 flagged for revision. Two refinement rounds were conducted, resulting in an average CR of 0.054, indicating strong agreement.\\u003c/p\\u003e \\u003cp\\u003eThe second stream involved collecting empirical compliance records from 26 contractor organizations operating under the EAGC in infrastructure, pipeline maintenance, and industrial service projects. Archival documentation was systematically reviewed, with each record mapped to its relevant WHSE sub-indicator and scored on a standardized 0\\u0026ndash;100 scale. Binary indicators were scored 0 or 100, while quantitative measures were normalized via linear transformation. Missing entries were scored as zero unless clarified via follow-up with site personnel. Two independent analysts assessed the records, resolving discrepancies by consensus. The resulting normalized compliance scores were integrated with the F-ANP-derived weights to compute CHPI scores for each contractor. This dataset was also used to train a Random Forest regression model, supporting the evaluation of the CHPI framework\\u0026rsquo;s predictive validity, and reinforcing its empirical robustness.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab1\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 1\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eProfile of Experts Involved in Content Validity Assessment\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"3\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eExpert Role\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eNumber of Experts (n)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eMean Years of Experience\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eProject-Level WHSE Managers\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e11.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eSenior Safety Officers\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e10.2\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEnvironmental Health Professionals\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e11.0\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eProcess Risk Assessors\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e9.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eCertified HSE Auditors\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e12.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eTotal\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e14\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e10.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec11\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.4 Indicator Development and Reduction Process\\u003c/h2\\u003e \\u003cp\\u003eThe development of the CHPI indicators followed a structured, multi-stage refinement process to ensure content validity, relevance, and operational applicability (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e). Initially, a comprehensive pool of 84 candidate indicators was compiled based on an extensive review of prior contractor safety evaluation studies and WHSE standards. In the first screening stage, redundancy and semantic overlap were addressed through expert review, resulting in the removal of three overlapping indicators and yielding a refined set of 81 indicators. In the second stage, content validity was assessed using the CVR and CVI. A CVR threshold of 0.51 was applied in accordance with Lawshe\\u0026rsquo;s table for a panel of 14 experts, ensuring that retained indicators were considered essential by most experts. Subsequently, indicators meeting a CVI threshold of 0.78 were retained to confirm clarity, relevance, and representativeness. This dual filtering process reduced the indicator set to 43 validated indicators. In the final stage, the retained indicators were conceptually grouped based on functional similarity and system-level interactions, resulting in seven thematic clusters: Awareness and Competency, Risk Management, Emergency Preparedness, Environmental Health Management, Documentation, Monitoring and Audit, and WHSE Management Systems. This seven-cluster structure was consistently adopted throughout the ANP modelling and empirical validation phases and represents the final organizational structure of the CHPI framework.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec12\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.5 ANP Network Construction and Weight Derivation\\u003c/h2\\u003e \\u003cp\\u003eGiven the interdependent nature of contractor WHSE indicators, a hierarchical method such as AHP was considered insufficient. Therefore, ANP was adopted to capture feedback and mutual influences among indicators. Based on the validated indicators, the decision network was organized into six clusters: Awareness and Competency, Documentation, Emergency Preparedness, Environmental Health Management, WHSE Management Systems, and Monitoring and Audit, collectively representing the key organizational, procedural, and operational dimensions of contractor safety performance. The evaluation of contractor WHSE performance involves multiple criteria that are inherently interdependent and dynamically linked rather than hierarchically independent. In such contexts, traditional hierarchical methods such as the AHP, which assume unidirectional relationships and independence among criteria, may oversimplify the underlying safety structure [\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e]. Also, the ANP is specifically designed to capture these complex interrelationships by allowing feedback loops and interdependencies among criteria [\\u003cspan citationid=\\\"CR58\\\" class=\\\"CitationRef\\\"\\u003e58\\u003c/span\\u003e]. Given the strong interdependencies among contractor WHSE indicators, a hierarchical decision structure was deemed insufficient; therefore, ANP was adopted to explicitly model feedback and mutual influence among safety dimensions.\\u003c/p\\u003e \\u003cp\\u003eTo quantify indicator interdependencies, a structured pairwise comparison questionnaire was developed in accordance with the ANP framework, capturing both intra- and inter-cluster influences (Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e). Expert judgments were processed using Super Decisions v3.2, with all consistency ratios remaining below the acceptable threshold of 0.10, confirming the reliability of the comparisons. ANP calculations were then performed to derive global priority weights through successive construction of the unweighted, weighted, and limit super-matrices. The results (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e) indicate that Awareness and Competency held the highest global priority (0.358), followed by WHSE Management Systems (0.148), while Documentation (0.086) and Environmental Health Management (0.093) exhibited comparatively lower influence on contractor WHSE performance differentiation.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003eTo ensure the consistency of expert judgments in the ANP process, CRs were calculated for all pairwise comparison matrices using Super Decisions software. A CR threshold of 0.10 was adopted as the acceptance criterion. Matrices exceeding this threshold were returned to experts for review and revision, and iterative refinement was performed until acceptable consistency was achieved. Despite these controls, experts may encounter practical challenges during ANP pairwise comparisons, including cognitive burden due to the number of comparisons, difficulty in judging interdependent relationships, and uncertainty when comparing qualitatively different safety indicators. To mitigate these challenges, experts were provided with clear definitions of indicators, structured comparison templates, and opportunities for clarification, which helped reduce ambiguity and improve judgment coherence.\\u003c/p\\u003e \\u003cp\\u003eTo address uncertainty and subjectivity in expert judgments, the ANP procedure was implemented using a fuzzy logic extension. Expert pairwise comparisons were expressed using triangular fuzzy numbers (TFNs), allowing linguistic preferences to be represented as bounded ranges rather than precise values. Fuzzification was performed by mapping linguistic importance scales to TFNs, and individual expert judgments were aggregated using the geometric mean to obtain consolidated fuzzy comparison matrices. Defuzzification was then conducted using the centroid method to derive crisp priorities suitable for ANP computation. All subsequent calculations, including consistency assessment and weight derivation, were performed using Super Decisions software, which ensured methodological rigor and computational reliability without requiring manual illustration of intermediate matrices.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab2\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 2\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eRaw, Normalized, and Ideal Priority Values and Limit Super-Matrix for HSE Clusters\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eHSE Cluster\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eRaw Score\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eNormalized\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eIdeal\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eGlobal Priority\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAwareness \\u0026amp; Competency\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.358\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.358\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e1.000\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.358\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.086\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.086\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.240\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.086\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEmergency Preparedness\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.097\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.097\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.271\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.097\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEnvironmental Health Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.093\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.093\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.260\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.093\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eManagement-System Existence\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.148\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.148\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.413\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.148\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMonitoring \\u0026amp; Audit\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.118\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.118\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.330\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.118\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec13\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e3.6 Empirical Evaluation of Contractor HSE Data\\u003c/h2\\u003e \\u003cp\\u003eFollowing weight derivation, the CHPI framework was empirically evaluated to assess its applicability and discriminative power across contractors. ANP-weighted indicators were used to compute composite WHSE scores, enabling benchmarking and revealing substantial variability in compliance, particularly in Emergency Preparedness and Documentation. Predictive validity was first examined using correlation and regression analyses, which showed that Awareness and Competency and WHSE Management Systems were most strongly associated with overall CHPI scores, consistent with expert prioritization. To further validate robustness, a Random Forest regression model was applied using normalized domain scores from 20 contractors. The model demonstrated strong predictive performance based on R\\u0026sup2;, MAE, and RMSE, and feature importance rankings closely aligned with ANP-derived global weights, confirming consistency between expert judgment and data-driven validation.