ABSTRACT
Background
Although occupational safety in the chemical industry is an area of ongoing research worldwide, chemical incidents continue to occur, indicating the need to review the causes, assist in safety management, and identify problems in the quality of reporting.
Methods
A systematic search of studies published in English between January 1, 2000 and August 31, 2025 was conducted using the electronic databases Medline and Web of Science Core Collection. Due to their heterogeneity, studies were evaluated in a scoping review framework, presenting quantitative data on causes of industrial chemical incidents.
Results
A total of 12 studies from the EU and 17 outside the EU were included. Both in the EU and outside, organizational factors were the most prominent cause of chemical accidents, and the most commonly mentioned specific causes were lack of clear work instructions such as lack of adequately written standard operating procedures (SOPs), deficiencies in risk analysis, deficiencies in planning and scheduling of procedures, especially during non‐routine operations, insufficient safety training, and poor design and engineering of chemical processes. Regarding human factors, operator errors were noted in all geographical locations; however, studies from Asia more often pointed out the contribution of inadequate supervision and intentional violations. Lastly, technical factors were overall not a leading cause of chemical accidents, but there was a somewhat higher percentage of accidents attributed to this category in non‐EU countries.
Conclusions
Together, these insights underscore the need for system‐level preventive strategies, supported by harmonized reporting and prospective studies, to strengthen evidence‐based prevention.
Keywords: causality, incident, occupational, overview, prevention
1. Introduction
Although occupational safety in the chemical industry is an area of ongoing study worldwide, with the most well‐known regulatory example being the EU Seveso III Directive [1], chemical incidents continue to occur. The National Institute for Public Health and the Environment, RIVM of the Netherlands, recently conducted a comprehensive study of occupational incidents investigated by the official inspectorate for the control of major hazards and found that 97% occurred at chemical production sites [2], while data from the US show that the average frequency of accidents in the chemical industry resulting in multiple fatalities and injuries has in recent decades been four per year [3].
Since there are numerous technical and human factors that can affect worker performance and facilitate an incident, several research methodologies have been developed to analyze the causes of accidents, such as MORT (Management and Oversight Risk Tree), MTO (Man Technology and Organization) analysis, the TRIPOD method based on the “Swiss Cheese” model, SOL (Safety through Organizational Learning), the Accimap tool, and TAMP (Systems‐Theoretic Accident Model and Processes) analysis, and others [4]. However, primary studies still often use self‐designed categorizations, which consequently make comparisons difficult, and although there are several reviews that attempt to summarize the available literature, this area remains an active field of research [5]. For example, the systematic review by Sadeghi et al. [6] only listed primary studies reporting specific reasons for chemical accidents, but no quantitative data were extracted, and some of the included studies had a broad period of analysis, for example, the majority of cases occurred before the 1980s. The systematic review by Dodoo and Al‐Samarraie [7] included industries other than the chemical, and again did not provide quantitative data on causes. The systematic review by Politis et al. [8] on occupational injuries in Greece across various industrial sectors only mentions that in the petrochemical industry, the majority of accidents were attributed to human error (46%), based on only two publications. In the literature review by Kyriakaki et al. [9], the causes of occupational eye injuries following chemical accidents were not analyzed separately from other causes such as mechanical damage.
Overall, we noted several problems that make a comprehensive review of chemical accidents difficult. Primary studies often group chemical accidents with other types of accidents that occur in the chemical industry but are not related to chemicals, such as falls, mechanical injuries, or electric shocks, among others. The aforementioned inclusion of broad time periods also limits the applicability of primary study results to the preventive measures relevant today. There is the additional challenge of data comparability due to the variability of the databases used as data sources in primary studies, such as the Chemical Safety Board (CSB) in the US, the EU eMARS database of the European Commission's Joint Research Centre (JRC), the Major Hazard Incident Data Service (MHIDAS) in the UK, and the Relational Chemical Accident Information System (RISCAD) database of the Japan Institute of Advanced Industrial Science and Technology. Finally, while there are many reports of explosions, fires, leaks and poisonings that unfortunately resulted in death and injury, there are rarely reports that provide statistical data on the factors leading to specific adverse health effects such as corrosive skin and eye injuries or respiratory effects [10, 11, 12, 13].
Therefore, to address this gap our aim was to conduct a scoping review of causes of chemical accidents worldwide, to assist in safety management and identify problems in the quality of reporting and analysis, with the ultimate goal of overcoming current limitations in improving occupational safety [4] through closer collaboration between science and industry.
2. Methods
This study was part of a project aimed to design recommendations for preventing occupational chemical accidents by addressing the most common causes, so this scoping review of recent scientific evidence was conducted as a first step in that process, using the Arksey and O'Malley framework [14] and the checklist of Preferred Reporting Items for Systematic Reviews – extension for Scoping Reviews (PRISMA‐ScR) [15]. Primary planned systematic review was not possible to conduct due to the heterogenicity of the outcome variables in the primary studies, which did not allow full systematization of conclusions. As this work was a review of published primary studies, no approval from an ethics review board was required.
2.1. Information Sources and Search
A systematic search of peer‐reviewed literature published in English between January 1, 2000 and August 31, 2025 was conducted using the electronic databases Medline and Web of Science Core Collection. The inclusion criteria were: studies providing results of a quantitative analysis of the causes of chemical accidents, based on at least 10 incidents, at least 70% of the accidents analyzed occurring in 1990 and later, or, if that information had not been reported, the analyzed time period being at least 70% in 1990 and later, and that the outcome studied was attributed to a chemical hazard, typically a fire, explosion or inhalational poisoning due to an unplanned chemical reaction or uncontrolled release, resulting in human injury, death, financial loss, or official measures such as evacuation. In addition, to allow comparison between studies, a study was not included if only complex statistical results were presented, such as regression analysis, data mining, Bayesian network analysis, and so on, but not the number and percentages of chemical accidents that contributed to each cause studied. Finally, we performed a manual search via Google Scholar to identify any potentially relevant articles that may have been missed in the first systematic search, and the references of the included publications were checked for possible additional literature. The detailed search syntax is shown in Table S1. The identification process is described in detail by the PRISMA flowchart in Figure 1.
