An interpretable clustering approach to safety climate analysis: Examining driver group distinctions.

The transportation industry, particularly the trucking sector, is prone to workplace accidents and fatalities. Accidents involving large trucks accounted for a considerable percentage of overall traffic fatalities. Recognizing the crucial role of safety climate in accident prevention, researchers have sought to understand its factors and measure its impact within organizations. While existing data-driven safety climate studies have made remarkable progress, clustering employees based on their safety climate perception is innovative and has not been extensively utilized in research. Identifying clusters of drivers based on their safety climate perception allows the organization to profile its workforce and devise more impactful interventions. The lack of utilizing the clustering approach could be due to difficulties interpreting or explaining the factors influencing employees' cluster membership. Moreover, existing safety-related studies did not compare multiple clustering algorithms, resulting in potential bias. To address these problems, this study introduces an interpretable clustering approach for safety climate analysis. This study compares five algorithms for clustering truck drivers based on their safety climate perceptions. It also proposes a novel method for quantitatively evaluating partial dependence plots (QPDP). Then, to better interpret the clustering results, this study introduces different interpretable machine learning measures (Shapley additive explanations, permutation feature importance, and QPDP). The Python code used in this study is available at https://github.com/NUS-DBE/truck-driver-safety-climate. This study explains the clusters based on the importance of different safety climate factors. Drawing on data collected from more than 7,000 American truck drivers, this study significantly contributes to the scientific literature. It highlights the critical role of supervisory care promotion in distinguishing various driver groups. Moreover, it showcases the advantages of employing machine learning techniques, such as cluster analysis, to enrich the scientific knowledge in this field. Future studies could involve experimental methods to assess strategies for enhancing supervisory care promotion, as well as integrating deep learning clustering techniques with safety climate evaluation.

The impact of COVID-19 pandemic on construction safety in China and the U.S.: A comparative study.

Many studies indicated that the COVID-19 pandemic had significant impacts on construction safety. However, the studies had differing views on whether the pandemic increased or decreased construction safety performance. Furthermore, past studies did not adopt a comparative time series approach to evaluate the impact of the pandemic on construction safety. Thus, this study explores the differences in the impact of the COVID-19 pandemic on construction safety in China and the United States. This study used SciNet to forecast the number of construction accidents without the pandemic. Subsequently, the forecast was compared with the actual number of accidents since the outbreak, and the analysis showed a reduced number of construction accidents during the pandemic. However, there were minimal changes and even a slight worsening of fatality rates. Moreover, the correlation analyses showed that the effect of the pandemic on construction safety was weak and lagging. Construction safety was significantly affected by the pandemic in China, and the impact is relatively rapid. In comparison, outbreaks did not have a major impact on construction safety in the U.S. in the early stage. Since the pandemic is still raging worldwide, this research helps governments or project stakeholders formulate more targeted and data-driven safety countermeasures to improve construction safety during the crisis. The study also helps nations respond to future pandemics and crises to prevent adverse effects on construction safety.

Prioritization of industry level interventions to improve implementation of design for safety regulations.

INTRODUCTION: Many countries introduced mandatory Design for Safety (DfS) or Prevention through Design (PtD) requirements to reduce construction accident rates. However, there is a knowledge gap on the relative importance of industry level interventions to improve the implementation of DfS regulations. Thus, this study aims to identify and prioritize a set of industry level interventions to help regulators and industry associations understand the industry's perceptions and improve the implementation of mandatory DfS. METHOD: A mixed method approach consisting of 59 semi-structured interviews, four focus group discussions, and an online poll was implemented. RESULTS: Key challenges faced during DfS implementations were identified (lack of guidelines, lack of commitment towards DfS, the inadequate capability of DfS team, and limited effectiveness of DfS Professionals (DfSPs)). The study elicited eight industry level interventions to overcome these challenges and ranked them based on effectiveness and ease of implementation. The ranked industry level interventions in descending order are continuing training for DfSPs, samples and guidelines, DfS training for non-DfSPs, Building Information Modelling (BIM) for DfS review, strengthening DfSP as a profession, DfS awards for developers, third party audits for DfS reviews, and submission of DfS Risk Register to regulator. CONCLUSIONS: Identified interventions were classified into four intervention categories: (a) improving competency/ knowledge; (b) technological tools; (c) checks/ audits; and (d) recognitions/ certifications. The key contributions of this study are the identification and prioritization of industry level interventions for DfS, and the classification of safety interventions available to industry associations and regulators. PRACTICAL APPLICATIONS: Findings from this study help regulators and industry associations prioritize their resources to improve the implementation of mandatory DfS. Moreover, regulators and industry associations can also use the generic framework of industry level interventions to identify possible interventions to improve other mandatory WSH processes.