\\u003c/p\\u003e \\u003cp\\u003eFollowing weight derivation, the CHPI framework was empirically evaluated to assess its applicability and discriminative capacity across contractors. ANP-weighted indicators were integrated with normalized archival compliance data to compute composite WHSE scores, enabling systematic benchmarking and revealing substantial variability across safety dimensions, particularly in Emergency Preparedness and Documentation. Predictive validity was initially examined using correlation and regression analyses, which indicated that Awareness and Competency and WHSE Management Systems exhibited the strongest associations with overall CHPI scores, consistent with expert-derived prioritization. To further examine robustness and empirical coherence, a Random Forest regression model was employed as a complementary validation tool rather than a standalone predictive model.\\u003c/p\\u003e \\u003cp\\u003eDue to data completeness requirements, only contractors with fully available and standardized archival records were included in the machine learning analysis (n\\u0026thinsp;=\\u0026thinsp;20). To mitigate overfitting risk under small-sample conditions, a 10-fold cross-validation strategy was adopted, and model performance metrics (R\\u0026sup2;, MAE, and RMSE) were computed as averages across folds. The Random Forest model was implemented using 500 trees with bootstrap sampling enabled and mean squared error as the splitting criterion, while default depth parameters were retained to avoid excessive tuning bias. The resulting feature importance rankings demonstrated a high degree of alignment with ANP-derived global weights, supporting consistency between expert judgment and data-driven relevance. Collectively, these results confirm the empirical robustness of the CHPI framework and support the use of machine learning as a triangulation and validation mechanism for indicator prioritization rather than as a generalizable prediction engine.\\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"4. Results\",\"content\":\"\\u003cdiv id=\\\"Sec15\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.1 WHSE Oversight Mechanisms\\u003c/h2\\u003e \\u003cp\\u003eDocument analysis and field review showed that the client organization operates a multi-level oversight system regulating contractor WHSE performance across the project lifecycle. Pre-contract evaluations shortlist contractors based on documented WHSE qualifications, followed by formal validation of appointed WHSE officers. Standardized templates and training are used to ensure consistent reporting. Monthly documentation submissions serve as operational checkpoints prior to invoice approval. Compliance is monitored through random field inspections and formal non-conformance protocols, which include financial deductions for repeated violations. At project closure, final WHSE scores are assigned using a structured form, and these scores are recorded for consideration in future contractor selection.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec16\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.2 Derivation of Core WHSE Indicators\\u003c/h2\\u003e \\u003cp\\u003eAn initial set of 84 candidate WHSE performance indicators was compiled from international frameworks, sectoral guidelines, and academic literature. These indicators underwent a two-stage expert review process using the CVI method with a minimum threshold of 0.78. Items that did not meet this threshold were removed, reducing the list to 43 sub-indicators. These were grouped into seven dimensions: Awareness and Competence, Monitoring and Auditing, Emergency Response, Environmental Health Management, Documentation, Risk Assessment and Management, and Management Systems. Cronbach\\u0026rsquo;s alpha was calculated for the final set, with an overall reliability coefficient of 0.91 and individual dimension values ranging from 0.81 to 0.92. Tables\\u0026nbsp;\\u003cspan refid=\\\"Tab3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e and \\u003cspan refid=\\\"Tab4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e present the CVI-based reduction process and the reliability statistics for each dimension.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab3\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 3\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eContent Validity Index (CVI) Results\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"4\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEvaluation Step\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eNumber of Items\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCVI Threshold\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eOutcome\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eInitial pool (based on literature)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e84\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026ndash;\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eAll items included\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003ePost expert review (relevant\\u0026thinsp;+\\u0026thinsp;clear)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e77\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCVI\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e7 items removed after expert screening\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eFinal CVI assessment (content validation panel)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e77\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCVI\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e34 items removed due to insufficient CVI\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eFinal retained sub-indicators\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e43\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e\\u0026ndash;\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eUsed in questionnaire and ANP modelling\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab4\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 4\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eWHSE Main and Sub-Indicators and Cronbach's Alpha\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"3\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMain Indicator\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSub-Indicator\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCronbach\\u0026rsquo;s Alpha\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eAwareness and Competence\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eOfficial WHSE training\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.88\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE Committee sessions\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSafety briefing meetings\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEvaluation and deployment of key WHSE personnel\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWork performance quality\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE safety qualification certificate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eTechnical safety certification for machinery\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"4\\\" rowspan=\\\"5\\\"\\u003e \\u003cp\\u003eMonitoring and Auditing\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePre-start-up audit\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"4\\\" rowspan=\\\"5\\\"\\u003e \\u003cp\\u003e0.84\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePeriodic WHSE audits\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePerformance monitoring checklist completion\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePerformance \\u0026amp; efficiency monitoring mechanism\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eContinuous monitoring of workplace pollutant levels\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"2\\\" rowspan=\\\"3\\\"\\u003e \\u003cp\\u003eEmergency Response\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEmergency response plan (ERP)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"2\\\" rowspan=\\\"3\\\"\\u003e \\u003cp\\u003e0.81\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEmergency preparedness drills\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eCommunication channels between client and contractor\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eEnvironmental Health Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePollution control and management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.86\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eComprehensive waste management\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEnergy consumption\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSanitation and hygiene conditions\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eProvision of site welfare facilities\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEnvironmental management plan implementation\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEnvironmental impact assessments\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eAccident and anomaly documentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE performance checklist\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE documentation and record control\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eOccupational disease count\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE warnings follow-up\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eOccupational health exams\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSeverity rate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"8\\\" rowspan=\\\"9\\\"\\u003e \\u003cp\\u003eRisk Assessment and Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eRisk identification and assessment system\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"8\\\" rowspan=\\\"9\\\"\\u003e \\u003cp\\u003e0.92\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSafe work permit procedures\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eCorrective actions for unsafe acts/conditions\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEssential safety equipment\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eFirst aid availability\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eFrequency rate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eSafety Data Sheet (SDS/MSDS)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePolicy statement for management systems\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eRisk control plan for unaccepted hazards\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"3\\\" rowspan=\\\"4\\\"\\u003e \\u003cp\\u003eManagement Systems\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eISO certifications\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\" morerows=\\\"3\\\" rowspan=\\\"4\\\"\\u003e \\u003cp\\u003e0.87\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE organizational structure and chart\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eWHSE budget\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003ePreparation of WHSE plan\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec17\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.3 Weighting of Indicators\\u003c/h2\\u003e \\u003cp\\u003eTo determine the relative importance of the main and sub-WHSE indicators, the ANP method was applied using structured pairwise comparisons from 14 domain experts. All computed CR were below the 0.05 threshold, confirming reliability of expert inputs. Table\\u0026nbsp;\\u003cspan refid=\\\"Tab5\\\" class=\\\"InternalRef\\\"\\u003e5\\u003c/span\\u003e presents the global priority weights for each main and sub-indicator. Among the seven main dimensions, Awareness and Competence achieved the highest weight (25.8%), followed by Risk Assessment and Management (21.9%) and Management Systems (14.8%). At the sub-indicator level, the highest weights were assigned to Formal WHSE Training (24.2% within its cluster), Risk Identification and Assessment Systems (21.0%), and ISO Certifications (25.1%). The lowest weights were observed in the Documentation dimension (6.7%).\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab5\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 5\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eANP-Derived Global Weights and Priorities\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMain Indicator\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eMain Weight (%)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSub-Indicator\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eSub Weight (%)\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eConsistency Ratio\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eAwareness \\u0026amp; Competence\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e25.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eFormal WHSE Training\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e24.2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.014\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCompetency Evaluation \\u0026amp; Deployment of Key WHSE Personnel\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e22.