Figure 1.

PRISMA flow diagram for the screening and selection of articles.
2.1.1. Screening of Articles
The retrieved articles were independently screened for inclusion by two authors (ŽB and EV) based on their abstracts, titles, and keywords. We used the web‐based literature screening tool Rayan [16]. Articles were approved by both authors based on the inclusion criteria, and in case of conflicts, the two authors discussed them with the third author (JM) until approval or disapproval. The full text of the included articles was retrieved for data extraction.
2.1.2. Data Extraction
Two authors (ŽB and EV) extracted the following data from the included articles: first author and year of publication, country of origin and study period, data source, studied adverse outcome, chemicals involved if reported, and results for causative factors. The reported causes were summarized into three categories: human factors (such as operator error, time pressure, fatigue etc.), technical factors (such as mechanical damage, corrosion, unexpected or uncontrolled reactions detection of anomalies etc.), and organizational factors (such as lack of work instructions, poor communication and planning etc.). The causes reported in the primary studies are summarized on Figure 2.
Figure 2.

Summarized causes of chemical accidents reported in primary studies. *Causes reported only outside EU.
2.1.3. Quality Assessment
Two authors (ŽB and EV) assessed the quality of included studies. The tool for the evaluation of study quality and risk of bias, with a maximal possible score of 15, was made by the authors using a mix of the methods research appraisal [17], Cochrane collaboration [18], and working group of the US EPA [19]. The criteria covered appropriate design, sampling and sample size, justification of methodology, and presentation of results, which are described in detail in Table S2. A score above 10 was considered as good quality. We developed this tool due to the heterogenicity of the primary studies covering various chemical incidents such as fire, explosion, or inhalation poisoning due to an unplanned chemical reaction or uncontrolled release, resulting in human injury, death, financial loss, or official measures such as evacuation. The existing standardized scoring systems e.g. JBI Critical Appraisal Checklists (https://jbi.global/critical-appraisal-tools) deal with various designs of epidemiological studies, economic studies or diagnostic test accuracy studies, and neither was completely fit for our purpose, so our tool contains questions formulated more flexibly to allow for the assessment of the quality of studies with heterogeneous definitions of chemical accidents while covering all important topics of quality as JBI Checklists.
3. Results
3.1. Studies in EU
Table S3 presents research studies in the European Union that report on the causal factors of chemical incidents at the workplace. Of the 12 studies in total from Norway, Sweden, Netherland, United Kingdom, France, Germany, Spain, Italy and Greece, which were predominantly of good quality with a low risk of bias, 10 were retrospective, in which data were extracted from various European databases of the chemical industry (such as MARS, ARIA, VARO etc.). The remaining two were prospective studies from the Netherlands in which Wijnands et al. [10, 11] collected data on the circumstances, causes, and exposures of occupational incidents reported to the Dutch Poison Information Centre during a 1‐year period. Furthermore, an interview was conducted with patients who gave their consent, followed by a detailed analysis of the main causes of occupational accidents.
Only three studies [10, 11, 13], from the Netherlands and Finland, focused on the health consequences of occupational accidents, while the other studies broadly defined chemical accidents, usually including explosions, fires, releases of hazardous substances, leaks, and so on. In the study of Wijnands et al. [11] health effects reported after accidents involving ingestion were pain in mouth or throat (37.9%), abdominal pain (20.7%), after ocular exposure pain (59.7%), redness (36.3%) or blurred vision (29.9%), and corneal damage (3.2%), after dermal contact pain (43.6%), redness (36.6%), burns (22.8%) and necrosis (5%), and after inhalation dyspnea (42.9%), coughing (33.3%) and headache (32.4%). In the other study by the same group of authors [10], focused on ocular injuries only, reported signs and symptoms were pain in the eyes (56.1%), eye irritation (55.0%), redness of the eyes (34.1%), temporary loss of vision (28.8%), and corneal abrasion (3.0%). The authors did not provide details on patients having multiple symptoms, as the percentages for specific exposures (e.g., ocular exposure) summed to well over 100%. In the third study focused on health effects [13], the studied outcome was occupational asthma, with following diagnostic criteria: exposure to a high concentration of an irritant, occurrence of asthma symptoms in a close temporal relationship to the exposure (usually < 24 h or after few days), reversible obstruction or non‐ specific airway hyper‐ responsiveness in methacholine or histamine challenge, persistence of symptoms > 3 months, no evidence of active asthma in adulthood before the exposure event and no other pulmonary disorder that would explain the symptoms. Overall, in the studies presented in Table S3, the chemicals most commonly reported were acids and alkalis, aldehydes and ketones, alcohols and glycols, hydrocarbons, natural gas and fuels, ammonia, nitrogen, hydrogen sulfide, chlorates and nitrates (salts). However, a quarter of studies, specifically studies from France, Sweden and Germany conducted by Dakkoune et al. [20], Jacobsson et al. [21], and Uth and Wiese [22], did not specify which chemicals were involved in the occupational accidents.
