Data-Driven Analysis of Construction Safety Dynamics: Regulatory Frameworks, Evolutionary Patterns, and Technological Innovations

Abstract

Construction remains one of the most hazardous industries, consistently reporting high rates of workplace injuries and fatalities. Despite advancements in safety regulations and technologies, significant risks persist due to hazardous tasks, including working at heights, operating heavy machinery, and exposure to harmful materials. The establishment of the Occupational Safety and Health Administration in 1971 marked a significant turning point in construction safety, resulting in a decline in workplace fatalities. However, evolving construction methodologies and digital transformations demand continuous research to enhance worker protection and mitigate emerging risks. This study conducts a longitudinal bibliometric analysis to examine the evolution of construction safety research from 1972 to 2025. Using a dataset of 14,174 journal publications from Scopus, the analysis identifies key research trends, technological advancements, and regulatory shifts that have shaped the field. Findings reveal a transition from basic safety regulations to AI-driven hazard detection, digital twins, and IoT-enabled safety monitoring. The study also identifies key contributors, including prominent countries. By tracing both historical and contemporary research trends, this study offers insights into knowledge gaps and provides guidance on future directions. The findings provide valuable insights for researchers, policymakers, and industry professionals, supporting the development of research-informed safety strategies and the integration of emerging technologies in an increasingly complex industry.

1. Introduction

Construction ranks as one of the most hazardous industries in the world, with a perennial position as one of the industries that record the highest cases of work-related injuries and fatalities. According to global estimates, work-related accidents result in approximately 380,500 deaths annually, with a substantial portion occurring in the construction sector due to high-risk activities such as working at heights, operating heavy machinery, and exposure to hazardous materials [1]. The rates of injury for the industry are similarly concerning. As of February 2025, the United States construction industry employed approximately 8.3 million workers, representing approximately 5.2% of the total U.S. workforce [2]. The United States recorded 5283 workplace fatalities in all industries in 2023, a decrease of 3.7% from 2022, or approximately one fatality every 99 min. The construction industry alone accounted for 1075 fatalities, representing approximately 20.3% of all work-related deaths [3]. Despite this reduction, the construction industry remains considerably hazardous, underscoring the urgent need for ongoing research into safety protocols, adherence to legislation, and technological advancements that help mitigate risks and protect worker health. Previously, safety regulations in construction were minimal until catastrophes revealed severe hazards, catalyzing regulatory change.

The most significant development in construction safety legislation was the establishment of the Occupational Safety and Health Administration (OSHA) in the United States in 1971, which was prompted by the high and persistent fatality and injury rates of the 1960s [4]. Workplace fatality rates have declined significantly since the establishment of OSHA—by approximately 60%—demonstrating the powerful impact of systematic regulation on occupational safety levels [5]. The research has made significant contributions to advancing these safety enhancements by providing the fundamental evidence necessary for effective regulatory policies and preventive measures. Knowledge of the trends in research development, the growth of regulatory policies, and the introduction of new safety technologies is vital for ongoing development. The ever-evolving construction environment with changing methodologies, materials, and technologies necessitates continuous research to counteract newly developing risks, particularly those associated with digitalization, sustainability, and evolving work practices [6,7]. Moreover, the identification of enduring issues such as human factors, safety culture and safety climate, and enforcement variability underscore the need for continuous research that determines root causes and develops effective mitigation measures [8].

Understanding how construction safety research has changed over time in response to technical developments, legislative changes, and industry transitions requires the use of a longitudinal bibliometric analysis. This approach systematically evaluates the evolution of research focus areas by dividing the dataset into multiple time periods, providing insights into emerging trends, knowledge development, and gaps in the existing literature [9]. Previous bibliometric studies such as those by [6,10,11,12] have clearly mapped important research themes and shown the shift from reactive accident management toward proactive, technologically advanced safety solutions. These reviews have highlighted general trends without systematically dividing studies into distinct historical phases to analyze the impact of key regulatory and technological watershed points. Additionally, given the speedy evolution of artificial intelligence, automation, and digital safety technologies, the time is ripe for an extensive analysis. This study bridges those gaps by conducting an exhaustive longitudinal bibliometric review spanning half a century (1972–2025), bringing into sharper focus the ways in which construction safety research has progressed and identifying areas requiring further investigations.

Given the above context, the primary objective of this study is to analyze the historical and thematic evolution of construction safety research. In particular, this study aims to: (1) conduct a longitudinal bibliometric analysis to examine the evolution and global distribution of construction safety research from 1972 to 2025, (2) identify major research themes, technological advancements, and regulatory shifts that have shaped the field, and (3) highlight the contributions of leading institutions and geographic regions in advancing construction safety knowledge. By presenting a systematic overview of previous research trends and future directions, this research provides a clearer picture of the field’s development, reveals current knowledge gaps, and provides strategic guidance to support future research, regulatory policies, and industry best practices.

2. Methodology

In this section, we present the methodologies utilized in the data collection and quantitative trend analysis phases of the study. Specifically, it explains the steps and systematic approach followed to identify, screen, and select the relevant literature from the Scopus database. Additionally, it describes the bibliometric techniques used to analyze the evolution of construction safety research from 1972 to 2025.

2.1. Data Collection

To achieve the research objectives and understand how the focus of construction safety research has evolved over the last five decades, this study conducts a longitudinal bibliometric literature review to examine the progression of research from 1972 to 2025. Longitudinal literature reviews are essential for identifying research trends, topic developments, and the growth of knowledge over extended periods. Unlike static reviews, which capture a field at a single point in time, longitudinal reviews divide the dataset into multiple time periods to track how research focus areas emerge, evolve, and fade over time [9]. Thus, it provides a structured overview of construction safety research development over the last five decades, in response to technological advancements, shifts in regulations, and industry demands.

The start year for data collection (i.e., 1972) was selected to correspond with the establishment of the Occupational Safety and Health Administration (OSHA) in 1971, a significant event that significantly impacted workplace safety regulations in the United States and influenced international safety research efforts. Since its inception, OSHA has reduced workplace fatalities by 60% and occupational illness and injury rates by 40%, making 1972 a logical starting point for this longitudinal review [13]. Moreover, the Scopus database was selected to be the main source of data due to its extensive coverage of construction and interdisciplinary journal publications [14]. Consequently, it ensures a comprehensive collection of information related to the topic of study. Figure 1 illustrates the procedure used to obtain the final set of publications employed in the analysis.

The keyword selection for this study was designed to comprehensively capture the scope of construction safety research while maintaining relevance to occupational safety principles and the construction sector. Consequently, the search query was formulated to cover the related literature by integrating terms associated with safety (e.g., “safety management”, “accidents”, “hazard identification”, “OSHA”, “preventive safety measures”, “automation in safety”) and keywords associated with construction (e.g., “construction”, “construction industry”, “built environment”, “offsite construction”). As a result, this formation of keywords encompassed essential themes in safety performance, regulatory compliance, risk mitigation, and technological advancements while maintaining a focus on the construction sector and excluding broader workplace safety topics related to other industries. The study covered articles from 1972 through early 2025 by searching within the title, abstract, and keywords sections of the literature. This resulted in the identification of 103,029 documents, with only 87,748 of them being in English (see Figure 1).

