Near-miss accidents data analysis and knowledge dissemination in water construction projects in Oman

Abstract

The strategic role of near-miss data in safety management is undeniably vital. This data provides early warnings of potential accidents, thus serving as a proactive tool for recognizing weak points in safety management and preventing disasters. Reinforcing this premise, an investigation was launched to delve into the classification and analysis of near-miss accident data, aiming to augment safety prevention knowledge in the Nama Water Services (NWS) Company. The study engaged in comprehensive activities, including gathering near-miss reports, engaging with HSE personnel, formulating a standardized near-miss data form, and analyzing the collected data. Additionally, it proposed a structured guideline for identifying, analyzing, and classifying near-miss incidents, hinging on various parameters like types of hazards, operations, time of occurrence, and project segments. The findings highlighted that the construction projects involving pipelines and reservoirs experienced the most incidents, predominantly occurring during weekday mornings and afternoons. Major hazards included falls from scaffolds, falls within trenches, and lax housekeeping. Most notably, the operations most frequently associated with near-miss incidents were excavation, lifting, and loading. A deep dive into the root causes revealed that poor supervision, inadequacies in the safe work system, insufficient control measures, and inadequate training were the primary culprits. Furthermore, the study indicated that workers and equipment bore the brunt of these incidents.

1. Introduction

Safety is commonly associated with the culture of an industry, organization, or individual. According to Ref. [1], safety refers to protecting people or property from harm or danger and controlling recognized hazards to achieve an acceptable level of risk. As a result, safety can be viewed as a state where loss of life or property, injury, or damage is prevented. A considerable amount of effort has been made to promote safe working environments, with the International Labor Organization (ILO) leading global efforts, while governments, academic institutions, private companies, and independent organizations like the Occupational Health and Safety Association (OHSA) are working towards the same goals at national and organizational levels. Companies have also established internal mechanisms and private organizations to create safe work environments and provide safety services to other companies.

The International Labor Organization (ILO) regularly collects statistics to monitor the safety of work environments. However, their data suggests that there is a significant issue of underreporting of occupational safety incidents in workplaces. Despite this challenge, the ILO continues to make available the data they have collected. In a recent report, they estimated that approximately 2.3 million women and men worldwide experience work-related accidents or diseases each year, resulting in 6000 deaths per day [2]. The International Labor Organization [2] also notes that there are roughly 340 million occupational accidents and 160 million cases of work-related illnesses worldwide annually. Furthermore, their data indicates that younger and older workers are particularly susceptible to safety incidents. These statistics highlight the need for continued efforts to improve workplace safety and prevent accidents and illnesses in the workplace, particularly for vulnerable groups.

The construction industry, much like the manufacturing industry, is involved in the production of various infrastructure and facilities, such as roads, bridges, dams, water supply systems, and buildings. Due to the nature of their work, construction companies are faced with safety issues and incidents on a daily basis, particularly in the field or within a factory setting. The type and intensity of these safety issues depend on the task, operation, and logistics involved, such as work method, equipment, and manpower. Safety issues can range from missed incidents to severe and fatal incidents. According to Alcumus [3], the construction industry has the highest number of worker deaths compared to any other industry, with 10 fatalities per 100,000 full-time equivalent (FTE) workers, which is a concerning number.

Missed incidents play a crucial role in minimizing or avoiding safety incidents, as they provide early warning signs that can be used to take corrective action without causing harm to property or lives. To harness the knowledge associated with missed incidents, clear documentation and record-keeping of each incident, along with the dynamics surrounding it, are essential. However, due to historical reasons, the construction industry has been lacking in a culture of documentation. Nevertheless, some companies exhibit good and mature documentation practices, and there are opportunities to learn from their missed incident records. For example, a company operating in the water sector in Oman was identified, and collaborative work was done to study their missed incident records and extract knowledge to inform preventative strategies.

Safety incidents in the construction industry have both direct and indirect consequences, such as injury or loss of life, property damage, and project delays and budget overruns. Poor or inadequate safety management practices can lead to an unsafe work environment, resulting in such consequences. Proper documentation and control are crucial aspects of safety management, yet the construction industry tends to have a poor culture of documentation and control, particularly regarding safety issues [4]. This problem is exacerbated by a lack of knowledge and awareness about how to utilize information from existing documents. Consequently, there is insufficient evidence to justify the benefits of good safety record keeping within the construction industry, which perpetuates the culture of poor documentation and control.

In the construction industry, a range of safety incidents can occur, from mild to severe incidents, as well as missed incidents, which provide valuable opportunities for learning and improvement [5]. Scholars have established a correlation between near-miss events and major accidents, indicating that missed incidents can provide insights into potential future accidents [6]. However, gaining precise insights into missed safety incidents requires proper documentation and analysis of each incident.

Understanding the nature and patterns of missed safety incidents is critical in preventing future incidents and ensuring the safety of workers and the general public. By characterizing missed incidents, the study can help to identify the root causes of these incidents and develop strategies to prevent their recurrence. Moreover, analyzing the reporting and handling process of missed incidents can reveal any shortcomings in the current practices and suggest improvements to the existing systems. Therefore, the study introduces a systematic guideline for identifying, analyzing, and classifying these near-miss incidents, considering factors such as hazard types, operations, occurrence time, and project segments. Furthermore, the study conducts a detailed examination of the most frequently encountered hazards and operations associated with these incidents, thereby providing valuable insights for targeting specific areas of improvement and reducing the likelihood of potential accidents.

The findings of this study can be used to enhance safety management practices in the construction industry, particularly in the water supply project domain. It can also provide valuable insights into the benefits of good safety record-keeping, which can help to create a culture of safety documentation and control in the construction industry.