The reported causes from all 12 studies presented in Table S3, regardless of whether they were prospective or retrospective, can be summarized into three categories: human factors, technical factors, and organizational factors. Half of the studies analyzed all three categories of causes. Regarding most common factors, the results of studies from Netherlands, Sweden, Norway, Finland, UK and Italy presented in Table S3 showed that organizational factors are the main and most prevalent causes of occupational accidents, particularly a lack of clear work instructions, deficiencies in risk analysis, deficiencies in planning and scheduling of procedures, and inadequate failure diagnostics [10, 11, 13, 21, 23, 24, 25]. These are mostly followed by human factors, especially operator errors (typically due to fatigue, inaccuracy, time pressure etc.), neglect of work hygiene measures, and breaching barriers while performing maintenance, as reported in studies from UK, Netherlands, France, Norway, Finland and Italy [2, 11, 13, 20, 23, 24, 25]. Technical factors were the leading cause in the minority of studies, and the most common in this category were equipment/mechanical failures, as reported in studies from Greece, Germany and Spain [22, 26, 27].
Additionally, the studies conducted by Okoh et al. (Norway) and Nivolianitou et al. (Greece) analyzed combinations of causes in the workplace, which were out of the scope of this review. Specifically, Okoh et al. [23] identified that one‐third of all cause combinations related to the work process were due to deficient planning, scheduling, and fault diagnosis, while Nivolianitou et al. [27] found that a fifth of occupational accidents was caused by a combination of equipment failure and human error.
Lastly, there were several studies that did not meet any inclusion criteria but provided data of interest. Specifically, the study conducted by Yalcin et al. [28] analyzing data from the entire EU using the MARS and national databases was excluded due to its lack of a specified time period during which the occupational accident data was collected. However, it is important to note that occupational accidents in that study were predominantly caused by organizational deficiencies during the production process (64%), in storage (71%), during shutdown/startup operations (92%), as well as skill‐based errors in maintenance and repairs (82%) and also in loading/unloading (66%) [28]. Also, the study from Finland conducted by Kidam et al. [29] did not specify the time period during which the occupational accident data was collected, but the results were particularly interesting for the reasons for design errors (definition: “a design error is deemed to have occurred if the design or operating procedures are changed after an incident has occurred”). The most critical design errors were poor layout (17%), insufficient consideration of chemical reactivity and incompatibility (16%), and incorrectly chosen process conditions (16%).
Most of the included studies made specific recommendations arising from their analysis of the causes of accidents. These are summarized in Table 1 and given in detail in Table S4, with the prevailing advice being that simply providing personal protective equipment is not sufficient without a primary focus on training systems and appropriate supervision, detailed maintenance and cleaning procedures, and emergency planning systems.
Table 1.
Summarized recommendations of preventive measures discussed in primary studies based on the study findings.
| Recommendations of preventive measures |
|---|
| Establish continuous training for employees handling hazardous chemicals |
| Ensure clear, accessible, and up‐to‐date operating procedures |
| Promote a strong safety culture through reporting and proper PPE use |
| Implement robust management of change and maintenance procedures |
| Strengthen management oversight and accountability to ensure appropriate workload and effective chemical risk control |
| Regular physical and mental health assessments of workers to reduce error‐ and accident‐related risksa |
| Structured systems of incentives and sanctions to reinforce safe and deter unsafe behaviours a |
| Mandatory use of technical safety indicators and alarm systems as primary protective measures a |
Recommendations of preventive measures reported only in outside EU studies. Detailed recommendations of preventive measures discussed in primary studies were described in Table S4.
3.1.1. Studies Outside the EU
Seventeen studies met the inclusion criteria and are given in Table S5. The majority of studies were conducted in China (seven studies [30, 31, 32, 33, 34, 35, 36]) and the US (five studies [3, 37, 38, 39, 40]). In addition to China, three more studies were conducted in Asia, specifically: Iran [41], Sri Lanka [42], South Korea [43]. Interestingly, two studies [44, 45] investigated incidents globally, with Asia being the most represented in both studies: about a third of incidents in Maniar et al. [45] and half in Kannan et al. [44]. Regarding data sources, official national sources were predominately used, such as the Hazardous Substances Emergency Events Surveillance (HSEES) system, Chemical Safety Board (CSB) database, and Occupational Safety and Health Administration (OSHA) database in USA, as well as official reports of the Ministry of Emergency Management of the People's Republic of China (PRC) and China Chemical Safety Association (CCSA). A smaller number of studies used other types of sources, such as newspaper articles in a Sri Lankan study [42], publicly available reports without further specification in a South Korean study [43], and online news in a global study by Kannan et al. [44], while the other global study by Maniar et al. [45] used both official databases and unofficial reports.
Eleven studies reported which chemicals caused the incidents, and similarly to the EU studies, incidents occurred due to a wide range of chemical hazards (Table S5), such as acids and alkalis, aldehydes and ketones, alcohols and glycols, hydrocarbons, natural gas and fuels, ammonia, nitrogen, hydrogen sulfide, chlorates and nitrates. Only one study [39] focused exclusively on the health outcomes (fatal inhalation poisoning), while most studies that provided descriptive statistics on the types of accidents, usually explosions, fires, uncontrolled releases to the environment, or poisoning, reported fire and explosion as the leading incidents [31, 32, 33, 35, 36, 42, 44].