The research team employed a systematic methodology to review and evaluate the literature. Each team member initially conducted an independent analysis of the literature’s relevance utilizing various Scopus filters, followed by a focus group meeting to discuss the findings and interpretations. The group conducted several review sessions to compare selections, resolve discrepancies, and integrate the insights, ensuring the inclusion of the most relevant publications. As shown in Figure 1, the first screening step focused on source type, limiting the dataset to journal publications to ensure the inclusion of only high-quality, peer-reviewed research. Consequently, conference proceedings, books, and other forms of literature were excluded, resulting in a total of 48,809 records. Furthermore, the publications were screened by subject areas that are closely related to safety studies within the built environment, including engineering, environmental science, social sciences, computer science, multidisciplinary studies, and decision sciences. Therefore, research articles outside these domains, such as those in architecture, physics, or unrelated medical fields, were excluded. Following this step, the dataset was reduced to 14,490 records, reflecting the unique number of documents after accounting for overlaps in Scopus subject categories. The last step in the screening process included aligning the dataset with the historical development of construction safety research, as mentioned earlier. Thus, the publication timeframe was limited to 1972–2025, corresponding to the post-OSHA era and reflecting regulatory and technological advancements in safety management, which yielded 14,174 records that served as the final dataset input for the longitudinal analysis using VOSviewer version 1.6.18. While this study did not adopt the full systematic literature review process, the large-scale nature of bibliometric analysis (14,174 documents) made individual abstract screening impractical. Instead, the research team applied a robust, systematic filtering process—including iterative discussions and consensus checks—to ensure dataset validity and maintain high relevance across all selected publications.

2.2. Quantitative Trend Analysis

Bibliometric analysis is a quantitative approach used to examine the structure and development of a research field through the application of statistical tools to academic publications [14]. It is widely applied in scientometric studies to examine knowledge domains, research trends, and interconnections among publications, authors, and keywords [15]. In this study, author keyword co-occurrence analysis was conducted to highlight research trends and connections over time, shifts in research focus, and the relationships between key topics in construction safety research from 1972 to 2025 based on the core content of publications represented by their keywords. Bibliometric studies of this nature provide a strong foundation for identifying knowledge gaps and guiding future research directions [16]. VOSviewer was chosen for this purpose due to its strong capabilities in constructing and visualizing bibliometric maps, particularly co-occurrence networks. Compared to other bibliometric tools, VOSviewer offers an intuitive, free-to-use interface with advanced interactive features such as zooming, searching, and clustering, making it especially effective for handling large datasets like the 14,174-document corpus used in this study [17]. Its built-in thesaurus and normalization functions further enabled precise keyword unification and network clarity, which aligned with the study’s goal of mapping evolving research themes in construction safety and supporting citation analysis to provide a comprehensive bibliometric perspective [18].

To effectively evaluate the evolution of construction safety research and to illustrate the distribution of research topics with greater clarity and convenience, the analysis of the collected dataset involved splitting the publications into five distinct subperiods: 1972–1985, 1986–1995, 1996–2005, 2006–2015, and 2016–2025. These five subperiods were selected based on significant regulatory and technological milestones influencing construction safety. Specifically, these intervals correspond to the initial establishment of OSHA and early safety standards (1972–1985), widespread regulatory reforms and early technological adoption (1986–1995), increased emphasis on computational tools and structured safety management (1996–2005), accelerated adoption of digital tools and BIM (2006–2015), and the recent rapid expansion of AI, IoT, and digital twin technologies in safety management (2016–2025). Although alternative interval selections could slightly influence the emphasis or frequency of certain themes, the overall evolutionary trends identified would likely remain robust. This segmentation approach adheres to recognized practices in longitudinal bibliometric research, in which the dataset is divided into various time intervals to identify changes in research emphasis, emerging themes, and the development of intellectual frameworks [6,9,11,19]. Thus, analyzing each subperiod’s publications allows for a comparative analysis of how the research focus has evolved, enabling us to track emerging topics and technology developments in construction safety research.

The dataset for this analysis, which included citation, abstract, and keywords information of the 14,174 documents, was extracted from Scopus in CSV format and imported into VOSviewer for visualization. However, before initiating the analysis and ensuring consistency across the dataset, data cleaning and normalization were conducted to consolidate terms with equivalent semantic meanings. This was performed by collecting all keywords generated by VOSviewer and manually evaluating the meaning of each one. For instance, variations such as “construction safety”, “construction worker safety”, and “construction site safety” were unified under the single term “safety” to provide a clearer and more accurate representation of research trends. This consolidation was part of a systematic process applied across the entire dataset to ensure consistency and precision in the keyword analysis. For example, terms like “building information model”, “building information model (BIM)”, and “building information modeling” were all standardized under the term “BIM” to unify terminology related to digital construction tools.

For each subperiod, author keyword co-occurrence analysis was conducted separately to examine how research themes evolved over time. The analysis generated a network of interconnected nodes that vary in color and size, where each node represents a keyword, with its size reflecting its frequency of occurrence in the dataset. Moreover, frequently co-occurring keywords were organized into research clusters with unique colors, where all nodes with the same color presented a research trend. Additionally, knowing that the connection between the nodes are called links, their thickness describes the strength (i.e., frequency) of their co-occurrence, which means a thicker line indicates a stronger connection, allowing for the identification of key research patterns and shifts in focus within each time period. Using this information and analyzing each subperiod independently, this analysis provides valuable insights into major research trends during each subperiod, academics’ interests, and how different research topics were intellectually connected. In addition, it draws the trajectory of construction safety research by highlighting knowledge progression, the rise of new safety technologies, and evolving regulatory concerns over time.

In addition to author keyword co-occurrence analysis, the bibliometric assessment included an evaluation of the geographical distribution of research publications, and an assessment of institutions contributing to construction safety research. The first author’s affiliation country was analyzed to determine which nations have contributed the most to construction safety research over the study period. This provided insights into which countries have led research advancements and had the highest contributions to this field. On the other hand, the institutional analysis was conducted to determine which research leading affiliations have made the most significant contributions to construction safety research, highlighting key academic and industry-affiliated organizations that have driven advancements in the field.

3. Results

This chapter examines the evolution of construction safety research from 1972 to 2025, utilizing a longitudinal bibliometric analysis. It discusses the international contributions of various countries, the leading research institutions, and thematic developments in safety research. Based on publication trends, keyword occurrences, and institutional output, this study highlights the development, thematic focus, and top contributors in the discipline.

3.1. Research Trends and Global Contributions

The bibliometric analysis of construction safety research between 1972 and 2025 reveals widespread contributions from diverse countries. As shown in Figure 2, research publications on construction safety have been contributed by 134 nations, comprising 14,174 documents over this period. The worldwide contributions indicate a broad academic and industrial interest in enhancing safety standards in the construction industry. China is the frontrunner with 3634 publications, a testament to its pioneering status in construction safety research. This is most likely a result of heightened urbanization, mega-infrastructure development, and policy measures aimed at ensuring safety in construction [6]. The United States is second with 2647 publications, an indicator of its well-established research facilities, deep-seated regulatory frameworks such as OSHA, and technological innovations in safety monitoring and automation. The United Kingdom, with 987 publications, and Australia, with 601 publications, are third and the fourth in the world. These countries have a long history of safety regulations and emphasize sustainable construction practices. Canada has 525 publications, making a significant contribution that reflects its strong research initiatives, government regulations, and industrial safety standards in the construction sector. Asian countries, such as India (544), South Korea (534), Hong Kong (399), and Malaysia (358), are also featured prominently, indicating a greater focus on safety in fast-developing construction industries. South Korea and Hong Kong, among the leading contributors, indicate that more funds are being invested in research to address safety issues in large-scale city buildings [20]. European contributors, such as Italy (394), also demonstrate ongoing research activities in occupational safety, workplace hazards, and compliance with regulations. The output of these countries suggests robust academic research settings and a government focus on construction safety.

Although the top ten nations account for a high percentage of publications, representing 74.9% of the total, the inclusion of an additional 124 nations suggests an expansion in this field of research. This rise in global involvement marks a shift toward cooperation in knowledge, cross-border collaboration, and harmonization of safety protocols within the construction industry. The expansion of research in various geographical areas reflects an increasing awareness of the significance of construction safety as a vital discipline in occupational health and engineering.