Management of safe distancing on construction sites during COVID-19: A smart real-time monitoring system.

The outbreak of Coronavirus Disease 2019 (COVID-19) poses a great threat to the world. One mandatory and efficient measure to prevent the spread of COVID-19 on construction sites is to ensure safe distancing during workers' daily activities. However, manual monitoring of safe distancing during construction activities can be toilsome and inconsistent. This study proposes a computer vision-based smart monitoring system to automatically detect worker breaching safe distancing rules. Our proposed system consists of three main modules: (1) worker detection module using CenterNet; (2) proximity determination module using Homography; and (3) warning alert and data collection module. To evaluate the system, it was implemented in a construction site as a case study. This study has two key contributions: (1) it is demonstrated that monitoring of safe distancing can be automated using our approach; and (2) CenterNet, an anchorless detection model, outperforms current state-of-the-art approaches in the real-time detection of workers.

Experimental validation of an energy balance approach for design of horizontal lifeline systems.

The energy balance approach is one of the design approaches approved in fall protection standards Z359.6, Z259.16 and SS 607 to ensure that horizontal lifeline systems (HLLSs) are adequately designed. However, this study found that theoretical calculations predicting the total fall distance (hTFD) and maximum arrest load (MAL) using an energy balance approach need to be corrected before they can be used safely. Based on the data from 48 drop tests, the authors determined that energy balance calculations differ significantly from the empirical hTFD and MAL values of HLLSs. As a result, further correction factors are introduced into the theoretical calculations to estimate hTFD and MAL conservatively. These correction factors are estimated from a regression equation derived based on experimental results and theoretical calculations.

A mixed-methods approach to examining safety climate among truck drivers.

The purpose of the current study was to use a mixed-methods approach to understanding safety climate and the strategies to improve safety climate among truck drivers. Using both survey (N = 7246) and interview (N = 18) responses provided by truck drivers regarding key safety climate items, the current study identified a number of positive and negative policies, procedures and practices that truck drivers perceived as the determinants of whether their organizations are committed to the promotion of safety at work. Item response theory (IRT) analyses were conducted to identify discrimination parameters indicating which safety climate items were most sensitive to the safety climate level. Discriminative items were identified at both the organization and group levels which can be used to evaluate safety climate and differentiate a high versus low safety climate across groups and organizations in the trucking industry. Based on our results, we also offer safety researchers and practitioners some recommendations on what and/or how to intervene with and promote organizational safety climate in the trucking industry.

Validation of the energy balance approach for design of vertical lifeline systems.

To ensure that vertical lifeline systems (VLLSs) are well designed, calculation methods are required to estimate the extension of a personal energy absorber (PEA) (xPEA) and the total fall distance (hTFD). Thus, the authors conducted 28 tests to validate the accuracy of the energy balance approach for estimating xPEA and hTFD of VLLSs and propose suitable correction factors to improve the accuracy and safety of the estimated xPEA and hTFD. For 9 out of 19 tests with a PEA, the difference between the theorical xPEA and empirical xPEA was 25% or higher, indicating that the energy balance approach is not accurate for estimation of xPEA. In contrast, theoretical values of hTFD are more accurate. Linear regression equations for estimating xPEA (R2 = 0.81) and hTFD (R2 = 0.99) were developed. The regression equations can be used to improve the accuracy and conservativeness of estimations of xPEA and hTFD during the design of VLLSs.