Similarly to our approach to the EU studies, the causes of accidents reported in the non‐EU studies were also summarized into three categories: human, technical, and organizational factors (Table S5). Around three‐quarters of the studies analyzed all three categories. The factors that generally predominated depended on the outcome studied, and the country. In particular, four well‐designed studies from the USA, specifically Anderson et al. [37] and Orr et al. [40] which focused on chemical releases recorded in the Health and Safety Executive (HSEES) system, and Baybutt et al. [38] and Fyffe et al. [3] which focused on chemical accidents of any type investigated by the US Chemical Safety Board, found that the causes were predominantly technical and organizational in nature, with the most prominent from the former category being equipment failure and deficiencies in protective measures, and from the latter the conduct of abnormal or non‐routine operations, lack of adequately written standard operating procedures (SOPs), and poor design and engineering of chemical processes. Technical factors (equipment or utility failures) were also the leading cause in a study from Iran [41] and Sri Lanka [42], both of which were at borderline acceptable quality. In contrast, in studies from China, which were of lower, but mostly still acceptable quality [30, 31, 32, 33, 34, 35], human and organizational factors were mentioned with highest percentages, especially inadequate supervision, insufficient safety culture, inadequate resource management, insufficient safety training specific to workplaces in chemical production, illegal operations and intentional violations. Furthermore, the study of acceptable quality from South Korea [43], which focused only on human factors, found that the most common in this category were lapse and rule‐based mistakes (misapplying good rules or applying bad rules, doing the wrong thing believing it is right). In addition, human factors such as overloading the vehicle or speeding, were almost exclusively the reason of accidents occurring during chemical transportation [34]. For the two studies that investigated chemical accidents globally [44, 45], organizational factors were again implicated as leading causes, specifically, deficiencies in hazard awareness and identification, such as adequate supervision of operations, providing process safety information to workers, recording incidents and near misses, and adequate risk management. Lastly, in a study from the US [39] which focused on fatal inhalational poisonings, as expected, human factors were the predominant cause, especially not adhering to confined space standards (i.e., permit‐required spaces).
In addition to the data presented in Table S5, some of the included studies investigated the complex interconnection of causes, which was out of the scope of this review. For example, Chen et al. [30], found that inappropriate planning of operations can make the occurrence probability of poor technological environment increase by nearly three times, which then in turn can double the possibility of decision errors. Vulnerabilities at an early stage of the organizational process (designing procedures, emergency plans, safety management system, supervision system, etc.) may increase the occurrence probability of failure in later stages of the accident cycle fivefold, by inadequately addressing problems. Furthermore, Jiang et al. [31] found that a poor organizational climate will increase the possibility of inadequate supervision by four times and the possibility of supervisory violations approximately eight times, and if a chemical enterprise suffers from inadequate supervision, this will increase the possibility of poor communication and coordination among workers approximately nine times, and the possibility of insufficient personal readiness at the time of the accident 14 times. Sadeghi Gelvardi et al. [41] found that providing safety data sheets (SDS), isolating chemical storage, training workers on chemical safety, and providing an emergency plan were associated with a significantly lower likelihood of workplace chemical accidents, while regular completion of a chemical accident prevention checklist by the employer and employer chemical training were not significantly associated. Finally, Anderson et al. [37] noted that the causes varied depending on the specific chemical that was released uncontrolledly leading to human injury or death, with equipment failure being the most important cause of releases of ammonia, carbon monoxide, and sulfuric acid, while human error was the most prominent contributor to releases of chlorine and hydrochloric acid.
Most of the included studies made specific recommendations arising from their analysis of the causes of accidents (Table 1 and Table S4), even designing a checklist of questions to be considered when evaluating updates to process safety management at a chemical facility.
4. Discussion
This scoping review highlighted that organizational causes are the most common leading causes of chemical accidents, both in the EU and outside it. Specific factors frequently mentioned were the lack of clear work instructions, such as the lack of adequately written standard operating procedures (SOPs), deficiencies in risk analysis, deficiencies in planning and scheduling procedures, especially during non‐routine operations, insufficient safety training, and poor design and engineering of chemical processes. In line with the specific recommendations for preventive measures discussed in primary studies (Table 1 and Table S4), frequent random inspections and updating of operating procedures, emphasizing proper on‐time evaluation and communication of process changes and non‐routine operations are recommended for the safety management systems [25, 38]. Safety procedures should include not only operation process but maintenance, cleaning, and proper use of personal protective equipment [10, 11]. Strong safety culture in chemical plants should involve repeated, continuously updated training program on safety procedures for workers handling hazardous chemicals, including temporary workers. It is beneficial to include case studies of previous accidents into the workers’ training, and to adopt various forms of teaching to gain workers’ interest (e.g., animations) [24, 25, 30, 41].
When analyzing accident characteristics globally, the first challenge is discrepancies in accident definitions. Typically, data are collected for major accidents, and the outcomes were usually mixed, including definitions such as “Accidental events in the chemical industry, ranging from catastrophic accidents to events that caused minor health effects and/or minor damages to property and the environment,” “Major accidents (toxic release, explosion and/or fire) involving chemical reactive substances” included in the MARS database. However, although the eMARS database for example represents a robust and well‐established system for reporting major industrial accidents under the Seveso III Directive, its scope is limited for health outcomes. In particular, minor and moderate incidents, which constitute the majority of workplace chemical exposures, are largely outside the remit of the system, and detailed clinical information on health effects is seldom available. In this context, data from poison control centers represent a valuable complementary source, as they capture real‐world chemical exposures, including less severe but frequent incidents and early health effects among workers, which are often not recorded in regulatory accident databases [11, 46]. Based on poison control center data [12], the majority of patients in occupational incidents with hazardous materials mainly develop only minor effects, such as dyspnea, irritation of mucous membranes, cough, headache, dizziness and nausea. In addition, it appears that for only about a quarter of the chemical accidents from the papers included in this review, information is available that they actually occurred in the workplace during work processes, while others vaguely defined the circumstances, or included improper storage, or transportation. This further highlights the need for coordinated medical data collection to strengthen surveillance, in addition to the need for near‐miss reporting, as it is estimated that for every 10,000 near‐misses, there will be 1000 minor injuries, 100 serious injuries, 10 deaths, and one catastrophic event [47].