Figure 3 illustrates the evolution of the leading 10 countries in construction safety research during five distinct periods between 1972 and 2025. In the early years (1972–2005), Western nations dominated safety research endeavors, with the United States consistently holding the top position, followed by such countries as the United Kingdom, Germany, Canada, and Japan. Most contributions during this period were confined to developed economies with well-established occupational safety legislation and robust research infrastructures. The regular appearance of nations such as Australia, Italy, Canada, and Japan across various time periods indicates sustained interest in construction safety research over the years. From 2006, the data reveal a profound geographic shift. China entered the top 10 in the 1996–2005 period and rapidly emerged as the leading contributor by 2016–2025. Similarly, countries such as South Korea, India, Malaysia, and Hong Kong have become key contributors, signaling a shift in the global research landscape. This transition is consistent with the acceleration of construction activities, increased research investment, and shifting regulatory environments in rapidly industrializing nations. The emergence of these countries marks a more globally cooperative stance in addressing construction safety issues and demonstrates the spread of academic contributions that were previously concentrated in a few locations.

Figure 4 illustrates the number of documents published annually related to construction safety, with contributions increasing gradually as safety research gained prominence in the construction sector. The collection comprises 14,474 papers. Although the count of publications in 2025 appears lower (116 documents), it is noteworthy that this review was conducted in January 2025, with many papers for the year still to be released. Between 1972 and 1985, the number of publications remained relatively low, with a total of 621 papers, accounting for 4.4% of all documents. Early in the period, annual publication numbers ranged from 12 to 111, depending on the year. With 111 documents, 1984 had the most publications in this decade, demonstrating a growing consciousness of safety in construction methods. Construction safety research was still in its early stages during this time, with few technological developments enabling new studies in this field and weak legislative systems [21]. Between 1986 and 1995, there was a gradual increase in construction safety research, resulting in 825 papers published throughout the ten-year period, which accounts for 5.9% of all documents. The number of yearly publications ranged from 69 to 107. Although development was consistent, it was quite sluggish, implying that safety issues still ranked low in research importance. This era marked the beginning of more organized safety protocols and a gradual acceptance of standardized regulations in the construction industry.

Between 1996 and 2005, publication rates increased significantly, resulting in 1125 documents (7.9% of all documents). Publications in 2005 totaled 159, indicating a growing research interest in safety protocols and regulatory evolution. The consistent rise in research publications is attributable to heightened global activity in occupational safety and the adoption of enhanced construction safety measures. The sector has also begun to incorporate emerging technologies to enhance safety monitoring and risk management strategies. The next decade (2006 to 2015) saw 2839 papers published over ten years, accounting for 20% of the total documents and indicating a significant increase in global construction safety research efforts. This decade witnessed a sharp upward trajectory, reaching 428 papers in 2015. The deployment of sophisticated safety monitoring devices, automation, and improved regulatory systems had presumably facilitated this surge [22]. Construction firms and regulatory agencies prioritized risk reduction measures, which resulted in a higher level of industrial and academic interest in construction safety. Between 2016 and 2025, research in construction safety grew exponentially, with 8764 publications within a decade, representing 61.8% of the total documents. The year 2024 saw an all-time high of 1871 publications, reflecting a peak in global research activity. The rapid growth of digital safety equipment, artificial intelligence, and IoT-based monitoring systems has been a major driver of this boom [23]. The impact of recent regulatory changes, growing industry awareness, and the increasing complexity of construction projects have also been significant motivators for this growth. As the ongoing digital revolution continues to impact the construction sector, safety research is likely to continue growing and improving, with more data-driven and predictive safety modeling being incorporated [24].

3.2. Thematic Evolution and Author Keyword Co-Occurrence Analysis

Supported by author keyword co-occurrence networks generated by VOSviewer, this section presents a comprehensive analysis of the evolution of construction safety research. From 1972 to 2025, the following bibliometric networks reveal how primary research subjects have evolved, emerged, and interconnected over the five decades. Beginning with the foundational period immediately following OSHA’s establishment in the United States, the following subsections individually review each subperiod.

3.2.1. Subperiod 1972–1985

The subperiod between 1972 and 1985 represents a critical phase in the study of construction safety, marking the establishment of OSHA and the creation of interdisciplinary safety fields such as process safety [25,26]. During this time, a turning point was marked by enforcing workplace safety regulations to reduce construction accidents, which is reflected in the keyword co-occurrence network generated through VOSviewer. Examining the network reveals interrelated subjects and provides insight into the field’s development during this period, as illustrated in Figure 5.

A notable focus during this period was the safety and accident prevention cluster (blue cluster), which emerged as the most central theme. From the figure, the keyword “safety” appeared 82 times, which is the highest occurrence in the network, reflecting its dominance in the research landscape. Moreover, it is closely linked to other keywords such as “accident prevention” (37 occurrences) and “accident” (19 occurrences), emphasizing proactive strategies to minimize risks in construction environments. These studies focused on developing structured frameworks and practices to reduce workplace hazards, ultimately creating safer operational environments for workers.

Focusing on the technological components of safety, the industrial processes and machinery cluster (green cluster) includes keywords such as “machinery” (10 occurrences), “equipment” (17 occurrences), and “maintenance” (7 occurrences). This cluster highlighted the crucial need to manage equipment-related risks during a period of rapid industrialization in construction practices. Research in this area has confirmed the significance of routine inspections and maintenance protocols in preventing equipment defects, enhancing operational safety, and minimizing mechanical failures [27].

The fire safety and regulatory standards cluster (yellow cluster) concentrated on fire-related hazards and the necessity for organized safety frameworks. A rising emphasis on fire dangers and regulatory frameworks is indicated by keywords like “fire safety” (5 occurrences), “fires” (12 occurrences), “standards” (23 occurrences), and “legal aspects” (14 occurrences). Research highlighted the use of fire-resistant materials, evacuation strategies, and emergency response methods, while also institutionalizing safety measures through regulations such as OSHA [28]. This cluster focuses on initial efforts to mitigate hazards and ensure compliance in complex urban construction projects.

The human factors cluster (red cluster) focused on the behavioral and psychological dimensions of construction safety. Terms such as “therapy” (5 occurrences) and “preventive actions” (8 occurrences) highlighted the need for organized interventions to mitigate workplace risks. This research cluster focused on techniques to enhance worker awareness, promote behavioral safety, and develop proactive strategies for accident avoidance [29,30].

The overall focus of each cluster signifies a transition from reactive to proactive safety strategies during this critical period. Significant improvements included the incorporation of safety into industrial design through the development of machinery protocols and maintenance methods, in conjunction with regulatory frameworks such as OSHA. The focus on fire safety protocols and preventive behavioral safety measures mitigated immediate hazards and laid the groundwork for advancements in construction safety research.

3.2.2. Subperiod 1986–1995

Construction safety research experienced a revolutionary period from 1986 to 1995, characterized by improvements in organized safety procedures, the integration of technology, and the development of regulatory frameworks. Figure 6 illustrates how safety research expanded this decade to include risk management frameworks, automation, and computational tools in addition to conventional accident prevention techniques.

As seen in the figure, one significant change from the previous subperiod was the greater focus on accident prevention (red cluster), which took precedence over the more general “safety” focus of 1972–1985. Thus, the keyword “accident prevention” appeared 131 times, making it the most frequently occurring term in the network. Moreover, emphasizing proactive strategies to minimize risks in building environments and underscoring the extension of regulatory enforcement and structured safety policies, the second, the third, and the fourth most occurring keywords were “construction” (51 occurrences), “building codes” (46 occurrences), and “safety” (43 occurrences). This shift was largely influenced by regulatory reforms, including the introduction of OSHA’s fall protection standards in 1994 [31], the EU’s Temporary or Mobile Construction Sites Directive in 1992 [32], and OSHA’s Voluntary Protection Program (VPP) in 1989 [33]. Furthermore, the gradual restriction of hazardous building materials, such as asbestos, contributed to long-term occupational health improvements.