2. Study objective

The primary objective of this study is to investigate the missed safety incidents that occur during the construction phase of water supply projects. The study aims to explore the patterns and nature of these incidents to identify their root causes and develop effective strategies for preventing their recurrence. The study has set out the following specific objectives to achieve this goal:

• -to significantly advance understanding of near-miss accidents, particularly through developing a systematic guideline for identifying, analyzing, and classifying these incidents. This guideline considered key factors such as types of hazards, operations, time of occurrence, and project segments, providing a comprehensive framework for assessing near-miss events and identifying their causes.
• -to offer valuable insights into the most frequent hazards and operations associated with near-miss accidents. These findings can be used to target specific areas for improvement and implement necessary safety measures, consequently reducing the likelihood of accidents in the future.
• -to unveil the primary root causes behind near-miss incidents.

The findings of this study can enhance safety management practices in the construction industry by promoting good safety record-keeping and creating a culture of safety documentation and control. The study's insights can help prevent future accidents and injuries in the construction industry, ultimately improving the safety and well-being of workers and the public.

3. Background

3.1. Safety management

Safety is a multi-facet phenomenon that can be abstracted using different constructs. These constructs include the safety logistic and resources, safety programs (e.g., trainings, measurement and diagnostics, investigations, etc.), the personnel in-charge of safety plus the workers, operational procedures, protocols and rules, etc. As such, customized management systems need to be developed and utilized in such a way that all these different aspects are taken care of in a reasonable fashion. These systems exist in different industries including the construction industry.

The purpose of construction safety management is to help organizations pinpoint gaps in their safety program and processes, so that they can proactively mitigate risks and protect the workforce. Safety management systems serve as an effective way to communicate safety messages to all the operations at an organization and a system that connects office and field employees as one from a safety perspective [3]. Safety management systems are comprised of different components such as the network of people responsible for developing, revising, and implementing it, the technological logistics (computing devices, monitoring devices), and the programs i.e., the rules, procedures, protocols, training, investigations, etc.

According to Flight Safety Foundation [7], scholars within the safety domain highlighted that safety management systems serve the following functions: ensuring that workers have and use safety equipment, allocating required safety resources, enforcing safety rules, the inclusion of safety in performance reviews, provision of safety training and coaching to workers, monitoring worker safety performance, and measurement, reporting and diagnostics, and carrying out comprehensive investigations into safety incidents. These may also be perceived as the roles and responsibilities within a safety management system.

Safety management protocols and systems have been successfully developed and applied in real practical situations to achieve desirable safety conditions. The Flight Safety Foundation [7] states that active involvement of workers at the different levels, i.e., frontline workers, supervisors, coordinators, and management, in the implementation of safety management programs is critical for their success. This can be achieved in several different ways such as following established safety procedures, reporting hazards and incidents, assisting in safety investigations, participating in safety committees tasked with development and reviewing safety programs.

3.2. Safety incidents

Safety has got three features associated with it just like any other risk event. It has got an event, a likelihood of occurrence of that event and an impact of the event occurring. In safety, the impacts are the aspect that has largely been studied and typically relate to the people and property that is in the vicinity of the safety event at the time of its occurrence. For people, impacts range from injury to fatalities while for property, the impacts range from damage to total loss.

Safety events have a root cause associated with their occurrence. In the construction industry, there are several possible root causes of safety events. The most commonly cited ones in literature include: falls from height, trapped by something collapsing or overturning, struck by a moving vehicle, contact with electricity or electrical discharge, struck by a flying/falling object during machine lifting of materials, contact with moving machinery or material being machined, exposure to a hot or harmful substance, etc. [8]. Those are low-level root causes. At a high-level, causes of safety incidents include: poor construction planning, a lack of design safety, insufficient safety training, worker behaviours, inherent security construction-related health and safety risks, as well as a lack of understanding of site rules [9].

3.3. Safety performance measurement

Performance measurement is an integral part of any sound management system as it provides a basis for assessing the state of a system and provides information that can support diagnostic and improvement work. This is particularly necessary in the domain of safety given the associated stakes. Several ways have been devised for tracking and measuring safety. The most popular of these is through the use of lagging and leading indicators. Lagging indicators are those which are tracked and measured after the occurrence of a safety incident. Leading indicators on the other hand track activities that are performed to prevent or reduce the occurrence of or severity of a safety incident in the present or future.

Popular examples of lagging indicators for safety include incident rate, lost time work injury. Incident rate refers to the frequency with which safety incidents occur. This is indicative of the type of safety issue that re-occurs often in the workplace. Typical examples of lead indicators for safety include safety training, safety audits, safety inspections, etc. Leading indicators have an important role in strengthening safety and health outcomes in a workplace through proactive preventive measures against the occurrence of fatalities, injuries and illnesses. Leading indicators have also been used to expose potential problems in a safety and health program and assess the effectiveness of such programs.

3.3.1. Leading indicators

3.3.1.1. Safety training

To ensure safe work practices, frontline workers, managers, and coordinators must have knowledge and skills in safety as well as familiarity with the materials, equipment, and processes involved in their work. Therefore, safety training plays a crucial role in establishing a safe work environment. Safety training programs typically cover safe work procedures and protocols, proper use of safety equipment such as safety vests, fall protection, boots, and hard hats, as well as jobsite materials and equipment. These programs also include instruction on inspection, supervision, measurement, reporting, and diagnostics of safety in the workplace. In addition, training may also involve teaching workers and supervisors to understand and implement predetermined safety standards [10].

