Unsafe behaviors have long been recognized as a contributing factor to construction accidents. In practice, the prevention of unsafe behavior-related risks mainly relies on pre-construction safety training intended to enhance safety awareness, together with manual inspection and intervention by supervisors during construction operations. However, the effectiveness of safety training typically diminishes over time, leading to a gradual decline in safety awareness, while manual supervision is labor-intensive and prone to omissions. These limitations make it difficult for conventional static safety assessment approaches to timely reflect evolving safety conditions, thereby constraining accurate identification of unsafe behavior-related risk states. To address this challenge, this study proposes a dynamic monitoring and assessment framework for identifying unsafe behavior-related risk states in construction. The framework combines expert-based subjective assessment, text mining-based objective analysis, and computer vision-based unsafe behavior detection to support continuous monitoring and assessment of evolving on-site risk states. In the case study, a combined subjective-objective weighting scheme was developed based on 36 expert surveys and 111 historical accident cases, and a computer vision-based real-time monitoring module was implemented for on-site unsafe behavior detection. The proposed framework was further validated using 7 days of synchronized on-site surveillance video streamed through a digital twin platform associated with a real-world bridge construction project. The results demonstrate the capability of the framework to dynamically identify variations in unsafe behavior-related risk states and to provide timely feedback reflecting evolving on-site safety conditions. Overall, the framework holds promise as a valuable tool for supporting on-site safety management and preventing accidents.
The increasing frequency and severity of natural and technological disasters underscore the urgent need for effective disaster management and emergency response strategies, particularly in Small Island Developing States (SIDS) such as the Maldives. This study examines the role of Information Systems (IS) in enhancing disaster preparedness, response, and resilience within the utility sector, focusing on electricity and water services in the Greater Malé Region. These sectors are critical for ensuring continuity of essential services during crises, and their disruption can severely hinder immediate response and long-term recovery. Using a qualitative approach, data was collected through semi-structured interviews with senior stakeholders from key utilities. The findings highlight systemic challenges, including fragmented coordination among agencies, limited integration of disaster data, and inadequate utilization of real-time information for decision-making. The study demonstrates that integrated IS can significantly strengthen situational awareness, streamline inter-agency communication, and support more efficient emergency operations. The adoption of digital platforms for data sharing and monitoring enhances both organisational preparedness and public service continuity. This research contributes to knowledge on disaster risk reduction by showing how IS can be leveraged to build resilience in critical utility infrastructure. As the first empirical study of IS integration in Maldivian utility disaster management, this research provides practical recommendations for policymakers and utility providers, with transferable insights for other SIDS facing similar vulnerabilities.
Background. Reliable performance in sociotechnical systems depends on continuous human adaptation. Resilience engineering and Safety-II established adaptation as the mechanism through which systems remain safe, yet governance monitors its products — accidents, mortality, delays, and compliance — far more readily than the compensatory work producing them. Stable performance may therefore conceal progressive deterioration in underlying system condition.,Model. This paper develops a formal dynamic model linking structural gaps G(t), compensatory effort C(t), adaptive reserves R(t), and observable output O(t). The model distinguishes demand-driven deficits from design-generated burdens that recruit the same finite reserves but require different interventions. The resulting compensation-dependent system state is termed Persistent Compensatory Engagement. The model predicts that compensatory effort rises before reserves decline, reserve depletion precedes outcome deterioration, and the sustainability boundary contracts as reserves are consumed.,Evidence. Convergent evidence from healthcare and commercial aviation reconstructs this predicted sequence across independent research programs. Both sectors document widening structural gaps, escalating compensatory effort, and reserve depletion while outcomes remained stable, with deterioration only now beginning to emerge. A three-tier measurement architecture links each model variable to candidate indicators for prospective monitoring.,Implications. In compensation-dependent systems, stable performance is not evidence of safety, it is evidence that compensatory work is succeeding. PCE reorients governance from monitoring the products of adaptation to measuring the processes sustaining them, and from reacting to failure after it emerges to intervening during the interval when structural correction remains possible. It provides a formal, falsifiable basis for distinguishing structural resilience from compensation-dependent stability in high-consequence systems.
