This paper examines how maritime accident investigation reports can be used more effectively to strengthen safety management and environmental protection. Building on earlier work focused on audit and Port State Control findings, it analyses accident reports to identify failures in operational control, Safety Management Systems, and organizational decision-making. The study finds that learning from accidents is limited by inconsistent investigation quality, weak evidential support, and insufficient causal explanation, reducing the effectiveness of corrective actions in preventing recurrence. A further contribution is the paper’s structured approach to distinguishing company-related failures embedded in quality assurance and management control from crew-related failures arising during operations. This distinction is used to support clearer accountability and more targeted prevention strategies, particularly in relation to management faults and manning issues. The study is based on a review of approximately 200 accident investigation reports since 2010, from which around 60 accidents were selected for detailed analysis. These cases span a range of vessel types and accident modes, including collisions, groundings, capsizing, falls from height, fire incidents, and enclosed-space events involving toxic exposure and fatalities. Each accident is summarized using a consistent set of descriptors covering vessel characteristics, accident type, causal statements, casualties, recommendations, recurrence likelihood, and the ISM element considered contributory. To impose structure on inconsistent reporting formats, the paper adopts an error-versus-mistake analytical logic supported by behavioral classification tools. A flowchart-based approach is used to support decisions on whether dominant causal pathways point to deficiencies in quality assurance and management control or to failures in operational execution and human performance. A central finding is that many accident investigation reports fail to support effective organizational learning. Despite prevention being their stated purpose, many lack sufficient evidential depth and causal clarity to reduce recurrence. Common weaknesses include poor supporting evidence, vague attribution of causality, and inconsistent identification of responsible organisations and decision-makers. Missing or unclear contextual information such as administrative responsibility, ownership, and the interface between company governance and shipboard management undermines accountability and obscures whether failures were systemic, operational, or both. These shortcomings align with concerns previously raised in analyses submitted to the IMO. The paper also reports difficulty in relying on global reporting systems such as GISIS for consistent learning, largely due to variability in the completeness and comparability of underlying reports. As a result, emphasis is placed on direct report review and triangulation with external analyses submitted through IMO processes. 1 To improve analytical clarity, the paper distinguishes between company-related errors and crew-related mistakes. Company-related errors refer to systemic deficiencies within quality assurance and management systems, including weaknesses in policies, procedures, risk assessment, supervision, audit processes, and resourcing. Crew-related mistakes are defined as failures in operational execution, including non-compliance with procedures, unsafe actions, communication breakdowns, poor situational awareness, and inappropriate decisions at the point of work. The paper stresses that this distinction is not intended to assign blame. Rather, it is used to identify where effective control should have existed and how organizational conditions may have shaped frontline performance. This paper presents a structured review of accident investigation reports to identify recurring deficiencies, weaknesses in reporting quality, and dominant human and organizational contributors to accidents. It finds that many reports lack the evidential depth and causal clarity required to prevent recurrence, with learning further degraded by vague recommendations and incomplete contextual information. Across the reviewed cases, ineffective communication, flawed decision-making, and unmanaged human vulnerability emerge as the most common contributors. The paper argues that meaningful improvement depends on rigorous learning systems, structured causal analysis, and stronger organizational accountability, supported by competence development and guidance that links lessons learned directly to Safety Management System control improvements.
Ziarati et al. (Mon,) studied this question.