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eWHSE Committee Sessions\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e12.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSafety Qualification Certificate\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e10.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eTechnical Safety Certificate for Machinery\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e10.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSafety Briefing Meetings\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e9.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eWork-Quality Performance\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e9.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"7\\\" rowspan=\\\"8\\\"\\u003e \\u003cp\\u003eRisk Assessment \\u0026amp; Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"7\\\" rowspan=\\\"8\\\"\\u003e \\u003cp\\u003e21.9\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eRisk Identification \\u0026amp; Assessment System\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e21\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"7\\\" rowspan=\\\"8\\\"\\u003e \\u003cp\\u003e0.019\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eControl of acceptable HSE Risks\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e19.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSafe Work Permit Procedures\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCorrective Actions for Unsafe Acts/Conditions\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e14.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eEssential Safety Equipment\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e9.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eFirst Aid Availability\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSafety Signage\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e6.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSafety Data Sheet (SDS/MSDS)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e6.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"4\\\" rowspan=\\\"5\\\"\\u003e \\u003cp\\u003eManagement Systems\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"4\\\" rowspan=\\\"5\\\"\\u003e \\u003cp\\u003e14.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eISO Certifications\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e25.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"4\\\" rowspan=\\\"5\\\"\\u003e \\u003cp\\u003e0.026\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eWHSE Budget\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e24.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePolicy Statement for Management Systems\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e21.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePreparation of HSE Plan\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e17\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eWHSE Organizational Structure \\u0026amp; Chart\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e11.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"2\\\" rowspan=\\\"3\\\"\\u003e \\u003cp\\u003eEmergency Response\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"2\\\" rowspan=\\\"3\\\"\\u003e \\u003cp\\u003e13.2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eEmergency Preparedness Drills\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e45.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"2\\\" rowspan=\\\"3\\\"\\u003e \\u003cp\\u003e0.035\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eEmergency Response Plan (ERP)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e37.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCommunication Channels (Client Contractor)\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e17.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"5\\\" rowspan=\\\"6\\\"\\u003e \\u003cp\\u003eMonitoring \\u0026amp; Auditing\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"5\\\" rowspan=\\\"6\\\"\\u003e \\u003cp\\u003e9.9\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePre-Start-up Audit\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e22.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"5\\\" rowspan=\\\"6\\\"\\u003e \\u003cp\\u003e0.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eContinuous Monitoring of Workplace Pollutants\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e17.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCompliance with WHSE Regulations\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e16.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePeriodic Audits\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePerformance \\u0026amp; Efficiency Monitoring Mechanism\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eCompletion of WHSE Performance Checklist\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e12\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eEnvironmental Health Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e7.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003ePollution Control \\u0026amp; Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e18.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.018\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eEnvironmental Impact Assessments\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e14.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eComprehensive Waste Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e14.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eProvision of Welfare Facilities\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e14.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eImplementation of Environmental Management Plans\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e13.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eEnergy Consumption\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e12.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSanitation \\u0026amp; Hygiene Conditions\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e12.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e6.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eSeverity Rate\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e20.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\" morerows=\\\"6\\\" rowspan=\\\"7\\\"\\u003e \\u003cp\\u003e0.036\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eFrequency Rate\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eOccupational Health Examinations\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e15.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eNear Misses \\u0026amp; Anomalies Documentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e13.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eRecord Control \\u0026amp; Documentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e12.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eOccupational Disease Count\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e11.9\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eFollow up on HSE Warnings\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e10.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec18\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.4 Sensitivity Analysis\\u003c/h2\\u003e \\u003cp\\u003eA sensitivity analysis was performed to evaluate the effect of \\u0026plusmn;\\u0026thinsp;10% perturbations in the ANP-derived weights for each of the seven WHSE clusters on the CHPI scores and rankings. For each perturbation scenario, the modified CHPI scores were recalculated while holding the remaining cluster weights constant. The results are presented in Table\\u0026nbsp;\\u003cspan refid=\\\"Tab6\\\" class=\\\"InternalRef\\\"\\u003e6\\u003c/span\\u003e. Across all perturbations, no contractor shifted between compliance tiers. The maximum observed rank change was two positions, occurring under the Documentation cluster perturbation. Spearman\\u0026rsquo;s rank correlation coefficients for all scenarios remained above 0.96.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab6\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 6\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eSensitivity Analysis Results for \\u0026plusmn;\\u0026thinsp;10% Weight Variation\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eCluster Perturbed\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eTier Changes Observed\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eMax Rank Shift\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eSpearman\\u0026rsquo;s ρ\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eModel Stability\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAwareness and Competence\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.985\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eVery High\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eRisk Assessment and Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.987\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eVery High\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eManagement Systems\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.981\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eHigh\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEmergency Response\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.978\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eHigh\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMonitoring and Auditing\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.976\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eHigh\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEnvironmental Health Management\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.968\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eHigh\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.962\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eModerate\\u0026ndash;High\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec19\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.5 Empirical Validation of the CHPI Model through Contractor Evaluation\\u003c/h2\\u003e \\u003cp\\u003eTo verify the practical applicability of the CHPI framework, a case study evaluation was conducted on five representative contractors. Normalized performance scores were assigned across the seven principal WHSE domains reflecting realistic safety maturity levels typically observed in gas infrastructure projects. Final CHPI scores were calculated through weighted summation using the ANP-derived global priorities. Based on the overall scores, contractors were classified into three compliance tiers: High (\\u0026ge;\\u0026thinsp;85), Moderate (75.0\\u0026ndash;84.9), and Low (\\u0026lt;\\u0026thinsp;75). As presented in Table\\u0026nbsp;\\u003cspan refid=\\\"Tab7\\\" class=\\\"InternalRef\\\"\\u003e7\\u003c/span\\u003e, all contractors were placed in the Moderate tier, with scores ranging from 75.45 to 81.18. Contractor C achieved the highest CHPI score, primarily due to strong performance in Risk Assessment and Management as well as Awareness and Competence. In contrast, Contractor E obtained the lowest score, influenced by relatively lower results in Monitoring and Auditing and Emergency Preparedness.