Search Results (310)

Construction accident reports contain rich causal evidence; however, their unstructured narratives make systematic analysis difficult. Recent advances in large language models (LLMs) have created new opportunities to leverage such information at scale. This study develops an integrated LLM–BERT–network framework for analyzing construction accident causation. Based on 347 official construction accident investigation reports, a DeepSeek-based pipeline with human-in-the-loop quality control was used to extract causal keywords describing direct and indirect causes, yielding 2572 keywords. A BERT-based semantic normalization procedure then consolidated synonymous expressions, reducing 811 deduplicated keywords to 104 normalized terms (an 87.2% reduction in vocabulary size). A manual sample-based evaluation further supported the reliability of the LLM-based extraction and BERT-based normalization procedures. The normalized keywords were further organized into a hierarchical taxonomy and used to construct a directed keyword-association network linking indirect and direct causes for structured relational analysis. To strengthen methodological rigor, additional validation and analytical experiments were conducted, including manual sample-based evaluation of keyword extraction, sensitivity analysis of normalization settings, and examination of representative failure cases. The results support the reliability and robustness of the proposed framework. The analysis indicates that behavior-related factors and management deficiencies occupy structurally important positions in the directed network. Overall, the findings suggest that construction accidents arise from the interaction of human, managerial, environmental, material, and technical factors rather than isolated single causes. Effective prevention therefore requires system-oriented interventions that strengthen worker competence, supervision, training, accountability, and hazard identification. Full article

Background: Burn injuries have both physical and psychological impacts on patients. Factors such as personal beliefs, prior experiences, and geographic, economic, or cultural barriers, as well as fear of hospitals, can contribute to delays in seeking specialized care. When combined with inadequate first aid or the inappropriate use of pharmaceutical or traditional remedies, these delays may worsen burn severity, prolong healing, and negatively affect quality of life. From a clinical perspective, delayed presentation following burn injury has been linked to burn wound progression, which increases the risk of local infection, hypertrophic scarring and prolonged functional impairment. Methods: This analytical cross-sectional study was conducted at the Clinical Emergency Hospital of Bucharest between January and September 2025. The primary objective was to characterize adult burn patients presenting more than 24 h after injury (Group A) and to describe self-reported psychosocial/behavioral characteristics and explore unadjusted patterns among delayed presenters. Data were collected from medical records and a structured questionnaire administered to delayed presenters. A secondary descriptive comparison was performed with patients presenting within 24 h (Group B) to provide contextual reference. Results: The majority of patients were male (62.2%) and of working age (18–65 years, 82.4%). Thermal burns from domestic accidents were most common (58.8%), with scalds predominating. Second-degree burns were the most frequent (60.5%), primarily affecting the upper and lower limbs. Mean total body surface area (TBSA) was low (2.86 ± 1.91%), although higher values were observed in radiation burns and closed-space accidents. More than half of the patients did not receive any first aid, while the remainder used various pharmaceutical or natural products, some of which were inappropriate for burn treatment. The main reasons for delaying specialized care were the expectation that injuries would heal spontaneously, negligence, and fear of the hospital. In contrast, escalating pain, edema, and family insistence were the primary motivators for seeking professional medical attention. Delayed presentation was associated with older burn lesions, higher burn severity and an increased likelihood of hospitalization or refusal of recommended admission. Conclusions: Burn injuries predominantly affect working-age males and most frequently arise from domestic thermal accidents. Delayed presentation and inadequate first aid are common and influenced by behavioral, social, and demographic factors. Targeted public education, improved first aid practices, and timely healthcare-seeking are essential to reduce burn severity and improve patient outcomes. Full article