On the other hand, the green cluster, which included keywords related to technological advancements and automation tools, focused on integrating new technology to improve construction safety. This cluster comprised such terms as “automation” (11 occurrences), “robotics” (9 occurrences), and “computer-aided design (CAD)” (7 occurrences), indicating an increasing reliance on computational tools to enhance risk management and hazard identification. In addition, the presence of the keywords “computer simulation” (11 occurrences) and “project management” (18 occurrences) underscores the adoption of digital modeling techniques to optimize safety decision making. The release of important software at that time, such as AutoCAD Release 10 in 1988, which advanced 3D modeling for risk assessment, and Primavera Project Planner in 1983, which became indispensable in the 1990s for scheduling and safety planning, can help to explain this growing focus in technological solutions [34,35]. Furthermore, the adoption of computer-controlled equipment in the late 1980s and early 1990s, such as automated bricklaying and welding, enhanced efficiency while minimizing worker exposure to hazardous tasks, laying the groundwork for the development of robotics systems in construction safety [36,37].

Focusing on mechanical hazards, excavation risks, and material safety measures, the blue and yellow clusters describe the construction equipment and environmental safety of this subperiod. The industry’s growing interest in machinery-related hazards is evident in the inclusion of such terms as “construction equipment” (15 occurrences), “excavation” (10 occurrences), and “hydraulics” (9 occurrences). This aligns with the increased mechanization of construction processes during this period, which introduced new safety concerns related to the operation of heavy machinery and excavation work. In addition, the development of geotechnical safety protocols became essential, as reflected in the occurrence of “soil” (8 occurrences), “soil mechanics” (8 occurrences), and “material testing” (7 occurrences), demonstrating the industry’s growing attention to ground stability and hazard prevention. This attention was reflected in OSHA’s excavation standards of 1989, which played a critical role in shaping safety regulations by mandating protective systems, such as trench shields and benching techniques, to reduce collapse risks [38]. As a result, advancements in hydraulic safety mechanisms and automated excavation equipment during this period significantly enhanced machine reliability and worker protection [39].

The construction industry and workforce safety (light blue cluster) concentrated on minimizing occupational risks, improving site safety management, and the growing role of governmental policies in shaping construction safety practices. The most relevant keywords in this cluster include “construction industry” (32 occurrences), “legislation” (21 occurrences), “personnel” (10 occurrences), and “occupational risks” (9 occurrences). This increasing focus on workforce-related risks aligns with regulatory changes that placed greater responsibility on employers to ensure safer working conditions. Specifically, new safety regulations for workers subjected to repetitive tasks and physical strain were adopted as a result of the emergence of ergonomic studies and occupational hazard evaluations in the late 1980s [40]. Additionally, programs like OSHA’s Voluntary Protection Program (VPP) encouraged construction companies to adopt industry-leading safety practices and exceed regulatory compliance requirements [41].

Finally, several of the other terms in the red cluster draw attention to a growing concentration on occupational safety in hazardous environments, especially with regard to hazardous material exposure, nuclear safety, and pollution control. These keywords include “environmental protection” (12 occurrences), “hazardous materials” (9 occurrences), and “nuclear power plants” (21 occurrences), which reflect an increasing awareness of workplace hazards stemming from environmental factors. During this period, concerns over worker exposure to toxic substances intensified, leading to stricter regulations on asbestos, lead-based paints, and other hazardous construction materials. Thus, OSHA expanded regulations on handling dangerous substances, reinforcing protective measures such as proper ventilation systems, personal protective equipment (PPE), and decontamination protocols [42,43]. Additionally, the increased construction of nuclear power plants in the late 1980s and early 1990s introduced new occupational risks for construction workers involved in reactor building and maintenance [44]. The consequences of the Chernobyl accident in 1986 heightened investigations into radiation exposure mitigation, protective shielding for workers, and emergency safety practices in nuclear construction sites. This period witnessed an intensified focus on worker training programs, ensuring that laborers in radioactive or contaminated environments were equipped with specific protective equipment and exposure monitoring systems [45].

3.2.3. Subperiod 1996–2005

Driven by creative technology, stricter regulatory standards, and more focus on environmental and worker safety, construction safety research made significant progress from 1996 to 2005. While regulatory systems gave a basis for enhanced compliance, Figure 7 shows how this decade encouraged the use of automation, robotics, and data systems to increase safety measures. Moreover, occupational health and environmental hazards became key concerns, indicating a transition toward more proactive and sustainable safety procedures.

With keywords such as “accidents” (377 occurrences), “construction” (338 occurrences), and “computers” (132 occurrences), the red cluster indicates an increased focus on utilizing technology to address workplace safety concerns and enhance accident prevention. Throughout this period, accidents remained a major concern, driving researchers to develop real-time monitoring systems and prediction models to minimize hazards [46,47]. The increasing integration of “systems” (100 occurrences), “technology” (33 occurrences), “automation” (25 occurrences), and “robotics” (23 occurrences) highlights the sector’s efforts to mitigate worker exposure to hazardous jobs. In response to these difficulties, researchers employed computational tools, including decision support systems and simulation models, to examine accident trends and develop proactive safety measures to enhance risk mitigation strategies [48]. While “databases” (22 occurrences) and “data” (43 occurrences) indicated the increasing dependence on structured information for decision making, the development of Computer-Aided Design (CAD) tools and safety software, like Primavera P3, permitted more efficient project scheduling and risk analysis [49]. Terms like “artificial intelligence” (7 occurrences) and “automated systems” (14 occurrences) indicate early attempts to incorporate intelligent technologies for safety management, thereby laying the foundation for later AI-based safety solutions.

The green cluster established by “management” (142 occurrences), “occupational safety” (50 occurrences), and “decision-making” (37 occurrences), emphasizes the increasing focus on structured safety strategies during this decade. The predominance of “education” (23 occurrences) and “personnel training” (9 occurrences) reflects the requirement of equipping staff with the skills and information necessary to function safely in high-risk building situations [50,51]. Furthermore, underscored by the prevalence of “healthcare” (36 occurrences) and “human” (88 occurrences) is the growing attention to the mental and physical health of building workers, thereby indicating a shift toward more comprehensive methods of occupational safety. Additionally, “Evaluation” (21 occurrences) also proved to be important in determining the effectiveness of safety regulations, thereby facilitating ongoing risk management enhancements [52], while the term “government” (20 occurrences) highlights the impact of regulatory authorities in developing safety rules.

Driven by the industry’s efforts to manage hazardous materials, pollution, and sustainability issues, the blue cluster highlights the increasing importance of environmental safety during this period. Keywords such as “environment” (111 occurrences), “safety” (156 occurrences), and “hazards” (50 occurrences) capture the increased awareness of hazards presented by environmental contamination and dangerous products. Moreover, the 132 occurrences of “materials” and the 18 occurrences of “construction materials” emphasize the importance of selecting and managing materials properly to minimize environmental impact and health risks. Additionally, keywords such as “pollution” (20 occurrences) and “radioactivity” (9 occurrences) underscore the industry’s commitment to addressing environmental hazards and ensuring greater sustainability in the construction process. To ensure compliance with new environmental regulations and pave the way for safer and more sustainable construction processes, “assessment” (96 occurrences) was crucial in evaluating the sustainability and safety of materials and techniques [53].

The yellow cluster highlights the standards and regulatory frameworks that shaped safety practices during this era. Keywords such as “standards” (72 occurrences), “building codes” (34 occurrences), and “compliance” (14 occurrences) illustrate the increasing emphasis on aligning construction methods with international safety standards. Regulatory milestones, such as the 1996 OSHA Scaffold Standards and the 2000 Construction Health and Safety Plan Directive in the European Union, established more stringent criteria for worker safety and project management [54]. These procedures were enhanced by improvements in inspection and monitoring, as evidenced by the term “inspection” appearing 21 times. The term “planning” (95 occurrences), associated with the purple cluster, emphasizes the importance of incorporating strategic planning into construction operations, ensuring that safety standards are systematically integrated into both the design and execution stages of projects.

3.2.4. Subperiod 2006–2015

Between 2006 and 2015, research began to shift beyond traditional accident prevention and regulatory compliance, focusing on digital technologies, real-time hazard monitoring, and environmental impact assessments. Figure 8 depicts the resulting keyword co-occurrence network, which illustrates these evolving themes and interconnections.