Another challenge is the inconsistency in definitions of causes. Several papers have categorized human factors according to the Human Factors Analysis and Classification System (HFACS) originally developed for US Air Force accident investigation [48], but most have used their own classification, often reported too vaguely, for example, process design deficiencies, inadequate management, or human error. A further challenge is the fact that chemical accidents often have multiple causes [43]. In most articles included in this scoping review, statistics on causes were presented based on the number of accidents to which that cause contributed, but it would probably be more informative to provide data on how many accidents had a particular combination of causes.
All in all, the limitations of this scoping review stem from the limitations of the data collection and analysis in the primary studies. In an attempt to gain a better overview of the main problems, we grouped the causes reported in the primary studies into one of three provisional categories: organizational factors, human factors, or technical factors. Admittedly, some vaguely described causes could be considered both human and organizational problems, so the categorization was made to the best of our knowledge. In addition, there was a language barrier in the Asian studies, although the papers were in English. Consequently, in these cases we may have misunderstood the true nature of the cause, for example, when it was not clear whether personal protective equipment was not worn because it was not available at the workplace or was available but the worker neglected to use it. This again points to the need for clearer and standardized reporting. Therefore, future studies should take care to define the causes in more detail, and to avoid broad categories such as “equipment failure,” “human error,” and so on. To avoid overlapping categories, we recommend using established cause categorization systems in future primary studies, such as AcciMap [49] and HFACS (Human Factors Analysis and Classification System) [50]. Specifically, HFACS system employ these major categories: (1) Unsafe acts, (2) Preconditions for unsafe acts, (3) Unsafe supervision, and (4) Organizational influences [50]. On the other hand, Accimap organizes factors in progressive levels: (1) Government, policy & budgeting, (2) Regulatory bodies and associations, (3) Local area goverment, planing & budgeting, company managment, (4) Technical & operational management, (5) Physical processes & actor activities, (6) Equipment & surroundings [49]. Furthermore, a sub‐analysis of causes for each type of chemical accident would help characterize the risks better, instead of grouping for example, fire, explosion and poisonings together, as would the reporting of exact chemicals which caused these outcomes. Conversely, reporting the number of accidents with multiple causes (and which combinations exactly) would help meta‐analyze the quantitative data in future reviews to derive the weighted average for each cause, which was not possible in this review. Another limitation of our review is the relatively small number of included studies primarily due to the criteria of including only studies published from 2000 onward and using comparable reporting of results, but on the other hand we believe our results better reflect the situation in the chemical industry today. Due to the heterogeneity of the primary studies, it was difficult for us to assess the quality of the studies in a comparable manner, which is why we designed our own quality assessment tool. Although we have tried to be as transparent as possible, providing details of the design process in Methods and a complete quality checklist in Table S2, the problem of the tool being unvalidated still remains and could be the subject of future studies. As for other limitations of our work, we did not include studies that only presented complex statistical results, but not the number and percentage of chemical accidents contributing to each cause studied, it is possible that we missed important information about the complex interplay of causal factors. We recognize that this is a significant limitation, however even the simpler quantitative data presented in our review could not be summarized in a meta‐analysis due to the extremely heterogeneous definitions of accidents and causes. Further inclusion of studies with complex statistical analysis was not feasible. However, this could be the focus of a new review, which will build on the results of this one.
One of the strengths of our review is including studies conducted worldwide. The transfer of occupational health problems following the relocation of hazardous industries to developing countries is a well‐documented phenomenon [51], while increased risk of chemical exposure among immigrant workers remains another ongoing concern [52]. In Europe there is currently a rapid increase of digital transformation in occupational safety management systems such as e‐training tools, big data systems, tools for monitoring work environment, electronic workers’ health surveillance systems, e‐safety guidelines and legislation, facilitating effective decisions [53]. We plan to use the findings presented in this paper in preventive actions aimed at workers and their employers, and we hope that our results will be used for this purpose by other safety experts. Our findings that organizational factors most often contribute to chemical accidents strengthen conclusions from previous studies using complex statistical analysis, such as Bayesian network‐based risk analysis [54], indicating that organizational level is the main level determining the success of a preventive action.
In conclusion, this scoping review demonstrates that organizational and human factors, rather than isolated technical failures, are the predominant causes of occupational chemical accidents worldwide. Deficiencies in safety management systems—particularly in procedures, planning, supervision, and training—consistently underlie adverse events across regions. The findings also highlight the added value of poison control center data, which can capture less severe incidents and early health effects that are often not detected by traditional accident reporting systems and thereby support earlier identification of safety gaps. Together, these insights underscore the need for system‐level preventive strategies, supported by harmonized reporting and prospective studies, to strengthen evidence‐based prevention of occupational chemical accidents.
Author Contributions
Ž. Babić, M. Brvar and J. Macan conceived the concept and design of the paper. Ž. Babić, E. Veledar, M. Brvar and J. Macan participated in the collection and analysis of data, and all authors participated in the interpretation of data. Ž. Babić and E. Veledar wrote the draft of the manuscript, which was critically revised by all authors. All authors gave final approval for the version to be published and agreed to be responsible for all aspects of the paper to ensure that questions regarding the accuracy or integrity of any part of the paper are adequately investigated and resolved.