Another essential safety practice is conducting periodic safety audits or inspections, which involve observing and supervising workers to ensure compliance with safety measures and protocols. Safety audits may also involve taking actual measurements of safety conditions and performance in the workplace. These diagnostics are used to gather information and data to help identify potential safety hazards and to establish proactive preventive measures [11].

Overall, training and auditing are critical components of a comprehensive safety program that helps ensure workers are knowledgeable, equipped, and empowered to work safely, while also providing supervisors and managers with information to continually improve and enhance the safety of the workplace.

3.3.2. Lagging indicators

3.3.2.1. Incident frequency rates

Sound safety management systems are crucial for any workplace that values the safety and well-being of its workers. One key aspect of these systems is the inclusion of safety measurement protocols that report incident details such as the type, cause, impact, location, and time of occurrence [12,13]. By analyzing the reported time of occurrence, organizations can conduct higher-level analysis to obtain incident rates of occurrence. Incident rates provide a temporal risk profile of specific incident types, including associated variables like space and activity, which are essential in proactive safety management.

One important safety performance measure that safety management systems track is the incident frequency rate. This measure indicates the frequency of incidents that have occurred within a specific timeframe and is calculated by dividing the total number of incidents by the total hours worked during that period. Although it is a lagging safety indicator and can only be determined after-the-fact, the incident frequency rate is still a critical measure for organizations [12,13]. It helps to identify areas of concern and supports the implementation of corrective measures to prevent future incidents.

Overall, safety measurement protocols are vital components of safety management systems that promote safe working environments. It is essential to track a range of safety performance measures, including the incident frequency rate, to ensure the safety of all employees in the workplace.

3.3.2.2. Lost time work injury

Safety incidents can lead to different impacts on both humans and property. While property damage is measured in terms of the extent of damage, the impact on human beings can manifest as an injury or a disease that may hinder their ability to perform their duties at work. "Lost time" due to an injury or disease is a common consequence of safety incidents. Workers may need to take time off work to seek treatment and recover fully before they can resume their duties. A "lost time" claim is typically made in such situations.

[14] defines a “lost time” claim as a claim made when a worker is off work past the day of the incident, experiences a loss of wages/earnings, or experiences permanent disability or impairment. The end of the lost time period varies and is determined by the worker's safety board. The board reviews clinical information related to the affected worker to decide whether the worker is fit to go back to their pre-injury work or suitable and available work.

Understanding the impacts of safety incidents and the process of making a “lost time” claim is essential to ensure a safe working environment. It is important to acknowledge that workers have the right to claim “lost time” and that such incidents can affect their ability to perform their duties at work. Therefore, it is crucial to implement measures that prevent safety incidents and reduce the potential impact on workers and property.

3.4. Near-miss definition

A near-miss can be defined as an incident that is close to happening and has the potential to cause harm or damage but lacks a few of the ingredients or generating mechanisms that may eventually lead to an accident. The causes of near-misses and accidents are often similar or identical, which means that learning from the former and addressing those causes can prevent the development of an accident. According to Ref. [15], a near-miss is an unsafe act, an unsafe condition, or a dangerous situation that could lead to an incident with consequences if opportunity factors are not disrupted.

The definition of a near-miss can vary depending on the industry, with different interpretations and emphasis on improving safety performance. Ritwik [16] defined a near-miss in the petrochemical industry as a state of unsafe act or behavior that could cause property damage, environmental release, or injuries, while Best et al. [17] defined near-miss events in clinical medicine as unplanned or unexpected events that have the potential to result in human death or serious consequences.

Muermann et al. [18] defined a near-miss as a signal or series of phenomena or circumstances of abnormal occurrence that offer a chance for system improvement by eliminating risks. NASA defines an accident precursor as off-nominal events that provide a warning about more severe consequences to come due to their discernible causes, while the National Academy of Engineering [19] defines a near-miss as an accident precursor, which is a condition, event, or sequence that precedes and can lead to an accident.

The Nuclear Regulatory Commission defines an accident precursor as an observed condition or event at a nuclear power plant that, in combination with other factors such as equipment failures or human error, may lead to core damage. Compared to an accident sequence, only a few ingredients or generating mechanisms are missing in a near-miss sequence, resulting in a few missing events or truncations in the accident sequence. Essentially, a near-miss is the same as an accident sequence, except that it lacks a few essential conditions and causal factors that could eventually lead to an accident. Learning from near-miss incidents and addressing the causes could prevent their development into accidents.

3.5. Key identification and reporting challenges of a NMS

Reporting near-miss incidents is a vital component of improving safety practices in any workplace. Haas and Yorio [20] discussed several approaches to reporting near-miss data, including voluntary or mandatory reporting by individuals who have experienced them or by designated personnel using observational methods or automated sensors. Encouraging anonymity in reporting near-misses can boost reporting rates, but concerns about liability may discourage individuals from reporting. Therefore, it is necessary to foster a culture of reporting and trust in the system by avoiding a punitive mind set when creating and maintaining a Near-Miss System (NMS) [21].

However, the literature has identified two significant obstacles to reporting near-miss data. The first is a lack of clarity about what should be reported. Van der Schaaf [22] highlights that without clear guidelines on what qualifies as a near-miss, individuals may not report incidents that they perceive as insignificant or not related to safety. Hence, defining near-miss incidents accurately is crucial to ensure that all relevant information is captured in reporting and to make the data meaningful.

The second obstacle to reporting near-miss data is concerns about the volume of data gathered and whether a decrease in report rates over time is desirable [22]. The intention of reporting near-misses should be to improve safety practices, not to gather a vast amount of data. While it is necessary to capture all relevant data, the reporting process should be streamlined and not overly burdensome. Moreover, the reporting culture should be reinforced by encouraging workers to report and respond to reported near-misses promptly.