While the dominant research focus remained safety and accident prevention, similar to previous decades (green and dark blue clusters), technological advancements (purple and orange clusters) gained prominence, including real-time monitoring tools, building information modeling (BIM), and simulation. In addition, the field saw a growing focus on sustainability and environmental impact (red cluster), indicating a continued emphasis on long-term hazard mitigation strategies. Moreover, occupational health, worker safety, and ergonomics (yellow cluster) emerged as key research areas, reflecting concerns about human factors and construction site conditions.

Based on Figure 7, the green cluster, which includes terms such as “safety” (366 occurrences), “modeling” (70 occurrences), and “simulation” (24 occurrences), reflects the increasing reliance on computational tools to enhance safety practices in construction. The integration of modeling and simulation techniques enabled researchers to test safety scenarios and develop predictive analytics, thereby improving risk assessment strategies and hazard prevention during this period [55]. Furthermore, the rise of “RFID” (15 occurrences), “GPS” (5 occurrences), and “location tracking” (5 occurrences) keywords underscore the early application of real-time tracking and automation technologies to improve safety monitoring and worker location tracking on construction sites [56]. These digital advancements significantly improved risk mitigation and streamlined safety management [57]. On the other hand, the emergence of keywords such as “virtual reality” (9 occurrences), “visualization” (10 occurrences), and “laser scanning” (6 occurrences) highlights the initial adoption of immersive technologies for safety training and hazard recognition in this decade [58,59]. Virtual reality (VR) has been particularly impactful in enhancing safety training, offering immersive simulations that enable workers to experience hazardous scenarios in a controlled environment [60]. Similarly, laser scanning technology facilitated site inspections and hazard detection, significantly reducing the need for manual safety assessments in hazardous conditions [61]. Finally, keywords such as “computer-aided design” (5 occurrences) and “PPE” (6 occurrences) align with the ongoing efforts during this period to improve worker safety as they were employed to minimize human errors and provide enhanced hazards protection.

The brown and orange clusters represent a technological transformation in construction safety research for this subperiod, characterized by the integration of automation, digital information management, and real-time monitoring. The widespread adoption of BIM (25 occurrences) and Geographic Information Systems (GIS) (12 occurrences), and the introduction of wireless sensor networks (10 occurrences) and sensor-based safety tracking (7 occurrences) enabled risk assessment, hazard prediction, and real-time site monitoring, significantly improving project safety and coordination [62,63,64]. Moreover, the roles of construction automation (20 occurrences) and robotics (6 occurrences) became critical in reducing manual labor risks, demonstrating how automated crane operations (19 occurrences) can mitigate risks associated with lifting hazards. Furthermore, the keywords “excavation” (8 occurrences) and “slope stability” (5 occurrences) appeared to reflect efforts to automate excavation and minimize geotechnical hazards [65,66]. Lastly, the presence of “inspection” (12 occurrences) and “image processing” (6 occurrences) in this cluster highlights how automation and image-processing technologies played a crucial role in enabling AI-driven analysis of site images to detect hazardous conditions in real time [67].

The red cluster highlights key themes, including sustainability (33 occurrences), the built environment (108 occurrences), physical activity (13 occurrences), urban design (9 occurrences), and public health (6 occurrences), which demonstrate a shift toward integrating safety with urban planning and environmental considerations. Research conducted during this time examined the impact of walkability, traffic safety, and site accessibility on construction risks, with studies highlighting the role of urban infrastructure in mitigating workplace hazards [68]. Nevertheless, sustainability-driven construction, while beneficial, introduces new risks, such as exposure to hazardous materials and increased fall hazards. This is due to the use of innovative or recycled materials with unknown safety profiles and the extended periods that workers spend at height compared to traditional projects [69]. These evolving hazards highlighted the need to evaluate building techniques from a broader, long-term perspective. To address these issues, the increased focus on life cycle assessment (10 occurrences) and material selection (6 occurrences) reinforced efforts to integrate long-term safety planning into construction practices [70]. In addition, keywords such as “developing countries” (7 occurrences) and “regulations” (7 occurrences) highlighted the ongoing challenge of implementing safety measures in regions with less stringent enforcement policies. Thus, this period’s research investigated how regulatory compliance in low- and middle-income nations was frequently affected by informal labor markets, inadequate enforcement, and unsafe building practices [71]. Overall, this cluster built upon the regulatory and technological advancements of previous decades, emphasizing the need for holistic planning approaches that balance urban development, worker well-being, and regulatory compliance to create safer and more resilient construction environments.

The blue and yellow clusters highlight the interconnected themes of risk management (115 occurrences), construction safety (179 occurrences), and accident prevention (44 occurrences), which highlight structured safety assessments and proactive hazard control measures. Studies during this period advanced occupational risk models, quantifying workplace hazards such as falls (6 occurrences), scaffold collapses (5 occurrences), and fire incidents (34 occurrences) to improve mitigation strategies [72]. This aligns with prior decades’ efforts to systematize safety regulations but with a stronger reliance on data-driven risk analysis [73]. Moreover, the rise of safety culture and safety climate (9 and 8 occurrences, respectively) highlighted the industry’s shift toward workplace attitudes and training initiatives (12 occurrences) as key factors in reducing injuries [74,75,76,77]. In addition, ergonomics (8 occurrences) became more prominent as musculoskeletal disorders among construction workers gained attention, particularly in scaffold-related work [78,79]. Lastly, the effects of climate change (10 occurrences) and flooding (5 occurrences) on construction safety introduced new concerns about extreme weather events and infrastructure resilience, further reinforcing the importance of risk assessment and proactive planning in ensuring worker safety [80,81]. Consequently, this period solidified risk-based safety management and workplace culture improvements as fundamental strategies for reducing construction-related injuries and fatalities.

In addition to the previous keywords, this subperiod includes keywords such as “monitoring” (18 occurrences) and “knowledge management” (11 occurrences), which reinforce earlier efforts of real-time hazard detection and data-driven decision making. Furthermore, as part of the purple cluster, the use of ontologies (9 occurrences) in safety planning helped standardize health and environmental risk assessments by bridging the gap between safety regulations and operational practices [23]. Moreover, human factors (8 occurrences) became a key concern, with studies identifying fatigue (5 occurrences) and cognitive stress as major contributors to accidents and reduced worker performance [82,83,84]. Thus, these findings align with previous decades’ focus on behavioral safety, predictive monitoring, and knowledge management.

3.2.5. Subperiod 2016–2025

Between 2016 and 2025, research in construction safety evolved significantly, shifting toward AI-driven technologies, digital automation, and real-time hazard detection, and moving beyond traditional risk management frameworks. Figure 9 illustrates the industry’s increasing reliance on machine learning, knowledge graphs, and extended reality for enhanced safety outcomes. Furthermore, emerging issues such as gender inequities, safety challenges in developing countries, and sustainability concerns demonstrate the expanding breadth of building safety research.

The brown and pink clusters in Figure 8 highlight the dominance of safety (788 occurrences), construction (579 occurrences), safety management (256 occurrences), and accident (102 occurrences), which reinforces the industry’s ongoing commitment to reducing workplace hazards. Keywords such as “ontology” (25 occurrences), “knowledge graphs” (23 occurrences), and “text mining” (21 occurrences) reflect the industry’s increasing reliance on artificial intelligence and machine learning for hazard identification and accident prevention. Thus, several studies integrated knowledge graph and text mining technologies to enhance hazard detection, build predictive safety models, and ensure regulatory compliance [85,86]. Moreover, the rise of Industry 4.0 (23 occurrences) and digital technology (19 occurrences) has further transformed safety management, with real-time monitoring and automation reducing workplace injuries [87]. However, barriers such as training gaps, implementation challenges, and resistance to technology remain as key challenges to digital adoption in construction safety [88]. Lastly, the keywords “developing countries” (29 occurrences) and “gender” (16 occurrences) emerged as notable themes, reflecting the increased safety risks faced by female construction workers and the workforce in developing countries. Several studies highlight the challenges faced by women in the construction industry, including a lack of gender-specific personal protective equipment, physically demanding tasks, and workplace harassment [89,90]. On the other hand, safety issues in developing countries remain a concern, with challenges related to weak regulations, inadequate PPE access, and lack of safety education [91].