Ethics Statement
As this work was a review of published primary studies, no approval from an ethics review board was required.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting File 1
Supporting File 2
Supporting File 3
Supporting File 4
Supporting File 5
Acknowledgments
This study was a part of the scientific project entitled “Science in the service of preventing chemical accidents” (Grant sponsors: Ministry of Science and Education of the Republic of Croatia, grant number: 910‐06/24‐01/00230; 533‐10‐25‐0003, and Slovenian Research and Innovation Agency, grant number: BI‐HR/25‐27‐026).
Babić Ž., Veledar E., Franić Z., et al., “Accidents in the Chemical Industry: A Scoping Review Focused on Causal Factors,” American Journal of Industrial Medicine 69 (2026): 1007‐1016, 10.1002/ajim.70130.
Institution at which the work was performed: Poison Control Centre, Division of Occupational and Environmental Health, Institute for Medical Research and Occupational Health (Ksaverska cesta 2, 10000 Zagreb, Croatia), and Centre for Clinical Toxicology and Pharmacology, The Ljubljana University Medical Centre (Zaloška cesta 7, 1000 Ljubljana, Slovenia).
Data Availability Statement
The data that support the findings of this study are available in the supporting material of this article.
References
- 1.DIRECTIVE 2012/18/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 4 July 2012 on the Control of Major‐Accident Hazards Involving Dangerous Substances, Amending and Subsequently Repealing Council Directive 96/82/EC.
- 2. Kooi E. S., Manuel H. J., Mud M., and Bellamy L. J., Fifteen Years of Incident Analysis: Causes, Consequences, and Other Characteristics of Incidents With Hazardous Substances at Major Hazard Companies in the Period 2004‐2018. RIVM, (2020). 10.21945/RIVM-2020-0115. [DOI] [Google Scholar]
- 3. Fyffe L., Krahn S., Clarke J., et al., “A Preliminary Analysis of Key Issues in Chemical Industry Accident Reports,” Safety Science 82 (2016): 368–373, 10.1016/j.ssci.2015.10.008. [DOI] [Google Scholar]
- 4. Dien Y., Dechy N., and Guillaume E., “Accident Investigation: From Searching Direct Causes to Finding In‐Depth Causes – Problem of Analysis or/and of Analyst?,” Safety Science 50 (2012): 1398–1407, 10.1016/j.ssci.2011.12.010. [DOI] [Google Scholar]
- 5. Sterud T., Tynes T., Mehlum I. S., et al., “A Systematic Review of Working Conditions and Occupational Health Among Immigrants in Europe and Canada,” BMC Public Health 18, no. 1 (2018): 770, 10.1186/s12889-018-5703-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Sadeghi F., Khoshmanesh B., Fatemi F., et al., “Effective Parameters on Chemical Accidents Occurrence: Findings From a Systematic Review,” Journal of Failure Analysis and Prevention 20 (2020): 1184–1191, 10.1007/s11668-020-00923-9. [DOI] [Google Scholar]
- 7. Dodoo J. and Al‐Samarraie H., “A Systematic Review of Factors Leading to Occupational Injuries and Fatalities,” Journal of Public Health 31 (2021): 99–113, 10.1007/s10389-020-01427-4. [DOI] [Google Scholar]
- 8. Politis P., Leotsinidis M., Jelastopulu E., et al., “Occupational Injuries in Greece: A Systematic Review of the Literature,” Cureus 17, no. 1 (2025): e77661, 10.7759/cureus.77661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kyriakaki E. D., Symvoulakis E. K., Chlouverakis G., et al., “Causes, Occupational Risk and Socio‐Economic Determinants of Eye Injuries: A Literature Review,” Medicine and Pharmacy Reports 94, no. 2 (2021): 131–144, 10.15386/mpr-1761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Wijnands A. P. G., de Vries I., Carlier M. P., et al., “Causal Factors of Work‐Related Chemical Eye Injuries Reported to the Dutch Poisons Information Center,” Journal of Occupational and Environmental Hygiene 20, no. 8 (2023): 315–321, 10.1080/15459624.2023.2210183. [DOI] [PubMed] [Google Scholar]
- 11. Wijnands A. P. G., de Vries I., Verbruggen T., et al., “Acute Occupational Exposures Reported to the Dutch Poisons Information Center: A Prospective Study on the Root Causes of Incidents at the Workplace,” Journal of Occupational Medicine and Toxicology (London, England) 17, no. 1 (2022): 19, 10.1186/s12995-022-00360-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wijnands A. P. G., Koppen A., de Vries I., et al., “Poison Center Surveillance of Occupational Incidents With Hazardous Materials (2016‐2023): Insights for Risk Mitigation and Incident Preparedness,” International Journal of Environmental Research and Public Health 22, no. 2 (2025): 158, 10.3390/ijerph22020158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Lindström I., Lantto J., Karvala K., et al., “Occupations and Exposure Events in Acute and Subacute Irritant‐Induced Asthma,” Occupational and Environmental Medicine 78, no. 11 (2021): 793–800, 10.1136/oemed-2020-107323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Arksey H. and O'Malley L., “Scoping Studies: Towards a Methodological Framework,” International Journal of Social Research Methodology 8 (2005): 19–32, 10.1080/1364557032000119616. [DOI] [Google Scholar]
- 15. Page M. J., McKenzie J. E., Bossuyt P. M., et al., “The PRISMA 2020 Statement: An Updated Guideline for Reporting Systematic Reviews,” BMJ 372 (2021): n71, 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Ouzzani M., Hammady H., Fedorowicz Z., et al., “Rayyan‐A Web and Mobile App for Systematic Reviews,” Systematic Reviews 5, no. 1 (2016): 210, 10.1186/s13643-016-0384-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Pluye P., Gagnon M.‐P., Griffiths F., et al., “A Scoring System for Appraising Mixed Methods Research, and Concomitantly Appraising Qualitative, Quantitative and Mixed Methods Primary Studies in Mixed Studies Reviews,” International Journal of Nursing Studies 46, no. 4 (2009): 529–546, 10.1016/j.ijnurstu.2009.01.009. [DOI] [PubMed] [Google Scholar]
- 18. Higgins J. P. T., Thomas J., Chandler J., et al., editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.5 (updated August 2024). Cochrane, 2024. www.cochrane.org/handbook [Last accessed: Feb 3, 2026].