3.5.1. Confusion regarding what is reportable to the NMS

Confusion among employees regarding near-miss or precursor events is a common issue and was found that 68 % of respondents were perplexed by the concept [23]. It is crucial to address this issue proactively when designing or implementing a Near-Miss Management System (NMS). The primary cause of confusion is the perception of a near-miss as an event-based idea rather than a pre-existing hazardous condition.

For instance, dropping a heavy load on a building site before it reaches its destination is an event-based near-miss incident that is reportable, even if no one was injured [24]. However, near-miss incidents resulting from adverse conditions, such as a poorly installed scaffold or a faulty safety valve in a plant, are less likely to be reported by employees. Thus, an effective NMS should not only focus on event-based incidents but also include adverse condition incidents to address safety blind spots and improve workplace safety.

3.5.2. Extent of near-miss data collected

Medium to large organizations often receive a significant volume of near-miss reports every year, ranging from hundreds to thousands. However, managing such a vast amount of data can be challenging without proper processes and filters in place to handle it. Near-miss reports are critical resources for improving safety awareness, conditions, and reflecting the organization's safety culture. Therefore, it is crucial to have appropriate filters in place downstream in the process to effectively interpret and evaluate the safety implications of the dataset.

In order to have a comprehensive near-miss program, organizations should have a clear definition of what qualifies as reportable incidents, data collection, storage, retrieve, and reporting protocols, tools and environments, an explanation of the program's scope, and thorough training on safety and risk awareness for employees. A critical issue when it comes to the number of reported near-miss incidents is whether the reporting initiative should aim to reduce or increase the number of reported near-misses. Reporting should be based on what should be reported, and employees should receive precise training to support this matter [25].

4. Methodology

The focus of this study was to analyze near-miss incidents in the water sector and disseminate knowledge about them. While near-miss incidents are commonly recorded in the construction sector, their utilization as proactive learning lessons is often limited. The study describes the methodology and steps taken to analyze raw near-miss data in the water sector, starting from topic selection to data analysis, as shown in Fig. 1. Previous research studies were reviewed to obtain a broad understanding of utilizing near-miss data to develop the occupational health and safety sector. A process of near-miss data collection was then implemented, alongside Health Safety and Environment (HSE) professional meetings to plan and agree on the data analysis and resolve any issues that arose during data analysis.

Fig. 1.

Research methodology flowchart.

As mentioned in the background, lagging indicators are widely recognized in safety management as retrospective measures that provide insights into past incidents, accidents, or undesirable outcomes. These indicators are often used to assess safety management systems' effectiveness and identify improvement areas. In this study, we utilized lagging indicators to analyze historical safety data and evaluate the performance of safety management practices within the studied context. By examining past incidents and their outcomes, we aimed to identify trends, patterns, and lessons learned that could inform future safety strategies. The study employed various lagging indicators to assess safety performance, such as:

• -Historical Incident Data: The analysis of past incidents provided valuable insights into the frequency, severity, and nature of safety events.
• -Injury and Illness Rates: this aims to assess the effectiveness of safety measures in reducing harm to workers by examining injury and illness rates over a specific period.
• -Property Damage and Loss Data: The study also considered property damage and loss data as lagging indicators to assess the impact of safety incidents on infrastructure, equipment, or resources.
• -Historical Compliance Records: This aims to identify recurring compliance issues or gaps that might contribute to safety incidents.

Due to the high variability of data sources, information, and formats of collected raw reports, the data had to be cleaned and restructured to fit the scope of this research. Finally, WEKA [26] software was used for data analysis. The study details all these steps in the following sections.

Before data analysis, several research studies were reviewed in occupational health and safety to identify the best practices in near-miss data management and dissemination of knowledge from such events. Different organizations and industries adopt a variety of near-miss management data analyses, and the design of a near-miss management system depends on the organization's needs, such as regulatory requirements or industry-specific implementation. This study implemented the best practices of near-miss management systems according to the data provided by NWS.

4.1. Data collection

During the construction phase of a project, there is continuous generation of data in real-time as the implementation progresses. It is the responsibility of those interested to set up and configure protocols to collect and store this data for various purposes. In this study, water supply projects in Oman were utilized as a case study, and safety data on near-miss incidents that occurred during the project's different components were used for analysis, inference, and subsequent discussion. The data was retrieved from the archives of the construction contracting company, which built the entire water supply project, but their identity was not disclosed for proprietary reasons.

The safety data was collected using standard safety reporting protocols set up by the construction company, which involved the collaboration of frontline workers, supervisors, safety coordinators, among others. Basic information was collected in hard copy, then scanned into softcopy, while others had data extracted and stored in MS Excel files. The safety data utilized in this case study was extracted from project safety reports into MS Excel files. Each column in the Excel file was dedicated to a unique feature, with columns utilized mapped directly to the feature vector for the safety dataset.

Four meetings with the HSE professional from Nama Water Services Company to plan the data analysis process and determine the necessary data. During these meetings, they discussed and agreed upon the procedure for data analysis. The raw data collected were in the form of near-miss reports and were collected from 2014 to 2020. The number of near-miss reports received per year is presented in Table 1. These meetings were crucial in ensuring that the collected data was relevant to the research question and that the data analysis process was comprehensive and accurate. The meetings also provided an opportunity for the researchers to clarify any issues that arose during the data collection process and to ensure that the data was of high quality. By collecting data from multiple years, the researchers were able to analyze trends in near-miss incidents over time, which is useful in identifying areas for improvement in occupational health and safety.