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab7\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 7\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eContractor CHPI Case Study Results\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"10\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c6\\\" colnum=\\\"6\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c7\\\" colnum=\\\"7\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c8\\\" colnum=\\\"8\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c9\\\" colnum=\\\"9\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c10\\\" colnum=\\\"10\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eContractor\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eAwareness \\u0026amp; Competence\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eRisk Assessment \\u0026amp; Management\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eManagement Systems\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eEmergency Response\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003eMonitoring \\u0026amp; Auditing\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003eEnvironmental Health Management\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003eCHPI Score\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eTier\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eA\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e63\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e87\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e75\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e75\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003e63\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003e76.07\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eModerate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eB\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e85\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e67\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e63\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e95\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e86\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003e87\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003e78.08\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eModerate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eC\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e85\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e88\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e64\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e76\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003e78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003e81.18\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eModerate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eD\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e73\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e66\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e93\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e88\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e74\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003e89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003e78.55\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eModerate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eE\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e75\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e76\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e68\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e63\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c8\\\"\\u003e \\u003cp\\u003e91\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c9\\\"\\u003e \\u003cp\\u003e75.45\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e \\u003cp\\u003eModerate\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003cp\\u003eThe fact that all five contractors were classified within the \\u0026ldquo;Moderate\\u0026rdquo; tier should be interpreted in light of the regulatory and operational context of the studied gas infrastructure projects, where minimum WHSE compliance standards are strictly enforced, and extreme performance deviations are uncommon among active contractors. Importantly, while tier-based classification places these contractors within the same compliance category, the CHPI demonstrates meaningful discriminative capability within this tier. As shown in Table\\u0026nbsp;\\u003cspan refid=\\\"Tab7\\\" class=\\\"InternalRef\\\"\\u003e7\\u003c/span\\u003e, CHPI scores vary by more than five points, and contractors exhibit distinct domain-level performance profiles across key dimensions such as Awareness and Competence, Risk Assessment and Management, and Monitoring and Auditing. These variations reveal substantive differences in safety maturity that would remain obscured under conventional checklist-based or threshold-driven evaluation approaches, which typically assign identical categorical ratings to contractors meeting minimum compliance requirements. In this regard, the added value of the CHPI lies not solely in categorical tier assignment, but in its ability to generate continuous, weighted performance scores that enable finer differentiation and more targeted interpretation of contractor WHSE performance, even within a single compliance tier.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec20\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.6 Descriptive Analysis of Contractor Data\\u003c/h2\\u003e \\u003cp\\u003eDescriptive statistics were calculated for the seven WHSE dimensions and the composite CHPI score based on evaluations of 200 contractors. As presented in Table\\u0026nbsp;\\u003cspan refid=\\\"Tab8\\\" class=\\\"InternalRef\\\"\\u003e8\\u003c/span\\u003e, CHPI scores ranged from 48.5 to 87.2, with a mean value of 67.87 (SD\\u0026thinsp;=\\u0026thinsp;9.69). Across the individual dimensions, Environmental Health recorded the highest mean score (79.01), while Risk Management had the lowest (61.16). Documentation and Awareness showed the greatest dispersion, with standard deviations of 14.79 and 14.06, respectively.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab8\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 8\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eDescriptive Analysis of Contractor Data\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eFactors\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eMean\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eStd\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eMin\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eMax\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eAwareness\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e67.49\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e14.06\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e39.4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e95.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eRisk Mgmt.\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e61.16\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e9.49\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e42.2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e80.1\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEmergency\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e71.97\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e8.14\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e55.7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e88.3\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eEnv. Health\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e79.01\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e12.21\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e54.6\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e100.0\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e63.12\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e14.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e33.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e90.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eMonitoring\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e67.33\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e12.25\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e42.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e91.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eManagement\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e63.12\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e13.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e35.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e90.8\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eCHPI\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e67.87\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e9.69\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e48.5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e87.2\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec21\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.7 Predictive Modelling Using Random Forest Regression\\u003c/h2\\u003e \\u003cp\\u003eTo validate the CHPI framework\\u0026rsquo;s predictive potential, a Random Forest Regression model was developed using the seven core WHSE indicators as input features. The model was trained on 80% of the dataset and tested on the remaining 20%. It yielded a coefficient of determination (R\\u0026sup2;) of 0.62, a MAE of 1.41, and a MSE of 3.33, indicating strong predictive performance with substantial explanatory power. As shown in Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig6\\\" class=\\\"InternalRef\\\"\\u003e5\\u003c/span\\u003e, actual and predicted CHPI scores exhibit a close alignment along the identity line, particularly in the mid-to-high performance ranges, with reduced dispersion compared to the lower-performing segment. The presence of natural prediction variance, without evidence of overfitting, affirms the model\\u0026rsquo;s generalizability and practical relevance in contractor safety assessments.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003eFeature importance analysis (See Table\\u0026nbsp;\\u003cspan refid=\\\"Tab9\\\" class=\\\"InternalRef\\\"\\u003e9\\u003c/span\\u003e) highlights Awareness \\u0026amp; Competence (28.1%), Environmental Health (22.1%), and Monitoring \\u0026amp; Auditing (16.8%) as the top predictors. This ranking mirrors the earlier ANP weightings and reinforces the central role of workforce capability, environmental management, and oversight in driving HSE outcomes. Across the 10-fold cross-validation procedure, model performance metrics exhibited low variability, indicating stable predictive behaviour despite the limited sample size. This consistency supports the use of Random Forest analysis as a robustness check for indicator prioritization rather than a generalizable prediction model.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab9\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 9\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eFeature Importance Analysis\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"3\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eNo.\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eFeature\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eImportance\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eMonitoring\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.1680\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEmergency\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.1078\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEnv. Health\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.2210\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eAwareness\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.2811\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.1023\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e6\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eRisk Mgmt.\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.0596\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eManagement\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.0602\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec22\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003e4.8 Alignment Between Data-Driven Insights and ANP Weights\\u003c/h2\\u003e \\u003cp\\u003eThe relative importance values obtained from the Random Forest Regression model were compared with the expert-derived global weights from the ANP process to evaluate consistency between empirical and judgment-based prioritization. As presented in Table\\u0026nbsp;\\u003cspan refid=\\\"Tab10\\\" class=\\\"InternalRef\\\"\\u003e10\\u003c/span\\u003e, Monitoring \\u0026amp; Auditing recorded the highest predictive importance in the ML model (37.6%) but received a lower weight in the ANP method (17.0%). Environmental Health Management ranked second in ML importance (22.5%) and displayed close agreement with its ANP weight (21.0%). Emergency Preparedness, Documentation, and Management Systems showed moderate differences between the two approaches. Risk Management and Awareness \\u0026amp; Competence exhibited the largest negative deviations, with their ANP weights exceeding their ML-derived importance.