The sustainable development of university safety governance is an important component of the national security management system and also serves as a fundamental safeguard for protecting the life and health of students and staff on campus. The improvement of university safety risk governance relies on analyzing the identification of various safety risks and maintaining an effective crisis management process for potential sudden safety risks. The 24Model and the 4R model have respectively demonstrated strong analytical advantages in the fields of accident causation analysis and emergency crisis management; however, few studies have examined the internal relationship between them. This study attempts to integrate the 24Model and the 4R crisis management framework to propose and analyze a 2-4-4R model for university safety risk management. Through a case study, the model is applied to analyze a laboratory explosion accident at a university. The results show that the risk factors leading to campus safety accidents can be analyzed from four aspects: safety culture, safety management system, individual factors, and unsafe acts and physical conditions. University safety management should comprehensively identify these four types of factors and propose governance measures sequentially from the four stages of reduction, readiness, response, and recovery in order to improve safety management capacity. The case analysis confirms that the 2-4-4R model has applicability and practical value in the identification and governance analysis of university safety risks. It provides a systematic research perspective for the identification and management of safety risks in universities, and is of great significance for promoting the sustainable development of universities. Full article

Traffic accidents are among the leading causes of injury worldwide, highlighting the urgent need to better understand the factors that contribute to accident occurrence and severity in order to improve road safety and reduce injuries and fatalities. This study analyzes the U.S. Accidents dataset, comprising data collected from 2016 to 2023, to identify the key determinants of accident severity and to evaluate feature-selection techniques for predictive modeling. To this end, several feature-selection methods are examined, including L1-regularized logistic regression, minimum redundancy maximum relevance (mRMR), conditional mutual information maximization (CMIM), ReliefF, and tree-based importance measures; these are compared with the Weighted Conditional Mutual Information (WCFR). The selected feature subsets are then evaluated using three machine learning models: logistic regression, random forest, and XGBoost. Experimental results show that WCFR consistently outperforms the competing methods, achieving higher validation accuracy (up to approximately 0.84) and Macro-F1 scores (up to approximately 0.55), while using fewer features and maintaining model interpretability. These results indicate that WCFR is particularly effective for accident severity modeling and highlight its potential as a robust feature selection strategy for large-scale transportation safety analytics and future severity prediction studies. Full article

Spontaneous combustion in goaf areas poses a significant threat to coal mine safety. Traditional safety management systems, reliant on passive response and single-indicator thresholds, often suffer from delayed warnings and lack cognitive decision support. To address this challenge, this study proposes a big-data-driven cognitive computing framework for dynamic risk prediction of goaf spontaneous combustion, based on a “Cloud-Edge-End” collaborative architecture. The method leverages multi-sensor big data streams (CO, C₂H₄, O₂, etc.) and deploys a lightweight Radial Basis Function (RBF) neural network on underground edge computing nodes (STM32) for real-time analytics. The model demonstrates excellent predictive performance on imbalanced datasets, with a PR-AUC of 0.910 and a recall of 99.7%. The edge-deployed RBF model achieves a single-pass inference time of only 0.62 ms, enabling real-time cognitive risk mapping. Field application at Z Coal Mine validated the system’s effectiveness, providing an average pre-warning time of 48.5 h, achieving zero spontaneous combustion accidents, and reducing the Total Recordable Injury Rate (TRIR) by 15.2%. This work illustrates how edge-based cognitive computing can transform safety management from passive response to proactive prevention, offering a scalable and interpretable framework for intelligent mine safety. Full article

The effect of alcohol on health is a controversial topic when it comes to the moderate or conscious consumption of fermented beverages. The recent claim by the World Health Organisation (WHO) and the European Heart Network (EHN) that the safe level of alcohol consumption is zero has compromised the efforts of the fermentation scientific community in developing healthier and more sustainable beverages. Therefore, the objective of this review was to assess the scientific background for such a claim that appears to be the result of recent scientific evidence. Using the meta-analytic data supporting WHO and EHN guidelines, it was possible to demonstrate that fermented beverages (e.g., wine and beer) have lower effects compared to spirits, that some population ethnicities have higher sensitivity to alcohol, and that drinking patterns influence the outcomes. Moreover, higher relative risks associated with younger individuals are mostly related to injuries (e.g., car accidents, self-inflicted injuries) and not with diseases. Sequential WHO studies produced significantly higher limits and emphasized that preventive policies should be tailored to populations at higher risk. In conclusion, the statement that “all alcohol is hazardous” has no scientific background and should be understood under the perspective that “one drink is too many and one thousand is never enough” used in alcoholism prevention. Fermentation researchers should continue their efforts on the promotion of healthier lifestyles, sustainable development and on the preservation of cultural heritage under the responsible drinking perspective. Full article