Red and orange clusters highlight major technological advancements, including BIM (207 occurrences), Internet of Things (IoT) (82 occurrences), automation (64 occurrences), and digital twins (53 occurrences), reinforcing the industry’s transition toward intelligent and real-time safety management. Thus, the adoption of BIM and IoT-driven safety monitoring systems has significantly enhanced hazard detection and construction site efficiency [92,93]. Similarly, knowing that a digital twin is the result of integrating BIM and IoT technologies, it has also been employed for real-time project monitoring, simulating potential risks, and enhancing safety decision making [94]. Moreover, with the need for secure, transparent, and trustworthy platforms for the generated data, blockchain (28 occurrences) and edge computing (14 occurrences) have emerged as critical tools in ensuring data security and decentralized safety management systems, providing real-time, tamper-proof safety records [95,96]. Furthermore, extended reality (XR) technologies, such as virtual reality (VR) (74 occurrences), augmented reality (AR) (28 occurrences), and mixed reality (MR) (15 occurrences), have been experiencing increased adoption. This increased attention follows earlier implementations of simulation and visualization models during previous periods and is primarily due to their ability to revolutionize safety training, enhance hazard awareness, and facilitate immersive risk simulations, thereby improving worker preparedness for high-risk scenarios [97,98]. Lastly, the clusters include other keywords such as “drones” (28 occurrences), unmanned aerial vehicles (36 occurrences), and “point cloud” technologies (20 occurrences), which reflect the employment of such innovations in reducing worker exposure to unsafe conditions through automated inspections and safety assessments of hazardous environments [99].

Purple and dark blue clusters denote significant advancements in machine learning (149 occurrences), deep learning (195 occurrences), and computer vision (95 occurrences), marking a transition toward real-time, AI-driven safety systems. These technologies have been widely adopted for the automatic detection of unsafe behaviors, hazard identification, and worker safety monitoring [100], building upon earlier digital modeling and risk prediction approaches from previous decades. Thus, the integration of robotics (65 occurrences), autonomous vehicle (17 occurrences), and reinforcement learning (16 occurrences) has further optimized automated safety monitoring and collision avoidance systems, especially in prefabrication and modular construction [101]. Additionally, object detection (55 occurrences) and CNN-based safety analytics (52 occurrences) have enhanced construction site surveillance by reducing the risks associated with heavy machinery and workforce interactions [102]. Moreover, the increased focus on wearable devices (42 occurrences) and PPE (40 occurrences) signifies an effort to integrate biometric monitoring and real-time hazard detection into safety compliance frameworks, particularly for high-risk activities like falls (38 occurrences) and confined space operations [103]. These findings extend the 2006–2015 focus on monitoring and risk assessment, now enhanced by AI-powered real-time detection systems. Lastly, fire safety (97 occurrences), optimization (73 occurrences), and tower crane safety (18 occurrences) have also been strengthened through AI-driven simulations and predictive modeling, reducing worksite hazards and improving safety decision making [104,105]. Overall, these advancements build upon previous decades’ regulatory frameworks and digital modeling efforts, which have been revolutionized by deep learning and automation technologies.

Building upon earlier advancements that aimed to shape workplace conditions, environmental safety, and risk mitigation strategies, the green and light blue clusters include some of the most occurring keywords in the network, including “built environment” (170 occurrences), “health and safety” (167 occurrences), “sustainability” (102 occurrences), and “construction management” (81 occurrences). Moreover, it discusses the impacts of COVID-19 (64 occurrences) on construction workers’ safety, hygiene, and awareness, which are reflected in the increasing concern for occupational health (14 occurrences), mental health (13 occurrences), and worker fatigue (13 occurrences) as responses to workplace stressors, pandemic-related safety measures, and ergonomic risk factors [106,107]. In addition, the inclusion of keywords such as walkability (20 occurrences) and physical activity (19 occurrences) in construction safety studies further emphasizes the role of ergonomic site layouts and movement-friendly environments in reducing fatigue-related accidents and improving worker performance [103]. Furthermore, in line with past data-driven advancements (1996–2005, 2006–2015), keywords such as GIS (27 occurrences), simulation (47 occurrences), and bibliometric analysis (23 occurrences) have been employed to map hazard-prone areas, track construction safety research trends, and develop predictive models for risk assessment [108]. Lastly, the integration of multi-criteria decision-making tools (14 occurrences), such as the Analytic Hierarchy Process (AHP) (62 occurrences), describes the continuing industry’s reliance on quantitative decision-making tools from previous decades in structuring safety risk assessments.

For the final cluster (yellow), risk management (334 occurrences) and reliability (99 occurrences) remain central to modern safety frameworks, advancing prior efforts in accident prevention and hazard mitigation [109]. Thus, recent research incorporates neural networks (31 occurrences) and deep learning-based risk assessment models to optimize predictive safety analyses for both material and human risk factors in construction sites [110]. In addition, this cluster reinforces the research focus on the stability and risk evaluation of temporary structures, as evident through such keywords as temporary structures (16 occurrences), progressive collapse (21 occurrences), and scaffolding (18 occurrences). Meanwhile, slope stability (20 occurrences), deep excavations (17 occurrences), and collapse (18 occurrences) highlight concerns about geotechnical safety, soil stability, and excavation hazards, which are an extension to previous work on hydraulic safety and excavation risk management (1996–2015) [111]. Lastly, the role of human error (16 occurrences) remains a significant safety challenge, with models identifying and integrating safety risk factors (30 occurrences) into AI-driven safety assessments [112].

4. Discussion

The following discussion synthesizes key findings from the review, highlighting the evolution of global construction safety frameworks, major research trends, and transformative technological shifts. It contextualizes the development of safety standards and research within historical, regulatory, and technological milestones that have shaped modern construction safety practices.

4.1. Major Regulatory Frameworks in Construction Safety

Formal regulatory frameworks have played a pivotal role in enhancing construction safety worldwide, with significant legal milestones unfolding chronologically across various regions. In the United States, the defining moment was the Occupational Safety and Health Act of 1970, which led to the establishment of OSHA in 1971. OSHA established mandatory safety standards and enforced compliance through inspections, dramatically lowering workplace fatality rates in construction by about 60% since its inception. This success inspired other nations to adopt similar regulatory structures. Thus, in the United Kingdom, the Health and Safety at Work Act of 1974 was introduced, creating the Health and Safety Executive (HSE) to enforce a comprehensive, goal-oriented safety system. UK legislation imposed clear responsibilities on employers to maintain safe workplaces. Its impact has been substantial, where annual construction-related fatalities have dropped by nearly 90% since the mid-1970s, illustrating the direct benefits of structured safety governance [113].

Moreover, the International Labor Organization advanced baseline standards with the 1988 Safety and Health in Construction Convention. This set of international guidelines, encourages countries to require hazard identification, worker training, and the use of proper protective equipment. Likewise, the European Union implemented broad-reaching safety legislation, most notably the Framework Directive of 1989, which mandated systematic workplace risk assessments across all EU member industries. In addition, specifically addressing construction, the EU Temporary or Mobile Construction Sites Directive 92/57/EEC was adopted in 1992, emphasizing proactive safety management from the early project-design stages and embedding principles aligned with what is now recognized as the principle of “prevention through design” in construction projects [114]. Consequently, research indicates that such regulatory interventions contributed to an overall decline in construction accident rates in many EU countries by emphasizing preventive management of safety risks [32].