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab10\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 10\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003eComparative Weights of WHSE Indicators from ANP and ML Models\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"5\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003eRank\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eFeature\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003eML Importance\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eANP Weight\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eDifference (ML - ANP)\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eMonitoring\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.376\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.170\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.206\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEnv. Health\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.225\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.210\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.015\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eEmergency\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.126\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.080\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.046\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eDocumentation\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.087\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.080\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e+\\u0026thinsp;0.007\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eManagement\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.073\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.130\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u0026ndash;0.057\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e6\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eRisk Mgmt.\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.060\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.100\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u0026ndash;0.040\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003e7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003eAwareness\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.053\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.230\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e\\u0026ndash;0.177\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"5. Discussion\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003e5.1 Findings Interpretation\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e5.1 Benchmarks with Previous Studies\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eExisting contractor safety evaluation models predominantly rely on checklist-based assessments, lagging indicators, or linear scoring systems that treat safety indicators as independent variables. Such approaches often fail to capture the interdependencies among organizational, behavioural, and operational safety factors. Moreover, prior models typically depend either on expert judgment without empirical validation, or on data-driven techniques without structured theoretical grounding. The lack of hybrid frameworks that integrate interdependency-aware weighting with predictive validation limits both the reliability and practical applicability of existing contractor safety evaluation tools. This study addresses these gaps by proposing a hybrid framework that combines fuzzy ANP to model indicator interdependencies with machine learning\\u0026ndash;based validation using real compliance data, enabling both conceptual rigor and empirical robustness in contractor safety performance assessment.\\u003c/p\\u003e\\n\\u003cp\\u003eTo ensure that the proposed framework is built upon a robust and conceptually valid indicator set, a systematic content validity assessment was conducted prior to weighting and modelling. Following Lawshe\\u0026rsquo;s content validity method, the minimum acceptable CVR was set at 0.51 for a panel of 14 experts, indicating that each retained indicator had to be judged as essential by a clear majority of the expert panel [59]. Applying this threshold directly influenced indicator selection by filtering out items with insufficient consensus and retaining only those reflecting strong domain agreement. This step reduced subjectivity in expert judgment, strengthened the theoretical grounding of the indicator pool, and ensured that subsequent ANP weighting, and machine learning validation were applied to a rigorously screened and context-relevant set of contractors WHSE indicators.\\u003c/p\\u003e\\n\\u003cp\\u003eThe integration of ANP-derived expert weights with machine learning\\u0026ndash;based feature importance analysis in the CHPI framework represents a methodological advancement over prior contractor HSE evaluation studies, which have traditionally relied on either expert judgment or statistical correlation in isolation [60-62]. By combining a structured decision-making approach with empirical data-driven validation, this study addresses the methodological gap noted by Hill et al. (2021) [63], Tan et al. (2016) [64] and Zhang et al. (2025) [65], where the absence of a cross-verification mechanism limited the reliability of contractor performance indices. To the best of our knowledge, this is the first study to operationalize the integration of ANP-derived expert weights with machine learning\\u0026ndash;based feature importance within the CHPI framework for contractor WHSE evaluation, thereby offering a validated, dual-perspective approach that bridges conceptual priorities and empirical predictive strength.\\u003c/p\\u003e\\n\\u003cp\\u003eThe comparative assessment of ANP-derived weights and Random Forest Regression feature importance values revealed a mixed pattern of convergence and divergence, each carrying distinct implications for the conceptual validity and operational realism of the CHPI framework. The strongest divergence was observed in Monitoring \\u0026amp; Auditing, which emerged as the top predictor in the ML model (37.6%) yet received a substantially lower weight in the ANP analysis (17%). This 20.6 percentage point difference indicates that while experts recognise monitoring as a relevant dimension, they may underestimate its direct operational leverage. Empirical evidence supports the ML perspective: Gharedaghi and Omidvari (2019) [28]\\u0026nbsp;demonstrated that in oil and gas projects, systematic site inspections, real-time compliance tracking, and documented follow-up actions significantly outperform policy-driven measures in predicting actual HSE incident reduction. Similarly, Sattari et al. (2021) [66] found that consistent monitoring of safety-critical processes explained more variance in safety outcomes than training investments or governance structures, highlighting its high-frequency, high-impact nature. This aligns with the ML finding that monitoring provides an immediate corrective mechanism that directly influences CHPI scores. Given the limited number of contractors with complete machine-readable records, the machine learning component is intentionally positioned as a triangulation and validation mechanism rather than a definitive predictive engine.\\u003c/p\\u003e\\n\\u003cp\\u003eIn contrast, Environmental Health Management displayed high consistency across methods, reflecting strong conceptual and empirical agreement. Both expert judgement and data-driven modelling point to environmental safeguards as a central pillar of contractor WHSE performance. This alignment echoes findings from Amir-Heidari et al. (2017) [14] and Soleimani and Ferdos (2017) [14], who reported that environmental controls were statistically associated with higher safety compliance ratings and lower regulatory penalties in petrochemical and pipeline projects. The present results reinforce that environmental dimension serve as both a regulatory requirement and a driver of client confidence, thereby justifying their stable weighting in integrated frameworks like CHPI. Complementary experimental evidence from construction-related environments further demonstrates that effective control of airborne pollutants, such as volatile organic compounds, plays a critical role in mitigating environmental and occupational health risks at contractor-operated sites [67]. The present results reinforce that environmental dimension serve as both a regulatory requirement and a driver of client confidence, thereby justifying their stable weighting in integrated frameworks like CHPI. Beyond routine environmental controls, evidence from safety-critical infrastructure highlights that major accident risks also emerge from structurally interacting technical, organizational, and environmental factors. Empirical structural modelling of external floating roof tanks has demonstrated that safety outcomes are shaped by interconnected system components rather than isolated control measures, reinforcing the need for interdependency-aware evaluation frameworks in contractor governance [68].\\u003c/p\\u003e\\n\\u003cp\\u003eThe most striking divergence in the opposite direction was seen in Awareness \\u0026amp; Competence, which carried the highest expert derived ANP weight but ranked lowest in the ML results. This disparity suggests that while experts view training, formal qualifications, and competency validation as foundational to safety culture, these factors may exert indirect, long-term effects that are not fully captured in predictive models based on short- to medium-term outcome data. Similar patterns have been reported in other safety-related ML applications, where factors perceived by domain experts as highly influential were assigned comparatively low predictive importance by data-driven models. For instance, Zhong et al. (2022) [69] found that certain situation awareness indicators, though prioritized by pipeline safety experts, contributed minimally to ML-based risk predictions; Abbasianjahromi and Aghakarimi (2023) [70] observed comparable misalignments between managerial judgments and ML-derived feature rankings in construction safety performance; and Saridewi and Sari (2021) [71] noted that awareness and training variables in information security, while emphasized by practitioners, exhibited limited direct predictive power in ML models. Khalilzadeh et al. (2021) [72] similarly found that WHSE training initiatives improved compliance behaviours over multi-year timelines, but short-term datasets showed weak correlations because immediate operational outcomes depend more on enforcement and oversight mechanisms. Likewise, Lee et al. (2024) [42] argued that training interventions often act as enablers, enhancing the impact of other operational controls rather than producing immediate standalone effects, explaining their lower ML-derived importance in the present study.\\u003c/p\\u003e\\n\\u003cp\\u003eManagement Systems and Risk Management also revealed weaker predictive influence in ML analysis compared to their ANP weights. For example, Risk Management held a 10% ANP weight but only 6% ML importance, while Management Systems dropped from 13% in ANP to 7.3% in ML. Previous research by Aderamo et al. (2024) [56] and Sun et al. (2018) [73] offers a plausible explanation: governance structures, ISO certification, and risk assessment protocols create an essential enabling environment but do not directly translate into immediate performance shifts unless actively integrated into monitoring, enforcement, and operational routines. This can lead to an overestimation of their short-term impact when weights are assigned based solely on expert perception.\\u003c/p\\u003e\\n\\u003cp\\u003eFor Emergency Preparedness and Documentation, the ANP\\u0026ndash;ML comparison revealed relatively balanced mid-tier importance, with only modest differences. This alignment suggests that these domains act as bridging factors, valued by experts for their role in structural readiness while also demonstrating measurable operational performance benefits in ML models. Similar findings are reflected in prior research. Kyrkou et al. (2022) [74], Phark et al. (2018) [75], Richardson (2021) [76], and Dama\\u0026scaron;evičius et al. (2023) [77] collectively show that emergency planning and preparedness not only provide an organizational safety backbone but also yield quantifiable improvements in response time and risk mitigation when integrated with machine learning and sensor-based systems. Likewise, in the domain of safety documentation, Khan et al. (2025) [78], Fenton and Simske (2021) [79], and Kumi et al. (2025) [40] demonstrate that intelligent document processing and digitalized record-keeping enhance both compliance and the efficiency of safety management workflows, aligning with the dual structural\\u0026ndash;operational role observed in our results.