Road traffic accidents remain a major public health concern, and effective safety management requires interpretable tools that integrate multiple causal dimensions. This study proposes a Traffic Accident Causation Index (TACI) to provide a holistic representation of severity-related drivers by combining six theoretically grounded domains: Accident Infrastructure, Driver, Pedestrian, Road Condition, Emergency and Response, and Severity. Using a national police-reported dataset from Türkiye (N = 13,639), operational variables are mapped to normalized risk scores, aggregated into domain indices, and combined into a 0–100 composite TACI score. To assess the robustness and compatibility of the proposed index framework, we develop ensemble machine learning models (Random Forest, Gradient Boosting, LightGBM, XGBoost, and CatBoost) under two feature configurations: an Extended Feature Set (EFS) with the original variables and a Core Feature Set (CFS) consisting of the six domain indices. The results indicate that domain-level aggregation improves predictive stability, and the best-performing boosting models (XGBoost/CatBoost) achieve near-perfect agreement with the constructed index (test R² > 0.99) and very high classification performance (AUC > 0.999). SHAP-based explainability highlights pedestrian exposure and vulnerability as the dominant contributors, followed by lighting/visibility conditions, road surface quality, and adverse road–environment factors, whereas emergency-response and infrastructural attributes show comparatively indirect effects. Overall, the proposed framework supports interpretable, domain-oriented evidence for prioritizing safety interventions and monitoring high-risk accident conditions. Full article

Dynamic positioning (DP) systems are essential to maritime operations, as they ensure precise station keeping. Yet human error remains a major contributor to DP incidents, often interacting with technical failures and environmental conditions. This study proposes an adaptive probabilistic framework to characterise human-error contributions to DP risk and support targeted mitigation. We compare integrated Bayesian network (BN)/fuzzy analytic hierarchy process (AHP) and Bayesian network (BN)/Dempster–Shafer (D-S) theory to model causal relationships, aggregate uncertain expert judgements, and prioritise risk factors. Historical incident narratives, accident reports, and expert elicitation inform the model to analyse failure propagation and quantify factor contributions. In a representative DP case application, insufficient training, operator fatigue, and reduced situational awareness—together with software anomalies and adverse environmental loads—emerge as dominant contributors; BN backward analysis corroborates their diagnostic relevance. The approach yields actionable insights for risk reduction, including tailored training programmes, strengthened safety protocols, and integration of real-time monitoring. It provides an auditable, updateable basis for scenario-based training, software/maintenance assurance, and environment-aware operating envelopes, and is readily extendable as new evidence becomes available. Overall, the framework offers practical value for improving safety, operational continuity, and system resilience in DP operations. Full article

Safety monitoring of scaffolding operations is essential for preventing accidents in high-altitude construction. This study proposes an integrated computer vision and multi-criterion decision-making (MCDM) framework that combines object detection, pose estimation, Analytic Network Process (ANP) and ELECTRE III methods to evaluate safety risks of construction workers. Specifically, computer vision techniques are employed to extract objective visual evidence related to workers’ behaviors, protective equipment (PPE) usage, and working environments, which serve as the basis for subsequent safety risk quantification. A four-criterion system, including action risk, PPE compliance, working height, and structural integrity, is established. Weights are determined via the ANP, and risk ranking is conducted using ELECTRE III. Experiments on a self-built dataset achieved an mAP@0.5 of 92.3%, a segmentation IoU of 67.2%, and a pose OKS@0.5 of 89.6%. The evaluation results correlate strongly with expert assessments (Kendall’s τ = 0.79). The proposed framework effectively identifies unsafe behaviors and quantifies safety risks, providing reliable decision support for intelligent construction safety management. Full article