Asian countries followed suit, developing rigorous construction safety legislation alongside rapid industrial growth. Japan introduced its Industrial Safety and Health Act in the early 1970s, while South Korea passed its Occupational Safety and Health Act in 1981, continually updating it to improve enforcement practices [20]. China’s experience was more complex, evolving from minimal early safety measures in the 1950s and 1960s, when fatality rates were extremely high, to the establishment of structured, comprehensive safety regulations by the 1990s [115]. Moreover, through major laws enacted between 1995 and 2002 (i.e., the Labor Law in 1995, the Construction Law in 1998, and the Work Safety Law in 2002), China established comprehensive duties for construction employers, contractors, and owners, and empowered agencies to enforce site safety [115]. However, although ongoing challenges persist due to the scale of its construction activities, China is steadily working to improve its safety metrics. Other rapidly developing regions, such as Singapore and countries in the Gulf region, have more recently adopted similar stringent regulations, often modeled after the proven frameworks developed by OSHA, EU directives, and ILO standards.

These global efforts have not only made construction sites safer but also driven innovation in safety technology, better training practices, and detailed safety management systems. Overall, the evolution of formal safety regulations worldwide has substantially reduced construction risks, reshaped industry practices, and continuously influenced new directions in occupational safety research.

4.2. Chronological Shifts in Construction Safety Research Themes

The evolution of construction safety research from 1972 to 2025 is characterized by an exponential growth in the volume of published papers as well as the thematic sophistication of the discipline. The longitudinal bibliometric analysis revealed a dramatic increase in construction safety research output from 1972 through 2025. Early in the 1970s and 1980s, relatively few studies were published annually, whereas the 2000s and 2010s saw an exponential surge. For example, the number of papers published from 1972 to 2000 is nearly equal to those published between 2001 and 2010 alone, reflecting heightened scholarly and industry attention to construction safety in recent decades. As a result, research emphasis has evolved significantly over time, shifting from fundamental safety standards and accident prevention to more complex, technology-driven strategies.

The analysis also reveals that research contributions are globally distributed but dominated by a few countries, including China, the United States, the United Kingdom, and Australia, which consistently lead the contributions. This leadership aligns with those nations’ large construction industries and strong research funding. Thus, China has led the way in research on automation and robotics, owing to its rapid urbanization and large-scale construction projects [6]. The United States, through OSHA standards, has significantly contributed to shaping regulatory environments and developing predictive safety systems. In addition, the European Union has been influential in the development of global safety standards through the implementation of rule-making models, such as the Temporary or Mobile Construction Sites Directive, thereby promoting a shared approach to ensuring worksite safety and protecting worker well-being [32].

As shown in Figure 10, during the period from 1972 to 1985, construction safety research focused on the development of fundamental safety paradigms and the establishment of regulatory requirements. This period was marked by the establishment of OSHA in the United States and other regulatory bodies, which laid the groundwork for improving construction safety and informed research efforts. Early research was largely reactive in nature, addressing short-term safety issues such as accident prevention, machinery maintenance, and the development of the earliest safety standards to mitigate risks on construction sites [21,28]. The thematic clusters from this period confirm a strong emphasis on machinery safety, fire safety standards, and the early recognition of human and behavioral factors influencing workplace safety.

Between 1986 and 1995, a significant shift occurred toward more structured safety regulations and the initial incorporation of technological innovations. Studies during this period further broadened their scope of conventional accident prevention to include more sophisticated risk management strategies, the design of safety automation systems, and the application of computational tools for hazard identification. It marks the beginning of a shift from reactive to more proactive safety measures, thereby laying the groundwork for the subsequent development of construction safety. The establishment of regulatory policies (e.g., OSHA’s fall protection standards enforced in 1994, and the EU’s Temporary or Mobile Construction Sites Directive adopted in 1992) was a driving force influencing the research on safety in this era [31,32]. Additionally, technological integration began emerging more notably, with the adoption of automation and computational tools (e.g., CAD and early robotics), signaling the industry’s initial steps toward digital transformation.

Furthermore, the decade from 1996 to 2005 was a transformative phase of evolution in the discipline, marked by the growing significance of technological innovations, including CAD and safety simulations. The era witnessed a heightened scholarly production, alongside a greater emphasis on how technology could enhance safety outcomes. Investigations into the feasibility of predictive safety models and real-time monitoring systems were among the trending topics as scholars sought to mitigate risk in construction environments. The increasing complexity of construction projects during this period necessitated the application of more sophisticated tools for managing safety, thereby incorporating early risk management technologies and automated safety systems [46,47].

Between 2006 and 2015, construction safety research notably expanded beyond traditional frameworks by increasingly incorporating advanced digital technologies such as BIM, wireless sensor networks, and early-stage automation. These technologies revolutionized safety planning, risk minimization, and site management. BIM allowed for improved visualization of construction sites, facilitating better hazard identification and safety analysis. Automation and robotics also reduced the exposure of workers to hazardous tasks by automating dangerous operations. This era witnessed an increased recognition of the significance of environmental and sustainability issues in relation to construction safety, resulting in a greater emphasis on environmental hazards associated with construction operations.

Finally, the period from 2016 to 2025 witnessed a revolution in the construction safety sector, marked by the massive adoption of new technologies, including AI, XR, IoT, digital twins, and wearable devices. These inventions enabled real-time hazard detection, predictive safety modeling, and enhanced monitoring systems. The integration of AI and IoT technologies enabled the development of more intelligent and efficient safety management systems, capable of detecting risks in real time and suggesting preventive measures. Simultaneously, there was an increased emphasis on human factors related to safety, specifically mental health, fatigue, and employee behavior, toward a more integrated safety management approach that combines technological advancements with consideration of employee welfare. The increase in the number of papers during this period, which totaled 8764 publications, reflects the growing global recognition of the significance of construction safety as a critical element in the construction process. The practice has shifted from a reactive, accident-minimization approach to a proactive, data-driven approach that integrates advanced technologies, emphasizing prediction, detection, and prevention of safety hazards. The latest advancements in AI-powered safety measures, real-time IoT monitoring, and predictive modeling are transforming construction safety, making it more efficient, effective, and better equipped to handle the increasing complexity of contemporary construction projects.

4.3. Analysis of Evolutionary Patterns Across Historical Periods

Academic research on construction safety has undergone significant evolution over the past several decades, characterized by distinct shifts in focus, methods, and theoretical foundations. Initially, in the 1970s and earlier, construction safety research was predominantly reactive, where researchers analyzed accidents after they occurred, focusing on identifying immediate causes and basic preventive measures. During this period, much of the scholarship was qualitative or descriptive. It documented prevalent hazards and recommended commonsense interventions (such as hard hats, guardrails, and improved training) to curb the high injury rates. Nevertheless, some researchers began recognizing the importance of human and organizational factors beyond individual worker mistakes, thus laying the groundwork for future studies. Therefore, by the late 1980s and into the 1990s, a significant shift occurred from a reactive, incident-focused perspective to a proactive, systems-based approach [6,10,11,12]. This transformation was largely driven by the introduction of stronger safety regulations in the U.S., Europe, and globally, which mandated risk management practices and preventive measures rather than just post-accident responses. Research in this era has increasingly focused on structured approaches to hazard identification, risk assessment, and safety planning at early project stages, contributing to the development of the prevention through design approach [10]. In addition, researchers began to quantitatively evaluate organizational elements, such as safety culture and managerial commitment, and found a link between positive safety cultures and reduced accident rates. Thus, this period also saw the rise of statistical methods to analyze accident data, aiming to predict and prevent incidents rather than respond to them after the fact.

Entering the 2000s, and particularly advancing into the 2010s and 2020s, construction safety research underwent another transition, now emphasizing technology-driven and data-centric strategies. The introduction of digital tools, such as Building Information Modeling (BIM), has enabled researchers to simulate construction processes and proactively identify potential safety hazards [22]. Studies explored automated hazard analysis through digital modeling and virtual simulations, enabling the identification and correction of potential risks before construction even began. Furthermore, rapid technological advances, including the Internet of Things (IoT), wearable sensors, real-time monitoring, and big data analytics, opened entirely new avenues for real-time safety management [24]. Additionally, recent research is increasingly leveraging artificial intelligence (AI) and machine learning to automatically identify hazards, predict accident risks, and provide intelligent safety interventions in real time, significantly enhancing predictive safety capabilities.