Table 1.

Number of received reports from NWS.

Year	2014	2015	2016	2017	2018	2019	2020
Number of Reports	148	175	402	645	768	424	413

4.2. Data preparation

The raw near-miss reports received were subject to an extensive review to identify the relevant information that could be extracted and utilized for further analysis. Although the data was retrieved from a single source, the reports containing this data came from various sources, typically the different contractors that reported their incidents to NWS - a client company responsible for gathering all the incident reports and storing them in a database. During the review process, several issues were encountered, making these reports unsuitable for direct analysis. The majority of the raw report formats varied considerably among different reporters, leading to significant data variations. In addition, many reports lacked essential details necessary for incident analysis. For example, some reports did not indicate project type, incident type, incident subcategory, or classification, which are vital pieces of information to identify the areas with the highest frequency of incidents and the type of risk that occurred more frequently than others.

To make the basic near-miss reports useful for analysis, a cleaning and preparation process was necessary to extract all relevant information and generate consistent data. This process involved identifying and resolving inconsistencies in the data, filling in missing information, and standardizing the format of the reports. By doing so, the resulting dataset provided a more accurate representation of the near-miss incidents and enabled a more comprehensive analysis of the underlying causes and contributing factors. This process was crucial to obtain meaningful insights and develop effective strategies to prevent similar incidents from occurring in the future.

It was essential to conduct a thorough investigation of near-miss incidents to identify the root and immediate causes of the incident and to assess the severity of the incident in terms of its actual and potential consequences. However, many near-miss reports lacked this vital information, making it difficult to gain a clear understanding of the incident scene and assess the risk involved. Additionally, the format of these reports varied significantly, with some in handwritten format that was barely legible.

4.3. Data restructuring and classification

4.3.1. Reports filtration and categorization

Between 2018 and 2020, NWS received more than 1605 near-miss incident reports from various contractors and projects, which represented more than half of all reported incidents. To gain an understanding of the latest incident reporting trends, the researchers selected raw near-miss reports from this period and filtered them to ensure they contained essential data. Reports that were missing crucial information were discarded, leaving 774 reports for further analysis. This filtration process was crucial to ensuring that the resulting dataset was reliable and accurate.

By analyzing these reports, the organization can identify areas for improvement and implement measures to prevent similar incidents from occurring in the future. The selected timeframe reflects the most recent incidents, providing insights into the current safety trends within the organization. By understanding these trends, the organization can make informed decisions to address areas that require improvement and take proactive steps to prevent future incidents.

To ensure a more effective analysis of the selected near-miss reports, it was crucial to establish a clear categorization and classification system. To accomplish this, a thorough review of prior studies was conducted to identify the most appropriate categories and classifications for the near-miss reports and their circumstances. This was necessary to ensure that the extracted data could be sorted into more reliable groups for analysis. The incident categorization consisted of two categories: accidents and near-misses. Accidents were further divided into minor or significant sub-categories, which were not applicable to near-misses. The report's categories and sub-categories are illustrated in Fig. 2. By establishing a clear categorization and classification system, the analysis of the data could be conducted more efficiently and accurately, enabling the identification of patterns and trends in the types of incidents that occurred.

Fig. 2.

Incident report category and sub-category.

4.3.2. Event categorization and classification

After conducting a comprehensive review of previous studies and existing industry best practices, the National Examination Board in Occupational Safety and Health (NEBOSH)[27] categorization and classification was found to be the most suitable for the near-miss reports. However, these categories and sub-categories required some modification to make them more applicable to construction hazards. To achieve this, amendments were made to the NEBOSH categories based on scientific research in the construction sector. Additional event classifications were also introduced along with other modifications to enable more effective risk identification.

The modifications made to the NEBOSH categorization and classification system were necessary to ensure that the resulting dataset was accurate and reliable. By using a system that was specifically tailored to construction hazards, the organization could more effectively identify underlying causes and contributing factors, leading to more effective incident prevention measures. Furthermore, the additional event classifications would help to identify specific risks associated with near-miss incidents.

4.3.3. Data analysis techniques

4.3.3.1. Action taken

To ensure the accuracy and reliability of the dataset, a feature datasheet was created by extracting all the corrective measures mentioned in the raw near-miss reports. However, some reports did not mention disciplinary actions, and similar corrective measures were described differently, which presented a challenge for the analysis. For instance, some reports mentioned corrective actions such as convening meetings to raise workers' awareness or assigning workers for training or information. To address this issue, all these actions were grouped under the broader category of "Training, Information, and Instruction" to ensure consistency in the analysis.

Standardizing the corrective measures was crucial to identify patterns and trends in the incidents, which would allow the organization to implement appropriate measures to prevent similar incidents from happening in the future. This process was undertaken with consultation with the safety engineer to ensure that the resulting dataset was comprehensive and accurate.

4.3.3.2. Immediate and root causes analysis

In order to effectively identify the cause of an incident and develop appropriate control measures and corrective action, a comprehensive investigation must be conducted to pinpoint the underlying reasons for the incident. This investigation should identify both the immediate or direct causes and the underlying or root causes, with the latter being the events or failures that set off all other causes or failures.

The received near-miss incident reports were analyzed based on their immediate and underlying causes. However, in some cases, the exact reasons for the incidents were unclear and could not be inferred from the provided incident description. This highlights the importance of conducting a thorough investigation at the time of the incident and gathering all relevant information from all parties involved in the incident. Unfortunately, this was not feasible for this study. By identifying the underlying causes of incidents, the organization can implement appropriate measures to prevent similar incidents from occurring in the future, ultimately promoting a safer work environment.