Factors influencing unsafe behaviors: A supervised learning approach.

Despite its potential, the use of machine learning in safety studies had been limited. Considering machine learning's advantage in predictive accuracy, this study used a supervised learning approach to evaluate the relative importance of different cognitive factors within the Theory of Reasoned Action (TRA) in influencing safety behavior. Data were collected from 80 workers in a tunnel construction project using a TRA-based questionnaire. At the same time, behavior-based safety (BBS) observation data, % unsafe behavior, was collected. Subsequently, with the TRA cognitive factors as the input attributes, six widely-used machine learning algorithms and logistic regression were used to develop models to predict % unsafe behavior. The receiver operating characteristic (ROC) curves show that decision tree provides the best prediction. It was found that intention and social norms have the biggest influence on whether a worker was observed to work safely or not. Thus, managers aiming to improve safety behaviors need to pay specific attention to social norms in the worksite. The study also showed that a TRA survey can be used to extend a BBS to facilitate more effective interventions. Lastly, the study showed that machine learning algorithms provide an alternative approach for analyzing the relationship between the cognitive factors and behavioral data.

Construction accident narrative classification: An evaluation of text mining techniques.

Learning from past accidents is fundamental to accident prevention. Thus, accident and near miss reporting are encouraged by organizations and regulators. However, for organizations managing large safety databases, the time taken to accurately classify accident and near miss narratives will be very significant. This study aims to evaluate the utility of various text mining classification techniques in classifying 1000 publicly available construction accident narratives obtained from the US OSHA website. The study evaluated six machine learning algorithms, including support vector machine (SVM), linear regression (LR), random forest (RF), k-nearest neighbor (KNN), decision tree (DT) and Naive Bayes (NB), and found that SVM produced the best performance in classifying the test set of 251 cases. Further experimentation with tokenization of the processed text and non-linear SVM were also conducted. In addition, a grid search was conducted on the hyperparameters of the SVM models. It was found that the best performing classifiers were linear SVM with unigram tokenization and radial basis function (RBF) SVM with uni-gram tokenization. In view of its relative simplicity, the linear SVM is recommended. Across the 11 labels of accident causes or types, the precision of the linear SVM ranged from 0.5 to 1, recall ranged from 0.36 to 0.9 and F1 score was between 0.45 and 0.92. The reasons for misclassification were discussed and suggestions on ways to improve the performance were provided.

Knowledge, attitude and practices for design for safety: A study on civil & structural engineers.

Design for safety (DfS) (also known as prevention through design, safe design and Construction (Design and Management)) promotes early consideration of safety and health hazards during the design phase of a construction project. With early intervention, hazards can be more effectively eliminated or controlled leading to safer worksites and construction processes. DfS is practiced in many countries, including Australia, the UK, and Singapore. In Singapore, the Manpower Ministry enacted the DfS Regulations in July 2015, which will be enforced from August 2016 onwards. Due to the critical role of civil and structural (C&S) engineers during design and construction, the DfS knowledge, attitude and practices (KAP) of C&S engineers have significant impact on the successful implementation of DfS. Thus, this study aims to explore the DfS KAP of C&S engineers so as to guide further research in measuring and improving DfS KAP of designers. During the study, it was found that there is a lack of KAP studies in construction management. Therefore, this study also aims to provide useful lessons for future applications of the KAP framework in construction management research. A questionnaire was developed to assess the DfS KAP of C&S engineers. The responses provided by 43 C&S engineers were analyzed. In addition, interviews with experienced construction professionals were carried out to further understand perceptions of DfS and related issues. The results suggest that C&S engineers are supportive of DfS, but the level of DfS knowledge and practices need to be improved. More DfS guidelines and training should be made available to the engineers. To ensure that DfS can be implemented successfully, there is a need to study the contractual arrangements between clients and designers and the effectiveness of different implementation approaches for the DfS process. The questionnaire and findings in this study provided the foundation for a baseline survey with larger sample size, which is currently being planned. In contrast to earlier studies, the study showed that the responding C&S engineers were supportive of the DfS. The study showed that the key to improving the DfS KAP of C&S engineers is by improving clients' motivation for DfS.