\\u003c/p\\u003e\\n\\u003cp\\u003eOverall, the combination of ANP and ML results highlights the need for a dual-perspective weighting strategy in future CHPI iterations. ANP effectively captures long-term, culture-building priorities as perceived by domain experts, while ML uncovers high-frequency, high-impact levers that directly influence measurable safety outcomes. As Trivedi et al. (2024) [80] recommend, hybrid weighting systems that reconcile conceptual importance with empirical predictive validity can yield more balanced and context-responsive performance indices. In practice, this means retaining the structural emphasis on training, management systems, and risk protocols while increasing the operational weight of monitoring and environmental controls to reflect their proven predictive strength. By explicitly unpacking these convergences and divergences, the CHPI framework offers both a validated measurement tool and a diagnostic lens to reconcile expert judgment with operational data, enhancing transparency, policy relevance, and transferability to high-risk industries beyond the gas sector.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e5.1.2 Reconciling Expert-Based and Data-Driven Weights in Future CHPI Development\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe observed discrepancies between ANP-derived weights and machine learning\\u0026ndash;based feature importance values carry important implications for the future design and refinement of the CHPI framework. While the ANP prioritization reflects expert judgment regarding long-term structural and cultural drivers of contractor WHSE performance, the Random Forest results emphasize operational indicators with more immediate and measurable influence on empirical performance outcomes. These differences highlight that expert-based and data-driven weighting approaches capture complementary, rather than conflicting, dimensions of safety performance. From an index design perspective, this divergence suggests that relying exclusively on either expert judgment or machine learning may lead to partial representations of contractor safety performance. Expert-based weighting is well suited for capturing latent, system-level factors such as competency development, governance structures, and safety culture, whose effects may unfold over longer time horizons. In contrast, data-driven models tend to prioritize high-frequency, enforcement-oriented indicators, such as monitoring, auditing, and environmental controls, that exert direct and short-term influence on observed compliance outcomes.\\u003c/p\\u003e\\n\\u003cp\\u003eFuture versions of the CHPI could address this methodological tension through the development of a hybrid weighting system that integrates expert-derived and data-driven weights in a structured manner. One possible approach involves combining normalized ANP weights and machine learning importance scores using adjustable blending coefficients, allowing decision-makers to balance strategic, long-term priorities against operational performance sensitivity. Alternatively, expert weights may be treated as baseline structural weights, while machine learning outputs are used as dynamic adjustment factors that respond to evolving performance data. Such hybrid weighting mechanisms would preserve the conceptual integrity and interpretability of the CHPI while enhancing its empirical responsiveness and adaptability. By explicitly integrating both judgment-based and data-driven perspectives, future CHPI implementations could achieve a more balanced, context-aware performance index capable of supporting both strategic safety governance and real-time contractor performance management\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e5.2 Theoretical and Methodological Contributions\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThis study contributes meaningfully to both the theoretical discourse and methodological advancement of WHSE performance evaluation. Theoretically, it reinforces and operationalizes the systems thinking paradigm within safety science, particularly in contractor governance. By modelling contractor safety performance as a complex, interdependent network of behavioural, organizational, and procedural factors, the CHPI framework aligns with socio-technical systems theory such as Dahl \\u0026amp; Kongsvik (2018) [81] and Lingard \\u0026amp; Pirzadeh (2025) [82]. Rather than treating safety as the outcome of isolated activities, the framework emphasizes how the effectiveness of each domain, such as emergency preparedness or documentation, depends on the strength of others like competence or audit mechanisms. This holistic view advances prior work that typically approached safety indicators as independent variables. The emphasis on interactions, captured through the ANP, mirrors arguments from Salas and Hallowell (2016) [51], who noted that safety leading indicators are only meaningful when situated within broader systemic functions, such as communication quality, behavioural adherence, and feedback structures.\\u003c/p\\u003e\\n\\u003cp\\u003eMethodologically, the integration of a three-stage framework, Delphi consensus, CVI filtering, and ANP modelling, offers a novel structure for developing and validating performance metrics. While Delphi panels and CVI have been used separately in occupational safety studies such as Rshidi et al (2019) [83], their combination with network-based weighting introduces a level of methodological rigor previously absent from contractor-specific frameworks. The result is an indicator set that is not only psychometrically validated but also hierarchically interlinked, enabling more realistic prioritization. In comparison to earlier studies, such as Ahmed (2016) [84] or Nassiri et al. (2016) [85], which relied on binary or frequency-based scoring, the CHPI model offers graded, empirically ranked prioritization. It further introduces the ability to model reciprocal feedback, studies like Hadidi \\u0026amp; Khater (2015) [86] could not address.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eIntegrating real audit data anchors, the CHPI model in practice, transforming it into an operational decision-support tool. Random Forest regression provides data-driven validation, while ANP\\u0026ndash;ML comparisons reveal both alignment and variation, demonstrating analytical depth. Mapping the ANP structure to Safety Performance Theory and Socio-Technical Systems Theory further strengthens its validity and bridges theoretical constructs with practical WHSE evaluation (Figure 6).\\u003c/p\\u003e\"},{\"header\":\"6. Conclusion\",\"content\":\"\\u003cp\\u003eThis study developed and validated the CHPI, a novel multi-criteria framework for evaluating contractor safety performance in high-risk industrial sectors. By integrating expert driven fuzzy ANP weightings with empirical compliance data and machine learning validation, the CHPI moves beyond traditional checklist or scorecard approaches. It captures the interdependent structure of safety indicators and links them with real-world performance data, operationalizing safety as a systemic property rather than isolated metrics. Empirical analysis revealed that indicators related to Monitoring and Environmental Health Management exerted the strongest influence on overall WHSE performance, and the strong alignment between ANP-derived weights and Random Forest feature importance confirmed the internal validity and conceptual soundness of the CHPI model. Predictive testing further showed that CHPI scores can reliably distinguish between contractors with differing safety risk profiles, demonstrating its practical value as a proactive decision-support tool.\\u003c/p\\u003e \\u003cp\\u003eBeyond its methodological contribution, the CHPI carries significant implications for policy and practice. It enables more strategic allocation of oversight resources, supports performance-based contractor selection, and offers a foundation for integration into digital procurement and compliance platforms. Its modular architecture allows adaptation across sectors and regulatory environments, contributing to the harmonization of safety standards and strengthening accountability in contractor governance. Overall, the CHPI represents a step change in contractor safety evaluation, bridging the gap between theoretical rigor and operational applicability. Future research should extend its validation across diverse industries, incorporate adaptive weighting mechanisms responsive to dynamic risk conditions, and develop real-time dashboard applications to support continuous safety intelligence. As safety challenges grow in complexity, tools like CHPI provide a pathway toward more intelligent, responsive, and accountable risk management in the contractor ecosystem.\\u003c/p\\u003e \\u003cp\\u003eFrom an applied standpoint, the results suggest that contractor safety governance should move beyond predominantly documentation-oriented evaluations toward more dynamic, performance-sensitive oversight strategies. In particular, the dominant influence of Monitoring and Environmental Health Management highlights the importance of continuous site inspections, real-time compliance tracking, and timely corrective actions as primary levers for improving WHSE outcomes. Client organizations and regulators can operationalize the CHPI by embedding it within contractor prequalification systems, periodic performance audits, and incentive-based management schemes, thereby enabling targeted interventions for high-risk contractors. Moreover, the observed divergence between expert-derived priorities and machine learning importance underscores the need for balanced safety strategies that integrate long-term capacity-building measures with short-term operational controls. By translating analytical findings into concrete governance mechanisms, the CHPI provides a practical pathway for strengthening accountability, enhancing transparency, and supporting evidence-based decision-making in complex contractor environments.\\u003c/p\\u003e \\u003cp\\u003eQuantitatively, the proposed CHPI framework demonstrated strong discriminatory and predictive capability. Indicators associated with Monitoring and Environmental Health Management consistently ranked among the top-weighted dimensions, accounting for the largest share of influence on overall WHSE performance. The Random Forest model achieved robust predictive performance, with CHPI scores explaining a substantial proportion of variance in contractor safety outcomes and enabling clear differentiation between higher- and lower-risk contractors. Importantly, the convergence between expert-derived ANP weights and data-driven feature importance provides quantitative confirmation that the identified priority indicators are not only conceptually meaningful but also empirically relevant. These results substantiate the practical utility of the CHPI as a measurable, evidence-based tool for performance-oriented contractor safety management in high-risk industrial settings.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003ch2\\u003eDeclaration of Competing Interest\\u003c/h2\\u003e\\n\\u003cp\\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\\u003c/p\\u003e\\n\\u003ch2\\u003eClinical Trial Number\\u003c/h2\\u003e\\n\\u003cp\\u003eNot applicable.\\u003c/p\\u003e\\n\\u003ch2\\u003e\\u003cstrong\\u003eEthics/Consent to Participate/Consent to Publish declarations\\u003c/strong\\u003e\\u003c/h2\\u003e\\n\\u003cp\\u003efor this study was obtained from the Ethics Committee of Tabriz University of Medical Sciences (IR.TBZMED.REC.1398.416) and the Research Coordination Office of EAGC. All participants provided informed consent to participate and agreed to the publication of anonymized data. The study procedures complied with the ethical standards of the institutional research committee and the principles of the Helsinki Declaration.\\u003c/p\\u003e\\n\\u003ch2\\u003eFunding\\u003c/h2\\u003e\\n\\u003cp\\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\\u003c/p\\u003e\\n\\u003ch2\\u003eAuthor Contribution\\u003c/h2\\u003e\\n\\u003cp\\u003eConceptualization: Payam Khordoustan, Omid Akbarzadeh, Seyed Shamseddin Alizadeh. Data curation: Payam Khordoustan, Omid Akbarzadeh, Neda Gilani. Formal analysis: Neda Gilani, Omid Akbarzadeh. Investigation: Payam Khordoustan, Neda Gilani, Parisa Moshashaei, Seyed Shamseddin Alizadeh. Methodology Payam Khordoustan, Omid Akbarzadeh, Neda Gilani, Seyed Shamseddin Alizadeh. Project administration: Payam Khordoustan, Seyed Shamseddin Alizadeh. Resources: Payam Khordoustan, Neda Gilani, Seyed Shamseddin Alizadeh. Software: Neda Gilani, Omid Akbarzadeh. Supervision: Neda Gilani, Seyed Shamseddin Alizadeh. Visualization: Omid Akbarzadeh, Parisa Moshashaei. Writing \\u0026ndash; original draft: Omid Akbarzadeh. Writing \\u0026ndash; review \\u0026amp; editing: Seyed Shamseddin Alizadeh, Parisa Moshashaei.