Alongside these technological advancements, construction safety research has also become highly interdisciplinary, integrating insights from psychology, data science, public health, and computer engineering. Modern studies often involve collaborations among engineers, statisticians, behavioral scientists, and computer scientists, reflecting a holistic and rigorous methodological approach. Researchers now use advanced methods such as controlled experiments, network analysis, machine learning models, and systematic reviews or meta-analyses, greatly enriching the depth and validity of safety insights [6,11,12].

4.4. Summary

In summary, construction safety research has undergone significant evolution over time, reflecting broader industry and regulatory developments. It has transitioned from basic reactive analyses of accidents to proactive, systems-oriented approaches, and now to predictive, technology-intensive solutions. Each phase has built upon and enriched the previous one, creating a diverse and interdisciplinary field dedicated to continuously reducing workplace hazards and striving for accident-free construction sites.

5. Future Directions

As discussed earlier, the field of construction safety research is evolving toward proactive, predictive, and technology-driven safety measures. The sector is adapting to new paradigms that enhance risk assessment, compliance, and worker protection, as AI, automation, digital safety systems, and sustainability initiatives continue to evolve. The following study themes are key areas that have the potential to significantly influence the future development of construction safety, as illustrated in Figure 11.

5.1. AI-Enhanced Safety Management and Predictive Hazard Detection

Artificial intelligence has proved its potential in transforming construction and occupational safety management by progressively automating hazard identification, risk assessment, and real-time safety monitoring. Thus, with AI-driven analytics and its ability to scrutinize enormous volumes of site data from sensors, worker behavior analysis, and digital twins, it can provide proactive safety interventions and early warning systems to mitigate risks before incidents occur [86]. As a result, future research should further explore the integration of AI into decision support systems, focus on the interpretability and trustworthiness of AI-generated safety advisories, and discuss the ethical and legal implications of AI-enabled workforce monitoring [85]. Moreover, advancements in computer vision and natural language processing technology are expected to continue enhancing automation in safety analysis and regulatory enforcement monitoring, leading to improved adaptive risk management strategies [100].

5.2. Digital Twins and IoT-Enabled Real-Time Safety Monitoring

Digital twin technology, such as the integration of BIM and IoT, has the potential to transform construction safety management by enabling proactive safety planning, real-time risk analysis, and virtual hazards [92,93,116]. By integrating sensor data, environmental conditions, and worker locations, digital twins can create dynamic, real-time representations of construction sites that identify potential safety hazards and facilitate predictive safety assessments [94]. However, challenges in data fusion accuracy, computing efficiency, integration with current safety standards, and the scalability of digital twin frameworks across projects of varying complexity still impede widespread adoption in safety management [117]. Consequently, future research should prioritize the development of advanced data harmonization techniques, the optimization of real-time processing capabilities, and the establishment of standardized integration frameworks to enhance the practical implementation of digital twins in construction safety. Additionally, investigating strategies to enhance stakeholder acceptance, training approaches, and cost-effectiveness will be crucial for the efficient implementation of digital twin solutions in various construction settings.

5.3. Automation, Robotics, and Wearable Technologies for Risk Reduction

Automation and robotics are poised to revolutionize construction safety by reducing worker exposure to high-risk activities, repetitive strain tasks, and hazardous environments [101]. Autonomous material handling, drone-based site inspections, and AI-integrated heavy machinery are examples of robot technologies that enable remote operation, predictive maintenance, and enhanced situational awareness, thereby improving both worker protection and overall productivity [99]. Furthermore, to prevent accidents caused by exhaustion and to monitor employee health in real time, wearable exoskeletons, smart PPE, and biometric health monitoring devices can play a critical role in reducing such accidents by minimizing fatigue and providing real-time health assessments [103]. Nevertheless, despite these advancements, several challenges remain, including limited interoperability across different systems, high implementation costs, worker resistance due to a lack of training or trust, and data privacy concerns [84]. These issues will require more in-depth research to ensure effective human–robot collaboration in safety-critical tasks, improve user adaptability in various construction environments, and enhance cybersecurity in networked safety devices.

5.4. Blockchain and Smart Contracts for Safety Compliance and Data Integrity

Smart contracts and blockchain technology have the potential to enhance construction safety compliance by automating auditing cycles, ensuring the validity of worker certifications, and providing secure, immutable records of safety protocols [95]. By leveraging blockchain’s immutable characteristics, construction companies can enhance regulatory compliance, reduce the risk of fraudulent safety documents, and improve the traceability of incident reporting. Moreover, blockchain-based decentralized compliance solutions may promote cross-border regulatory compliance, enhance real-time event transparency, and establish accountability for liabilities, thereby reducing human error and improving safety governance [96]. However, further research is needed to overcome the challenges that hinder blockchain’s adoption, including scalability limitations, integration complexity with existing construction management systems, and legal ambiguities surrounding decentralized compliance systems, particularly in large-scale megaprojects [118]. Thus, future research should focus on developing scalable blockchain architectures for real-time safety data processing, standardization frameworks for industry-wide adoption, and legal models that address jurisdiction conflicts in decentralized compliance enforcement.

6. Conclusions

Through a longitudinal bibliometric review, this research offered a comprehensive evaluation of the evolution of construction safety research from 1972 to 2025. The analysis of the collected articles highlighted significant thematic shifts, legislative developments, and technological integrations that have shaped safety management practices globally. Over the past five decades, the research field has transitioned from initial reactive hazard control approaches, characterized primarily by regulatory compliance and basic safety standards, to sophisticated, proactive strategies driven by innovations in artificial intelligence, the internet of things, digital twins, robotics, and real-time monitoring technologies. This transition reflects both industry’s demands and regulatory developments in response to the increasingly complex construction environment. Moreover, with contributions from 134 nations, the extensive global engagement identified demonstrates that there has been widespread international recognition of construction safety as a critical discipline. China, the United States, the United Kingdom, Australia, and Canada emerged as the top leading contributors. Thus, these countries exemplify the correlation between robust regulatory frameworks, robust research infrastructure, and high scholarly output. China’s leadership is a result of the country’s proactive approach toward construction safety in response to its fast urbanization and infrastructure expansion. At the same time, OSHA’s persistent impact highlights the United States’ vital role in creating international occupational safety standards.

The study also revealed and highlighted that several gaps and challenges in construction safety research still persist. Consequently, future studies should focus on reducing obstacles to technology adoption, including enhancing cybersecurity in IoT-enabled monitoring systems, improving compatibility between AI-driven tools and existing regulatory systems, and facilitating human–robot collaboration in safety-critical tasks. Furthermore, for greater adoption and stronger risk management, scalable and standardized frameworks for blockchain applications and digital twins are essential. Moreover, further interdisciplinary research into human factors, ergonomic considerations, and sustainable construction practices is essential for thoroughly addressing emerging risks in dynamic construction environments.

While this study provided a comprehensive longitudinal bibliometric overview, several limitations must be acknowledged to contextualize the findings. First, the analysis was constrained by the data sources and search strategy. Similar to other bibliometric studies, this research focused on indexed journal articles available in the Scopus database. Consequently, important information included in books, conference papers, technical reports, or industry publications, in various forms or databases, may have been overlooked. Moreover, the reliance on database indexing means that some of the early or non-English literature, especially from the 1970s and 1980s, may not have been captured if it has not been digitally archived. Additionally, the keyword strings used in the search and filtering operations may have omitted certain pieces of relevant literature due to the extensive volume of identified publications. Finally, it is important to note that bibliometric analysis, by definition, focuses on published academic research. There is a broader context of on-the-ground safety improvements, technological advancements, and policy changes that occur outside of academia. Consequently, although our research identifies scholarly trends, they do not necessarily translate into proportional enhancements in construction safety outcomes in practice.

In summary, while construction safety research has undergone significant evolution, driven by technological innovations and shifting regulatory demands, ongoing investment in interdisciplinary research and international collaboration remains crucial. Addressing current limitations and exploring the outlined future directions will ultimately enhance safety outcomes, protect worker well-being, and ensure resilience within an increasingly complex and digitally integrated global construction industry.