4.3.3.3. Constructed feature vector data sheet

All the data extracted from the previous steps have been compiled into a construction feature vector datasheet, as shown in Table 2. This datasheet contains all the relevant data required for future analysis, including:

• -Project type
• -Time: this is to know the exact time of the event.
• -Weekday: range from Sunday to Saturday
• -Construction operation: There are various tasks and operations in each of the project segments mentioned above. Investigating the kind of operation underway when the incident occurred is paramount.
• -Root causes are the initiating causes that lead to a single or a sequence of outcomes. To eliminate such events from the occurrence, it is crucial to conduct an effective incident investigation to identify the immediate and root causes and propose the appropriate risk control.
• -Affected entity: this includes the main party affected by the incident. A person who is present in the workplace is injured or dies. Environmental pollution can occur depending on accidents, such as an oil spill. Company equipment and machines may be damaged, or public service may be disrupted.
• -Report Category: The description received of the incident may be either a near-miss report or an accident report.
• -Report Sub-category applies only to the accident, describing it as either significant (High loss or damage) or minor (low loss or damage).
• -Event category: in this part, the risk will be identified
• -Event sub-category
• -Event classification
• -Action taken

Table 2.

Near-miss features data sheet.

Report No	Project Type	Time	Weekday	construction operation/process/tasks	Root Couse	Affected	Report Category	Report Subcategory	Event category	Event Subcategory	Event Classification	Action Taken
1	Tank	9:30 a.m.	Thursday, January 2, 2020	Lifting/Loading	Failure to Provide adequate Training	Person	Near Miss	NA	Work Equipment	Load Handling Equipment	Load Falling	Training, Information, and Instruction
2	Pipeline	8:30 a.m.	Tuesday, January 7, 2020	Excavation	Failure to carry out Proper Risk Assessment	Service	Near Miss	NA	Service	Cable or Utility Strike	Electric Cable damage	Training, Information, and Instruction
3	Tank	10:30 a.m.	Sunday, January 19, 2020	Plastering	Failure to Provide adequate Training	Person	Near Miss	NA	Work Equipment	Hand Tools and Portable Power Tools	Loos handle	Training, Information, and Instruction
4	Pipeline	9:45 a.m.	Monday, January 27, 2020	Excavation	Failure to carry out Proper Risk Assessment	Service	Near Miss	NA	Service	Cable or Utility Strike	Electric Cable damage	Training, Information, and Instruction
5	Workshop	10:00 a.m.	Saturday, January 4, 2020	Steel reinforcement	Poor Supervision	Person	Near Miss	NA	Workplace Issues	Slips and Trips	Object on the floor	Training, Information, and Instruction
6	Workshop	3:00 a.m.	Saturday, January 11, 2020	Vehicle Related	Failure to Provide adequate Training	Person	Near Miss	NA	Work Equipment	Hand Tools and Portable Power Tools	Loos handle	Training, Information, and Instruction
7	Pipeline	9:00 a.m.	Wednesday, January 1, 2020	Excavation	Failure to carry out Proper Risk Assessment	Service	Near Miss	NA	Service	Cable or Utility Strike	Electric Cable damage	Supervision of Process
8	Workshop	10:00 a.m.	Friday, January 10, 2020	Lifting/Loading	Poor Supervision	Property	Near Miss	NA	Workplace Issues	Working at Height	Fall from Vehicle	Training, Information, and Instruction
9	Pipeline	12:00 p.m.	Wednesday, January 15, 2020	Hammering	Failure to Provide adequate Training	Person	Near Miss	NA	Work Equipment	Hand Tools and Portable Power Tools	Throwing off	Training, Information, and Instruction
10	Pipeline	9:00 a.m.	Sunday, January 26, 2020	Thrust Boring	Failure of Safe Systems of Work	Environment	Near Miss	NA	Chemical and Biological Agents	Chemical Agent	Oil Spill	Training, Information, and Instruction
11	Pipeline	10:30 a.m.	Thursday, January 23, 2020	Welding	Failure to carry out Proper Risk Assessment	Person	Near Miss	NA	Workplace Issues	Working at Height	Fall in Trench	Supervision of Process
12	Reservoir	5:20 p.m.	Wednesday, January 8, 2020	Electric task	Poor Supervision	Person	Near Miss	NA	Workplace Issues	Slips and Trips	Wet Surface	Training, Information, and Instruction
13	Reservoir	10:10 a.m.	Tuesday, January 21, 2020	Other	Inadequate Control Measures	Person	Near Miss	NA	Workplace Issues	Access and Egress	Unauthorize person	Safe System of Work
14	Pumping station	7:10 a.m.	Sunday, January 5, 2020	Dismantling/Installing	Failure of Safe Systems of Work	Person	Near Miss	NA	Workplace Issues	Access and Egress	No Handrail	Safe System of Work

5. Results and discussion

5.1. General data analysis

The study conducted data analysis on near-miss incident reports spanning from 2018 to 2020. During this period, a total of 768 near-miss reports were received in 2018, 424 in 2019, and 413 in 2020. However, not all of these reports were included in the analysis. Only 392 reports from 2018, 217 from 2019, to 165 from 2020 were analyzed, representing a decline in the number of analyzed reports over the years.

The analysis showed a decreasing trend in the percentage of recorded near-miss incidents, from 50.6 % in 2018 to 31.3 % in 2020, indicating an improvement in the organization's occupational health and safety practices. Fig. 3 illustrates this trend. The reduction in near-miss incidents is a positive sign that the organization is taking proactive measures to prevent workplace accidents and injuries.