Background: Both home accidents and sleep problems are prevalent health issues among young children. This study aimed to investigate the association between the sleep characteristics of both preschool children and their mothers and the occurrence of home accidents among children. Methods: In this analytical cross-sectional study, the home accident group consisted of 90 children who presented to the Mersin University Hospital Pediatric Emergency Department due to home accidents. The control group comprised 90 healthy children, matched for age and sex with the home accident group. Sleep patterns of both children aged 12–72 months and their mothers, as primary caregivers, were evaluated through face-to-face interviews with the mothers. Results: Each one-hour increase in the child’s total nocturnal sleep duration increased the risk of being in the home accident group by 1.63 times (95% CI: 1.19–2.21, p = 0.002). Conversely, each one-hour increase in the mother’s total nocturnal sleep duration reduced the risk of child home accidents by a factor of 0.72 (95% CI: 0.58–0.91, p = 0.006). Maternal excessive daytime sleepiness increased the risk of home accidents in children by 11.35 times (95% CI: 2.38–54.26, p = 0.002). Conclusions: Preschool children who have had home accidents and their mothers should be evaluated for sleep problems. To reduce the frequency and severity of injuries associated with home accidents, greater focus must be placed on improving the sleep hygiene of both children and their mothers. Full article

Unsafe behavior remains a dominant contributor to accidents in complex socio-technical systems (STSs), yet it is still frequently interpreted as an individual-level information failure. This study argues that unsafe behavior is more accurately understood as a systemic outcome shaped by multi-level technological, organizational, and environmental conditions. To address this gap, an integrated human factor risk analysis framework is proposed by combining the STS perspective with safety information cognition (SIC) theory. The framework conceptualizes unsafe behavior as the result of risk transmission through safety information flows, linking system-level risk sources to individual perception, cognition, decision-making, and action. Within this perspective, human factor risk does not arise directly from individual error, but from deficiencies and asymmetries in the generation, transmission, and utilization of safety-related information embedded in the STS. Based on this conceptualization, a system-oriented human factor risk analysis (HRFA) approach is developed to support the identification, assessment, and control of unsafe behaviors across both accident scenarios and operational contexts. The framework is applied to road transportation of dangerous goods in China, a typical high-risk STS. The application results demonstrate that the proposed approach can effectively distinguish the comprehensive risk characteristics of different unsafe behaviors and reveal their underlying systemic causes. This study contributes to systems thinking in safety governance by shifting the analytical focus from individual behavior correction to upstream system conditions and information processes. The proposed framework provides a transferable approach for understanding and managing human factor risk in complex STSs and offers practical implications for proactive, system-oriented safety governance. Full article

Lagrangian particle-tracking models are increasingly used to simulate radionuclide transport in marine environments, especially for assessing the consequences of accidental releases. However, existing models generally neglect radioactive decay chains, limiting their ability to reproduce the complete behavior of radionuclides and their progeny. To the authors’ knowledge, this work presents the first implementation of radioactive decay chains within a fully three-dimensional Lagrangian marine radionuclide transport model, explicitly coupling stochastic particle tracking with decay kinetics and dynamic sediment–water interactions, enabling a realistic simulation of parent–daughter transformations in the ocean. The approach is tested for the chain in the Western Mediterranean Sea, following a hypothetical nuclear accident. Results confirm that the stochastic treatment accurately reproduces analytical decay solutions and can be seamlessly incorporated into operational-scale transport simulations. The framework can be extended to other radionuclide series and marine domains, providing a versatile and computationally efficient tool for emergency response, environmental impact assessment, and safety analysis in nuclear engineering applications. Full article