\\u003c/p\\u003e\\n\\u003ch2\\u003eAcknowledgement\\u003c/h2\\u003e\\n\\u003cp\\u003eThe authors gratefully acknowledge the West Azerbaijan Gas Company for their support during data collection and Tabriz University of Medical Sciences for their academic and logistical assistance.\\u003c/p\\u003e\\n\\u003ch2\\u003eData Availability\\u003c/h2\\u003e\\n\\u003cp\\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request.\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\u003cli\\u003e\\u003cspan\\u003eMearns K, Yule S. The role of national culture in determining safety performance: Challenges for the global oil and gas industry. Saf Sci. 2009;47(6):777\\u0026ndash;85.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAlNoaimi FA, Mazzuchi TA. Risk management application in an oil and gas company for projects. Int J Bus Ethics Gov (IJBEG), 2021: pp. 1\\u0026ndash;30.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAbdul Razak N, et al. Identification of health and safety prequalification criteria for contractor selection in construction projects: a systematic review. Energies. 2021;14(21):7244.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSmallwood J. The impact of a client contractor health and safety (H\\u0026amp;S) programme on contractor H\\u0026amp;S performance. Procedia Eng. 2017;196:996\\u0026ndash;1002.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTamim N, et al. Roles of contractors in process safety. J Loss Prev Process Ind. 2017;48:358\\u0026ndash;66.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWilliams SO, et al. Management of construction safety: the failure and success of stakeholders. Int J Acad Res Bus Soc Sci. 2018;8:58\\u0026ndash;68.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHalim SZ, Mannan MS. A journey to excellence in process safety management. J Loss Prev Process Ind. 2018;55:71\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eNowrouzi-Kia B, et al. A systematic review of lost-time injuries in the global mining industry. Work. 2018;60(1):49\\u0026ndash;61.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKhan S, Yaqoob S. \\u003cem\\u003eSafety Recordable Incident Cost Calculator\\u003c/em\\u003e. in \\u003cem\\u003eAbu Dhabi International Petroleum Exhibition and Conference\\u003c/em\\u003e. 2020. SPE.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLingard H, et al. Leading or lagging? Temporal analysis of safety indicators on a large infrastructure construction project. Saf Sci. 2017;91:206\\u0026ndash;20.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eXu J, et al. Implementing safety leading indicators in construction: Toward a proactive approach to safety management. Saf Sci. 2023;157:105929.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLeveson N. A systems approach to risk management through leading safety indicators. Reliab Eng Syst Saf. 2015;136:17\\u0026ndash;34.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSARKHEIL H, RAHBARI S. HSE Key Performance indicators in HSE-MS establishment and sustainability: a case of south pars gas complex, Iran. Int J Occup Hygiene. 2016;8(1):45\\u0026ndash;53.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAmir-Heidari P, et al. A new framework for HSE performance measurement and monitoring. Saf Sci. 2017;100:157\\u0026ndash;67.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eIghravwe DE, Oke SA. A fuzzy-grey-weighted aggregate sum product assessment methodical approach for multi-criteria analysis of maintenance performance systems. Int J Syst Assur Eng Manage. 2017;8(Suppl 2):961\\u0026ndash;73.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSimukonda W. \\u003cem\\u003eOccupational health and safety practices among contractors in Malawi: A generic overview.\\u003c/em\\u003e Proceedings of the Institution of Civil Engineers-Management, Procurement and Law, 2019. 172(3): pp. 118\\u0026ndash;124.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eValluru CT, Dekker S, Rae A. How and why do subcontractors experience different safety on high-risk work sites? Technol Work. 2017;19(4):785\\u0026ndash;94. Cognition.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eFalahati M, et al. Multi-dimensional model for determining the leading performance indicators of safety management systems. Work. 2020;67(4):959\\u0026ndash;69.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGhaleh S, et al. Presenting a conceptual pattern of HSE performance of oil trucks. Environ Monit Assess. 2018;190(2):97.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eEbrahimi MH, et al. Effects of administrative interventions on improvement of safety and health in workplace: A case study in an oil company in Iran (2011\\u0026ndash;2015). J Eng Appl Sci. 2016;11(3):346\\u0026ndash;51.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAkbarzadeh O, Golzad H, Moshashaei P. Health Promotion Model Insights on Determinants of PPE Utilization in Occupational Settings. Health Educ Health Promotion. 2025;13(1):1001\\u0026ndash;15.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWalter J. Safety management at the frontier: Cooperation with contractors in oil and gas companies. Saf Sci. 2017;91:394\\u0026ndash;404.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHajipour V, et al. An integrated process-based HSE management system: A case study. Saf Sci. 2021;133:104993.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAsbury S. Health and Safety, Environment and Quality Audits: A risk-based approach. CRC; 2023.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTaherdoost H, Madanchian M. Multi-criteria decision making (MCDM) methods and concepts. Encyclopedia. 2023;3(1):77\\u0026ndash;87.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKadoić N, Begičević Ređep N, Divjak B. \\u003cem\\u003eDecision making with the analytic network process\\u003c/em\\u003e. in \\u003cem\\u003eThe 14th International Symposium on Operational Research in Slovenia, SOR 2017\\u003c/em\\u003e. 2017.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDarko A, et al. Review of application of analytic hierarchy process (AHP) in construction. Int J Constr Manage. 2019;19(5):436\\u0026ndash;52.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGharedaghi G, Omidvari M. A pattern of contractor selection for oil and gas industries in a safety approach using ANP-DEMATEL in a Grey environment. Int J Occup Saf Ergon. 2019;25(4):510\\u0026ndash;23.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKaramoozian A, et al. An approach for risk prioritization in construction projects using analytic network process and decision making trial and evaluation laboratory. IEEE Access. 2019;7:159842\\u0026ndash;54.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGunduz M, Khader BK. Construction Project Safety Performance Management Using Analytic Network Process (ANP) as a Multicriteria Decision-Making (MCDM) Tool. Comput Intell Neurosci. 2020;2020(1):2610306.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eEskandari D, et al. Development of a scale for assessing the organization's safety performance based fuzzy ANP. J Loss Prev Process Ind. 2021;69:104342.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGul A et al. STATISTICAL MODELING AND PREDICTION OF PANDEMIC USING CLASSICAL AND MACHINE LEARNING APPROACH. Pakistan J Stat, 2025. 41(3).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDahl \\u0026Oslash;, Starren A. The future role of big data and machine learning for health and safety inspection efficiency. Bilbao, Spain: EU-OSHA; 2019.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKoklonis K, et al. Utilization of machine learning in supporting occupational safety and health decisions in hospital workplace. Eng Technol Appl Sci Res. 2021;11(3):7262\\u0026ndash;72.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSarwar J, et al. A comparative analysis of feature selection models for spatial analysis of floods using hybrid metaheuristic and machine learning models. Environ Sci Pollut Res. 2024;31(23):33495\\u0026ndash;514.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZermane A, et al. Predicting fatal fall from heights accidents using random forest classification machine learning model. Saf Sci. 2023;159:106023.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWu M et al. \\u003cem\\u003eA bayesian network model for real-time crash prediction based on selected variables by random forest\\u003c/em\\u003e. in. 2019 \\u003cem\\u003e5th International Conference on Transportation Information and Safety (ICTIS)\\u003c/em\\u003e. 2019. IEEE.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAhmed S, et al. A study on road accident prediction and contributing factors using explainable machine learning models: analysis and performance. Transp Res interdisciplinary Perspect. 2023;19:100814.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZheng X, et al. Comparison of machine learning methods for potential active landslide hazards identification with multi-source data. ISPRS Int J Geo-Information. 2021;10(4):253.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKumi L, Jeong J, Jeong J. Proactive approach to enhancing safety management using deep learning classifiers for construction safety documentation. Eng Appl Artif Intell. 2025;153:110889.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eOnukwulu EC, et al. Redefining contractor safety management in oil and gas: A new process-driven model. Int J Multidisciplinary Res Growth Evaluation. 2024;5(5):2582\\u0026ndash;7138.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLee S, et al. Beyond boundaries: The mechanisms by which host organizations\\u0026rsquo; safety climates impact on-site contractors\\u0026rsquo; safety and retention. J Saf Res. 2024;90:9\\u0026ndash;18.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eNg ST, Cheng KP, Skitmore RM. A framework for evaluating the safety performance of construction contractors. Build Environ. 2005;40(10):1347\\u0026ndash;55.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWu X, et al. Prospective safety performance evaluation on construction sites. Accid Anal Prev. 2015;78:58\\u0026ndash;72.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKarakhan AA, et al. Measuring and evaluating safety maturity of construction contractors: Multicriteria decision-making approach. J Constr Eng Manag. 2018;144(7):04018054.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMohammadfam I, et al. Evaluation of the quality of occupational health and safety management systems based on key performance indicators in certified organizations. Saf health work. 2017;8(2):156\\u0026ndash;61.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTixier AJ-P, et al. Application of machine learning to construction injury prediction. Autom Constr. 2016;69:102\\u0026ndash;14.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMostofi F, Toğan V. A data-driven recommendation system for construction safety risk assessment. J Constr Eng Manag. 2023;149(12):04023139.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWang B, et al. Using data-driven safety decision-making to realize smart safety management in the era of big data: A theoretical perspective on basic questions and their answers. J Clean Prod. 2019;210:1595\\u0026ndash;604.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTrist B, Murray H, Trist E. Characteristics of socio-technical systems. The Social Engagement of Social Science, a Tavistock Anthology, Volume 2: The Socio-Technical Perspective. University of Pennsylvania; 2016. pp. 157\\u0026ndash;86.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSalas R, Hallowell M. Predictive validity of safety leading indicators: Empirical assessment in the oil and gas sector. J Constr Eng Manag. 2016;142(10):04016052.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAcheampong T, Kemp AG. Health, safety and environmental (HSE) regulation and outcomes in the offshore oil and gas industry: Performance review of trends in the United Kingdom Continental Shelf. Saf Sci. 2022;148:105634.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eYu X, Xu S, Ashton M. Antecedents and outcomes of artificial intelligence adoption and application in the workplace: the socio-technical system theory perspective. Inform Technol People. 2023;36(1):454\\u0026ndash;74.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSackey E, Tuuli M, Dainty A. Sociotechnical systems approach to BIM implementation in a multidisciplinary construction context. J Manag Eng. 2015;31(1):A4014005.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDehdasht G, et al. DEMATEL-ANP risk assessment in oil and gas construction projects. Sustainability. 2017;9(8):1420.