Fig. 3.

Number of incidents analyzed per year.

The near-miss incident reports analyzed in this study revealed that the majority of incidents occurred on weekdays, with the highest number reported on Sunday and Thursday, ranging from 14.47 % to 20.28 %. In contrast, Friday and Saturday had the lowest number of reported incidents, with only 2.33 % and 8.4 % respectively.

It is logical that the number of incidents decreases on weekends since these are non-working days for many employees. However, it is noteworthy that incidents still occur on these days due to shift patterns or emergency work that needs to be completed promptly. In some projects, workers only work half a day on Fridays, which could explain the low incidence rate during that time period.

Analyzing incident patterns by day of the week is essential to identify the frequency and potential risks of incidents. This information can help organizations plan and implement appropriate measures to reduce the likelihood of incidents occurring during high-risk days or periods. Fig. 4 provides a visual representation of the number of near-miss incidents reported on each day of the week, highlighting the distribution and trend over the three-year period.

Fig. 4.

Time (Day) of incident.

The analysis of near-miss incidents presented in Fig. 5 revealed a steep increase in incidents from 7:00 a.m. to a peak at 10:00 a.m., which accounted for 21.3 % of all incidents. The number of incidents then began to decrease steadily until midnight, where it reached a low of 4.7 %. From 14:00 to 18:00, there was another increase in the number of incidents, with the highest percentage of incidents occurring at 10:00 a.m. This can be attributed to the fact that actual operations started at 10:00 a.m., with preparations such as conducting a toolbox meeting taking place earlier in the morning.

Fig. 5.

Time (Hour) of incident.

The period from 20:00 to 6:00 a.m. saw a low percentage of near-miss incidents, ranging from 0.1 % to 0.9 %. This can be attributed to a significant reduction in operations during this period, except for some work done in shift patterns and emergency jobs. The low number of incidents during this period could also be attributed to the fact that workers may be more alert during night shifts due to the higher risk involved.

5.2. Project segment analysis

The analysis revealed that pipeline construction activities had the highest number of reported near-miss incidents, accounting for 53.2 % of all incidents. This is due to the increased risk associated with activities such as excavation, lifting, and access/egress. The reservoir was the second most common project segment, accounting for 21.4 % of reported incidents, followed by the pumping station (6.8 %), tank (5.7 %), and administration building (4.0 %). Fig. 6 illustrates the distribution of near-miss incidents across the different project segments, providing a clear visual representation of the data. This information can be valuable in identifying areas that require improved safety measures and helping organizations to prioritize their efforts to mitigate risks in construction activities.

Fig. 6.

Project segments.

5.3. Operation and process analysis

Improving occupational health and safety is a critical goal for any organization, and investigating the type of operation in which an incident occurred is a crucial step towards achieving this goal. While incidents can occur in any operation or task, certain activities are more prone to incidents than others. Near-miss reports are categorized based on the type of operation, tasks, and processes involved, as illustrated in Fig. 7.

Fig. 7.

Construction operation, task, and process.

Excavation and trenching are among the most hazardous construction operations, and they are responsible for a significant proportion of incidents in the construction industry. In fact, excavation operations account for 27.4 % of all incidents, as shown in Fig. 7. Excavation-related incidents such as cave-ins during water pipeline construction have a higher likelihood of severe consequences, including fatalities. The high number of near-miss reports in excavation operations can be attributed to various factors that contribute to hazards, such as:

• -loose rock or soil that could fall
• -roll from an excavation face, materials or equipment that can pose a hazard by falling
• -rolling inside the excavation, loads being handled by lifting
• -excavating equipment that can fall on the excavation site
• -length and location of the excavation and exposure time.

In some cases, the water pipeline's mainline can be excavated for hundreds of kilometres, crossing various topographies and environments, such as roads, pedestrian areas, wadis, and mountains. This excavation can generate numerous hazards for residents, road users, and wildlife. Therefore, it is crucial to plan the length, location, and time of excavation and implement appropriate control measures to mitigate these hazards.

The loading and unloading process is the second most frequent operation involved in reported near-miss incidents, accounting for 8.3 % of incidents. Loading and unloading materials from one location to another is a common practice in construction work, and it can be performed either manually or mechanically. Among the loading and unloading activities, the transportation of pipes is the phase where the highest number of near-miss events is recorded. This can be attributed to several factors, such as:

• -The weight of the lifted load
• -Range of the lift
• -Location of the pipe
• -Size and shape of the pipe
• -Number and frequency of lifts performed
• -Proper selection of equipment for the loading and unloading tasks
• -Pinch points/line-of-fire area
• -Overloaded or improperly loaded (poor weight distribution) trucks/rails/trailers
• -Uneven and/or shifted loads reduce the overall stability of the load/vehicle
• -Missing or damaged strapping/tie-downs
• -means of securement (tie-down strap, banding, etc.)
• -Proximity of the operation to overhead power lines or other structures,
• -Presence of wind, fog, or other environmental hazards
• -improper selection and inspection of lifting rigging (chains, slings, straps etc.)
• -Qualified operators must run the lifting equipment and change ground conditions related to equipment stability.

5.4. Affected entity analysis

The assets that could be impacted by incidents are divided into five categories: worker, service, property, environment, and equipment. Workers make up the majority of the incidents analyzed, accounting for approximately 67.7 % of all reports. They are involved in nearly all processes, operations, and tasks, from excavation and lifting to loading and housekeeping. While some tasks can be automated, many still require human intervention. Leaving trenches open for an extended period, especially in residential areas, can pose a significant hazard to both people and road users. Some excavation activities lack barricades, isolation, or safety signs to indicate the presence of deep excavation. Although machinery is used for loading and unloading, some worker behaviours can put them in hazardous situations, such as setting or moving under a suspended load. Inadequate housekeeping can also contribute to slip and trip hazards.