- 19. A Summary of General Assessment Factors for Evaluating the Quality of Scientific and Technical Information ‐ Prepared for the U.S. Environmental Protection Agency by Members of the Assessment Factors Workgroup, a Group of the EPA's Science Policy Council (Washington, DC: US EPA, 2003). [Google Scholar]
- 20. Dakkoune A., Vernières‐Hassimi L., Leveneur S., et al., “Risk Analysis of French Chemical Industry,” Safety Science 105 (2018): 77–85, 10.1016/j.ssci.2018.02.003. [DOI] [Google Scholar]
- 21. Jacobsson A., Sales J., and Mushtaq F., “A Sequential Method to Identify Underlying Causes From Industrial Accidents Reported to the MARS Database,” JRC54720 22, no. 2 (2009): 197–203, 10.1016/j.jlp.2008.12.009. [DOI] [Google Scholar]
- 22. Uth H. J. and Wiese N., “Central Collecting and Evaluating of Major Accidents and Near‐Miss‐Events in the Federal Republic of Germany‐‐Results, Experiences, Perspectives,” Journal of Hazardous Materials 111, no. 1–3 (2004): 139–145, 10.1016/j.jhazmat.2004.02.022. [DOI] [PubMed] [Google Scholar]
- 23. Okoh P. and Haugen S., “A Study of Maintenance‐Related Major Accident Cases in the 21st Century,” Process Safety and Environmental Protection 92, no. 4 (2014): 346–356, 10.1016/j.psep.2014.03.001. [DOI] [Google Scholar]
- 24. Sales J., Mushtaq F., Christou M. D., et al., “Study of Major Accidents Involving Chemical Reactive Substances: Analysis and Lessons Learned,” Process Safety and Environmental Protection 85, no. 2 (2007): 117–124, 10.1205/psep06012. [DOI] [Google Scholar]
- 25. Wen J. X., Marono M., Moretto P., et al., “Statistics, Lessons Learned and Recommendations From Analysis of HIAD 2.0 Database,” International Journal of Hydrogen Energy 47, no. 38 (2022): 17082–17096, 10.1016/j.ijhydene.2022.03.170. [DOI] [Google Scholar]
- 26. Darbra R. M., Palacios A., and Casal J., “Domino Effect in Chemical Accidents: Main Features and Accident Sequences,” Journal of Hazardous Materials 183, no. 1–3 (2010): 565–573, 10.1016/j.jhazmat.2010.07.061. [DOI] [PubMed] [Google Scholar]
- 27. Nivolianitou Z., Konstandinidou M., and Michalis C., “Statistical Analysis of Major Accidents in Petrochemical Industry Notified to the Major Accident Reporting System (MARS),” Journal of Hazardous Materials 137, no. 1 (2006): 1–7, 10.1016/j.jhazmat.2004.12.042. [DOI] [PubMed] [Google Scholar]
- 28. Yalcin E., Ciftcioglu G. A., and Guzel B. H., “Human Factors Analysis by Classifying Chemical Accidents Into Operations,” Sustainability 15 (2023): 8129, 10.3390/su15108129. [DOI] [Google Scholar]
- 29. Kidam K. and Hurme M., “Design as a Contributor to Chemical Process Accidents,” Journal of Loss Prevention in the Process Industries 25, no. 4 (2012): 655–666, 10.1016/j.jlp.2012.02.002. [DOI] [Google Scholar]
- 30. Chen M., Wang K., Guo H., et al., “Human Factors of Fire and Explosion Accidents in Petrochemical Enterprises,” Process Safety Progress 38 (2019): e12043, 10.1002/prs.12043. [DOI] [Google Scholar]
- 31. Jiang W., Han W., Zhou J., et al., “Analysis of Human Factors Relationship in Hazardous Chemical Storage Accidents,” International Journal of Environmental Research and Public Health 17 (2020): 6217, 10.3390/ijerph17176217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Wang B., Li D., and Wu C., “Characteristics of Hazardous Chemical Accidents During Hot Season in China From 1989 to 2019: A Statistical Investigation,” Safety Science 129 (2020): 104788, 10.1016/j.ssci.2020.104788. [DOI] [Google Scholar]
- 33. Yang D., Zheng Y., Peng K., et al., “Characteristics and Statistical Analysis of Large and above Hazardous Chemical Accidents in China From 2000 to 2020,” International Journal of Environmental Research and Public Health 19, no. 23 (2022): 15603, 10.3390/ijerph192315603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Zhang H. D. and Zheng X. P., “Characteristics of Hazardous Chemical Accidents in China: A Statistical Investigation,” Journal of Loss Prevention in the Process Industries 25 (2012): 686–693, 10.1016/j.jlp.2012.03.001. [DOI] [Google Scholar]
- 35. Zhao L., Qian Y., Hu Q. M., et al., “An Analysis of Hazardous Chemical Accidents in China Between 2006 and 2017,” Sustainability 10, no. 8 (2018): 2935, 10.3390/su10082935. [DOI] [Google Scholar]
- 36. Zhou K., Xiao L., Lin Y., et al., “A Statistical Analysis of Hazardous Chemical Fatalities (HCFs) in China Between 2015 and 2021,” Sustainability 14, no. 4 (2022): 2435, 10.3390/su14042435. [DOI] [Google Scholar]