A hybrid simulation approach for integrating safety behavior into construction planning: An earthmoving case study.

One of the key challenges in improving construction safety and health is the management of safety behavior. From a system point of view, workers work unsafely due to system level issues such as poor safety culture, excessive production pressure, inadequate allocation of resources and time and lack of training. These systemic issues should be eradicated or minimized during planning. However, there is a lack of detailed planning tools to help managers assess the impact of their upstream decisions on worker safety behavior. Even though simulation had been used in construction planning, the review conducted in this study showed that construction safety management research had not been exploiting the potential of simulation techniques. Thus, a hybrid simulation framework is proposed to facilitate integration of safety management considerations into construction activity simulation. The hybrid framework consists of discrete event simulation (DES) as the core, but heterogeneous, interactive and intelligent (able to make decisions) agents replace traditional entities and resources. In addition, some of the cognitive processes and physiological aspects of agents are captured using system dynamics (SD) approach. The combination of DES, agent-based simulation (ABS) and SD allows a more "natural" representation of the complex dynamics in construction activities. The proposed hybrid framework was demonstrated using a hypothetical case study. In addition, due to the lack of application of factorial experiment approach in safety management simulation, the case study demonstrated sensitivity analysis and factorial experiment to guide future research.

Modelling and simulation of complex sociotechnical systems: envisioning and analysing work environments.

Accurate comprehension and analysis of complex sociotechnical systems is a daunting task. Empirically examining, or simply envisioning the structure and behaviour of such systems challenges traditional analytic and experimental approaches as well as our everyday cognitive capabilities. Computer-based models and simulations afford potentially useful means of accomplishing sociotechnical system design and analysis objectives. From a design perspective, they can provide a basis for a common mental model among stakeholders, thereby facilitating accurate comprehension of factors impacting system performance and potential effects of system modifications. From a research perspective, models and simulations afford the means to study aspects of sociotechnical system design and operation, including the potential impact of modifications to structural and dynamic system properties, in ways not feasible with traditional experimental approaches. This paper describes issues involved in the design and use of such models and simulations and describes a proposed path forward to their development and implementation. PRACTITIONER SUMMARY: The size and complexity of real-world sociotechnical systems can present significant barriers to their design, comprehension and empirical analysis. This article describes the potential advantages of computer-based models and simulations for understanding factors that impact sociotechnical system design and operation, particularly with respect to process and occupational safety.

Dynamics of safety performance and culture: a group model building approach.

The management of occupational health and safety (OHS) including safety culture interventions is comprised of complex problems that are often hard to scope and define. Due to the dynamic nature and complexity of OHS management, the concept of system dynamics (SD) is used to analyze accident prevention. In this paper, a system dynamics group model building (GMB) approach is used to create a causal loop diagram of the underlying factors influencing the OHS performance of a major drilling and mining contractor in Australia. While the organization has invested considerable resources into OHS their disabling injury frequency rate (DIFR) has not been decreasing. With this in mind, rich individualistic knowledge about the dynamics influencing the DIFR was acquired from experienced employees with operations, health and safety and training background using a GMB workshop. Findings derived from the workshop were used to develop a series of causal loop diagrams that includes a wide range of dynamics that can assist in better understanding the causal influences OHS performance. The causal loop diagram provides a tool for organizations to hypothesize the dynamics influencing effectiveness of OHS management, particularly the impact on DIFR. In addition the paper demonstrates that the SD GMB approach has significant potential in understanding and improving OHS management.