This research focuses on the challenge of measuring the socio-economic impact of road traffic accidents (RTAs) by examining how losses are redistributed across major institutional sectors, including the government, businesses, and households. Unlike traditional cost-based approaches, the analysis relies on a modified input–output framework that captures not only the direct losses but also the indirect damage flows transmitted from one sector to another. This methodology makes it possible to reveal the multiplicative propagation of losses, determine the proportion of net costs, and quantify the transfer dependencies between institutional agents. Using compiled and adapted data for the Azerbaijani economy, the study estimates the net economic damage from RTAs at 2268.17 million manats after adjusting for internal transfers. The results show that households bear more than 47% of total losses, the enterprise sector accounts for approximately 39%, and the government absorbs nearly 13%. The model also isolates an “additional damage” component, reflecting lost income, profits, and tax revenues, and demonstrates that every 1000 RTA generates a chain reaction of interlinked costs that substantially amplifies the overall effect. The findings highlight the necessity of integrating input–output analytical approaches into the practical assessment of RTA-related economic consequences, particularly in countries with limited statistical capacity and structurally diverse institutional linkages. Full article

In recent years, fall-from-height accidents have frequently occurred in construction activities, posing severe risks to workers’ safety and impeding the sustainable development of construction enterprises as well as social stability. Due to the complexity and multifactorial nature of such accidents, traditional safety risk assessment methods face significant limitations in uncovering their underlying causes. To address this issue, this study develops a novel analytical framework that integrates text mining with an improved Apriori algorithm. A standardized text preprocessing pipeline is established, including data collection, construction of a domain-specific lexicon, and synonym-based term unification. Key features are extracted using the TF-IDF method, while thematic patterns are identified through LDA topic modeling. To overcome the contextual insensitivity of conventional association rule mining, the Apriori algorithm is enhanced by introducing time-based constraints, enabling the discovery of accident causation patterns that differ between daytime and nighttime. Using 1064 accident reports from 22 provinces in China, the framework extracted 40 high-frequency accident-causing features and generated a richer set of meaningful association rules compared to the standard algorithm. The results indicate that insufficient safety protection, inadequate worker training, and management deficiencies are the predominant causes of fall-from-height accidents. Building on these insights, the study proposes targeted preventive measures. The findings make significant theoretical contributions by enhancing methodological frameworks for accident analysis while also providing practical insights to improve safety management practices in the construction industry. Full article

To improve the fire safety performance of fire protection renovation projects for existing public buildings, this paper systematically sorts out and analyzes relevant research studies, accident reports, and fire protection renovation codes and guidelines. It constructs a fire performance evaluation system for such projects, including 4 first-level indicators—”Building Characteristics”, “Building Fire Protection and Rescue”, “Fire Facilities and Equipment”, and “Heating, Ventilation, Air Conditioning (HVAC) and Electrical Systems”—and 19 second-level indicators such as “Building Usage Function”. The subjective–objective combined weighting method of Analytic Hierarchy Process (AHP)-CRITIC is adopted to determine the weights of indicators at all levels. Four high-weight second-level indicators are selected as core remediation objects: average fire load density, floor layout, automatic fire alarm and linkage control system, and electrical systems. Meanwhile, the evaluation system is converted into a Bayesian Network model, with an empirical verification analysis carried out on a shopping mall in Chaoyang District, Beijing, as a case study. Results show that the approach of combining partial codes with the rectification of high-weight indicators can reduce the fire occurrence probability of the mall from 78%, before renovation, to 24%. Therefore, the constructed evaluation system and Bayesian Network model can realize the accurate quantification of fire risks, provide scientific and feasible technical schemes for the fire protection renovation of existing public buildings, and lay a foundation for enriching and improving fire protection assessment theories. Full article