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAderamo AT, et al. AI-driven HSE management systems for risk mitigation in the oil and gas industry. Compr Res Reviews Eng Technol. 2024;2(1):1\\u0026ndash;22.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eBjerga T, Aven T. Adaptive risk management using new risk perspectives\\u0026ndash;an example from the oil and gas industry. Reliab Eng Syst Saf. 2015;134:75\\u0026ndash;82.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTaherdoost H, Madanchian M. Analytic Network Process (ANP) method: A comprehensive review of applications, advantages, and limitations. J Data Sci Intell Syst. 2023;1(1):12\\u0026ndash;8.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRomero Jeldres M, D\\u0026iacute;az E, Costa, Faouzi Nadim T. \\u003cem\\u003eA review of Lawshe\\u0026rsquo;s method for calculating content validity in the social sciences\\u003c/em\\u003e. in \\u003cem\\u003eFrontiers in Education\\u003c/em\\u003e. Frontiers Media SA; 2023.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eQabazard A, Mohanty AK, Adivi BSS. \\u003cem\\u003eContractors\\u0026rsquo; HSE Performance: Pragmatic Evaluation, Benchmarking and Way for Improvement\\u003c/em\\u003e. in \\u003cem\\u003eASSE Professional Development Conference and Exposition\\u003c/em\\u003e. 2017. ASSE.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDORMOHAMMADI A, et al. Providing an Approach for Evaluating the Organizational Level of Contractors in Terms of Safety, Health and Environment (HSE). Occupational Hygiene and Health Promotion; 2021.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eWang Y et al. \\u003cem\\u003eThe Influencing Factors and Methods of HSE Management in International General Contract Projects\\u003c/em\\u003e. in \\u003cem\\u003eInternational Field Exploration and Development Conference\\u003c/em\\u003e. 2022. Springer.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHill CB, Yadav OP, Khan E. Systemic risk analyses for potential impacts of onshore unconventional oil and gas development on public health and the environment: A critical review. Sci Total Environ. 2021;786:147512.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTan KH, Ortiz-Gallardo VG, Perrons RK. Using Big Data to manage safety-related risk in the upstream oil \\u0026amp; gas industry: A research agenda. Energy Explor Exploit. 2016;34(2):282\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZhang X et al. Research progress and prospects of risk assessment methods for oil and gas development. Geoenergy Sci Eng, 2025: p. 213992.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSattari F, et al. Application of Bayesian network and artificial intelligence to reduce accident/incident rates in oil \\u0026amp; gas companies. Saf Sci. 2021;133:104981.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAkbarzadeh O, et al. Investigation of Synergistic Influence of Ultrasound and Co-Doping to Degrade Toluene from Polluted Air in Construction Sites\\u0026mdash;An Experimental Approach. Buildings. 2024;14(9):2876.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMoshashaei P et al. A structural framework for fire and explosion risk in EFRTs: empirical validation using EFA, CFA, and path analysis. Sci Rep, 2025.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eZhong X, Zhang X, Zhang P. Pipeline risk big data intelligent decision-making system based on machine learning and situation awareness. Neural Comput Appl. 2022;34(18):15221\\u0026ndash;39.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAbbasianjahromi H, Aghakarimi M. Safety performance prediction and modification strategies for construction projects via machine learning techniques. Eng Constr architectural Manage. 2023;30(3):1146\\u0026ndash;64.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSaridewi VS, Sari RF. Implementation of machine learning for human aspect in information security awareness. J Appl Eng Sci. 2021;19(4):1126\\u0026ndash;42.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKhalilzadeh M, Shakeri H, Zohrehvandi S. Risk identification and assessment with the fuzzy DEMATEL-ANP method in oil and gas projects under uncertainty. Procedia Comput Sci. 2021;181:277\\u0026ndash;84.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSun Z-Y, Zhou J-L, Gan L-F. Safety assessment in oil drilling work system based on empirical study and Analytic Network Process. Saf Sci. 2018;105:86\\u0026ndash;97.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKyrkou C et al. \\u003cem\\u003eMachine learning for emergency management: A survey and future outlook.\\u003c/em\\u003e Proceedings of the IEEE, 2022. 111(1): pp. 19\\u0026ndash;41.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003ePhark C, et al. Prediction of issuance of emergency evacuation orders for chemical accidents using machine learning algorithm. J Loss Prev Process Ind. 2018;56:162\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRichardson N. Emergency Response Planning: Leveraging Machine Learning for Real-Time Decision-Making. Emergency. 2021;4:14.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDamaševičius R, Bacanin N, Misra S. From sensors to safety: Internet of Emergency Services (IoES) for emergency response and disaster management. J Sens actuator networks. 2023;12(3):41.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eKhan M et al. A machine learning driven automated system to extract multiple information fields from safety data sheet documents. Heliyon, 2025. 11(4).\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eFenton K, Simske S. \\u003cem\\u003eEngineering of an artificial intelligence safety data sheet document processing system for environmental, health, and safety compliance\\u003c/em\\u003e. in \\u003cem\\u003eProceedings of the 21st ACM Symposium on Document Engineering\\u003c/em\\u003e. 2021.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eTrivedi P, et al. A hybrid best-worst method (BWM)\\u0026mdash;technique for order of preference by similarity to ideal solution (TOPSIS) approach for prioritizing road safety improvements. IEEe Access. 2024;12:30054\\u0026ndash;65.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDahl \\u0026Oslash;, Kongsvik T. Safety climate and mindful safety practices in the oil and gas industry. J Saf Res. 2018;64:29\\u0026ndash;36.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLingard H, Pirzadeh P. Workplace health and safety performance at the client-contractor interface: measurement, management and behaviour. Saf Sci. 2025;184:106753.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eRshidi S, et al. Ranking Key Performance Indicators of Health, Safety, Environment, and Energy Education Using Multi-criteria Decision-making Techniques. J Occup Hygiene Eng. 2019;6(1):26\\u0026ndash;34.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eAhmed GA. \\u003cem\\u003eDevelopment of a health safety and environment (HSE) performance review. Methodology for the oil and gas industry in Libya\\u003c/em\\u003e. 2016.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eNassiri P, et al. Health, safety, and environmental management system operation in contracting companies: A case study. Arch Environ Occup Health. 2016;71(3):178\\u0026ndash;85.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eHadidi LA, Khater MA. Loss prevention in turnaround maintenance projects by selecting contractors based on safety criteria using the analytic hierarchy process (AHP). J Loss Prev Process Ind. 2015;34:115\\u0026ndash;26.\\u003c/span\\u003e\\u003c/li\\u003e\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":true,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"discover-analytics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"Learn more about [Discover Analytics](https://www.springer.com/journal/44257)\",\"snPcode\":\"44257\",\"submissionUrl\":\"https://submission.nature.com/new-submission/44257/3\",\"title\":\"Discover Analytics\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"stoa\",\"reportingPortfolio\":\"Discover Series\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true},\"keywords\":\"Contractor Safety Performance, Analytic Network Process, Machine Learning, Performance Index, Proactive Approach\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-8639261/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-8639261/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003cp\\u003eContractor-related deficiencies remain a critical challenge in workplace health and safety management within high-risk industries, particularly in the oil and gas sector, where heterogeneous safety practices and fragmented oversight mechanisms undermine effective risk control. Existing contractor evaluation approaches often rely on checklist-based or lagging indicators, offering limited ability to capture interdependencies among safety dimensions or to differentiate contractor performance in a meaningful and decision-relevant manner. This study develops and validates the Contractor Health, Safety, Environment, and Quality Performance Index (CHPI) as an integrated, evidence-based framework for systematic contractor performance evaluation. The study adopted a multi-method approach integrating expert judgment and archival compliance data. Health and safety indicators were identified through literature review and expert consultation, refined using Content Validity Index assessment, and weighted using a Fuzzy Analytic Network Process to capture interdependencies. The resulting weights were combined with standardized contractor records to compute CHPI scores. Robustness was confirmed through sensitivity analysis demonstrating stable contractor rankings, while a Random Forest\\u0026ndash;based analysis was used as a complementary validation to assess alignment between expert-based weights and data-driven importance. The results show that the CHPI enables nuanced differentiation of contractor performance, supports pre-contract screening and targeted intervention strategies, and enhances transparency in performance-based regulation. By integrating interdependency-aware weighting with empirical validation, the CHPI provides a scalable and adaptive decision-support tool that can strengthen contractor governance, improve safety performance monitoring, and support societal progress through more accountable risk management in high-risk industrial environments.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Developing the Contractor Health, Safety, and Risk Performance Index: A Hybrid ANP–Machine Learning Approach\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2026-01-29 16:01:34\",\"doi\":\"10.21203/rs.3.rs-8639261/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0},{\"type\":\"decision\",\"content\":\"Revision requested\",\"date\":\"2026-02-09T14:13:46+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2026-02-01T14:39:10+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2026-01-28T10:25:26+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"99226188894432034016497992934887949888\",\"date\":\"2026-01-27T13:21:57+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"181817891176325856292229779042895573615\",\"date\":\"2026-01-27T12:17:34+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewersInvited\",\"content\":\"\",\"date\":\"2026-01-27T11:48:48+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorAssigned\",\"content\":\"\",\"date\":\"2026-01-21T09:57:01+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"checksComplete\",\"content\":\"\",\"date\":\"2026-01-21T09:51:11+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"submitted\",\"content\":\"Discover Analytics\",\"date\":\"2026-01-19T11:50:19+00:00\",\"index\":\"\",\"fulltext\":\"\"}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"discover-analytics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"Learn more about [Discover Analytics](https://www.springer.com/journal/44257)\",\"snPcode\":\"44257\",\"submissionUrl\":\"https://submission.nature.com/new-submission/44257/3\",\"title\":\"Discover Analytics\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"stoa\",\"reportingPortfolio\":\"Discover Series\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"7d8d1aa4-538e-42b1-9db7-e42141319d4e\",\"owner\":[],\"postedDate\":\"January 29th, 2026\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"under-review\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2026-03-03T11:24:26+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2026-01-29 16:01:34\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-8639261\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-8639261\",\"identity\":\"rs-8639261\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}