- 37. Anderson A. R., and Centers for Disease Control and Prevention (CDC), “Top Five Chemicals Resulting in Injuries From Acute Chemical Incidents—Hazardous Substances Emergency Events Surveillance, Nine States, 1999‐2008,” MMWR Supplements 64, no. 2 (2015): 39–46. [PubMed] [Google Scholar]
- 38. Baybutt P., “Insights Into Process Safety Incidents From an Analysis of CSB Investigations,” Industries. Journal of Loss Prevention in the Process Industries 43 (2016): 537–548, 10.1016/j.jlp.2016.07.002. [DOI] [Google Scholar]
- 39. Dorevitch S., Forst L., Conroy L., et al., “Toxic Inhalation Fatalities of US Construction Workers, 1990 to 1999,” Journal of Occupational and Environmental Medicine 44, no. 7 (2002): 657–662, 10.1097/00043764-200207000-00012. [DOI] [PubMed] [Google Scholar]
- 40. Orr M. F., Wu J., Sloop S. L., and Centers for Disease Control and Prevention C. D. C., “Acute Chemical Incidents Surveillance—Hazardous Substances Emergency Events Surveillance, Nine States, 1999‐2008,” MMWR Supplements 64, no. 2 (2015): 1–9. [PubMed] [Google Scholar]
- 41. Sadeghi Gelvardi F., Khoshmanesh B., Fatemi F., et al., “Analyzing the Hazardous Chemical Accident in Iran: A Statistical Investigation,” IJOH 12, no. 2 (2020): 140–153. [Google Scholar]
- 42. Gunasekera M. and de Alwis A., “Process Industry Accidents in Sri Lanka: Analysis and Basic Lessons Learnt,” Process Safety and Environmental Protection 86 (2008): 421–426, 10.1016/j.psep.2008.05.002. [DOI] [Google Scholar]
- 43. Jung S., Pak S., Lee K., et al., “Classification of Human Failure in Chemical Plants: Case Study of Various Types of Chemical Accidents in South Korea From 2010 to 2017,” International Journal of Environmental Research and Public Health 18, no. 21 (2021): 11216, 10.3390/ijerph182111216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Kannan P., Flechas T., Mendez E., et al., “A Web‐Based Collection and Analysis of Process Safety Incidents,” Journal of Loss Prevention in the Process Industries 44 (2016): 171–192, 10.1016/j.jlp.2016.08.021. [DOI] [Google Scholar]
- 45. Maniar M., Kumar A., and Mentzer R., “Global Process Safety Incidents in the Pharmaceutical Industry,” Journal of Loss Prevention in the Process Industries 68 (2020): 104279, 10.1016/j.jlp.2020.104279. [DOI] [Google Scholar]
- 46. Schenk L. and Öberg M., “Comparing Data From the Poisons Information Centre With Employers’ Accident Reports Reveal Under‐Recognized Hazards at the Workplace,” Annals of Work Exposures and Health 62, no. 5 (2018): 517–529, 10.1093/annweh/wxy009. [DOI] [PubMed] [Google Scholar]
- 47. Sutton I., “Use Root Cause Analysis to Understand and Improve Process Safety Culture,” Process Safety Progress 27 (2008): 274–279, 10.1002/prs.10271. [DOI] [Google Scholar]
- 48. Shappell S., Detwiler C., Holcomb K., et al., “Human Error and Commercial Aviation Accidents: An Analysis Using the Human Factors Analysis and Classification System,” Human Factors: The Journal of the Human Factors and Ergonomics Society 49 (2007): 227–242, 10.1518/001872007X312469. [DOI] [PubMed] [Google Scholar]
- 49. Svedung I. and Rasmussen J., “Graphic Representation of Accident Scenarios: Mapping System Structure and the Causation of Accidents,” Safety Science ‐ SAF SCI 40 (2002): 397–417, 10.1016/S0925-7535(00)00036-9. [DOI] [Google Scholar]
- 50. Shappell S. and Wiegmann D., The Human Factors Analysis and Classification System‐HFACS (2000). [PubMed] [Google Scholar]
- 51. Park J., Hisanaga N., and Kim Y., “Transfer of Occupational Health Problems From a Developed to a Developing Country: Lessons From the Japan‐South Korea Experience,” American Journal of Industrial Medicine 52, no. 8 (2009): 625–632, 10.1002/ajim.20723. [DOI] [PubMed] [Google Scholar]
- 52. Moyce S. C. and Schenker M., “Occupational Exposures and Health Outcomes Among Immigrants in the USA,” Current Environmental Health Reports 4, no. 3 (2017): 349–354, 10.1007/s40572-017-0152-1. [DOI] [PubMed] [Google Scholar]
- 53. Demikhov O., Opanasiuk Y., Demikhova N., et al., “A Digital Transformation Into Occupational Health and Safety Systems: A Review of the Best Practices in Europe,” Agronomy Research 21, no. 2 (2003): 674–692, 10.15159/AR.23.083. [DOI] [Google Scholar]
- 54. Lu Y., Wang T., and Liu T., “Bayesian Network‐Based Risk Analysis of Chemical Plant Explosion Accidents,” International Journal of Environmental Research and Public Health 17, no. 15 (2020): 5364, 10.3390/ijerph17155364. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting File 1
Supporting File 2
Supporting File 3
Supporting File 4
Supporting File 5
Data Availability Statement
The data that support the findings of this study are available in the supporting material of this article.