Equipment is the second asset affected by incidents, accounting for 15.3 % of all reports. The machine operator's carelessness could be attributed to poor knowledge of how to operate equipment safely. Damage to service pipelines and cables, such as water pipeline damage, communication cable damage, and electric cable damage, accounts for 9.8 % of incidents. Most of this damage occurs during excavation due to poor trial techniques, as shown in Fig. 8.

Fig. 8.

Affected asset.

5.5. Event category

Upon analysing the near-miss reports, workplace issues were found to be the most frequent event category, accounting for 50.6 % of all incidents, followed by equipment categories (13.7 %) and services categories (9.4 %), as illustrated in Fig. 9. The workplace issue category is further divided into sub-categories, including working at height, confined space, lone working, slip and trip, vehicle movement, and driving. Among these sub-categories, working at height accounted for the highest number of reported near-miss incidents compared to other sub-categories. Working at height activities such as painting, installation, demounting scaffolds, and wall demolition require special attention due to the increased risk of accidents.

Fig. 9.

Event category.

Work equipment is another event category, with sub-categories including Hand Tools and Portable Power Tools, Machinery Hazards, and Load Handling Equipment. Different types of work equipment are used in construction activities, ranging from carpentry tools to heavy machinery like forklifts and welding machines. Due to the high number of incidents recorded in the excavation process, there is a high probability of damaging public services, such as water pipelines, sewage systems, electricity and communication cables, which are classified under the services categories.

5.6. Event sub-category

The twenty-seven sub-categories are grouped under the ten main categories. However, only twenty sub-categories were observed in the near-miss reports analyzed in this study, as depicted in Fig. 10. Although there were no reported incidents in the unmentioned sub-categories, they can still pose hazards in the construction sector. For instance, lone working, radiation, and mental health hazards. Nonetheless, working at height remains the most commonly reported sub-category in the analyzed near-miss reports for the years 2018, 2019, and 2020. This can be attributed to the nature of construction operations in the NWS projects, which involves activities like constructing water transmission lines, reservoirs, tanks, and pumping stations that require work to be done at height. Working at height poses a significant risk due to the severe consequences that may arise, including fatalities. Workers in the construction industry are often exposed to work at height, and this can include,

• -Workers are erecting the steel structure for a building
• -Erection and dismantling of scaffolds
• -Steel-framed buildings have their roofs clad with slates
• -A demolition crew uses machinery to demolish a multi-storey building
• -At the side of deep excavation, welders are working
• -Pipefitters installing pipework in the ceiling of a factory workshop.
• -Painters paint the external walls of a building.

Fig. 10.

Event sub-category.

Various types of equipment are used by workers in construction activities, including scaffolds, mobile elevating work platforms, and ladders, to facilitate their work. Scaffolds, in particular, are widely used in the construction industry due to their cost-effectiveness and simplicity. However, working at heights poses significant risks to workers that can be influenced by various factors such as:

• -The vertical distance is the main factor that should be considered and has the most potential risk-more than 2 m are considered as working at height since falling from such a distance can cause severe injuries.
• -Deteriorated materials: structure materials are exposed to different weather conditions and deteriorate with time. The structure should be in good condition to be used in operation. Structure materials can be broken when the person puts their weight on them, causing him to fall or materials broken and drop in person below.
• -Unprotected edge: the risk of falling people or objects increases when the working surface edges are not protected. Protected boundaries are necessary for scaffolds, elevated walkways, roofs and access platforms.
• -Unstable or Poorly Maintained Access Equipment: this can create a high risk if it is not stable or not secured very well. Access equipment suffers an unstable condition if it is poorly built, improperly sited or unsecured correctly, which may cause a catastrophic collapse or topple.
• -Weather: risk of slipping can be increased significantly in the freeze and rain conditions. During high wind conditions, materials are unstable, and they can be blown out.
• -Falling materials: considerable damage can be caused by falling materials from height, such as tools, equipment, waste materials or structural materials. Different factors can increase the likelihood of falling materials, such as deterioration of the structure, the unsafe position of storage materials, accumulation of waste leading to failing from the platform, gaps between two adjacent platforms or between wall and platform and wrong methods of getting materials.

5.7. Event category and sub-category comparisons

Event categories and event sub-categories were the top level the hierarchical constructs used in reporting safety incidents. Comparisons showing the extent of sub-category safety event occurrences for each category are summarized in the following Table 3. In the generation of these clusters from section 4.8 and 4.9, certain event categories were combined for convenience.

Table 3.

Clustering safety incident occurrence based on event categories and event sub-categories.

Event Category	Event Sub-Category	Percentage Occurrence
Work place issues	Collapse	5.6
	Fire	3.7
	Working in confined spaces	0.8
	PPE	5.6
	Access and egress	14.1
	Working at heights	17.3
	Strips and trips	13
	Vehicle movement	5.4
Environmental hazard	Biological agent	0.3
	Chemical agent	1.9
	Weather	1.3
	Wildlife hazard	0.8
Electricity	Cable/utility strike	9.4
	Electric shock	4.8
Work equipment	Machinery hazard	0.5
	Hand and portable power tools	5.4
	Load handling equipment	7.8
Physical, Psychological health and Muscoskeletal hazard	Hygiene	1.0

When safety incident data was clustered based on event category and event sub-category criteria, and the relative frequency of safety incidents occurring within each sub-category assigned, Table 3 and Fig. 11 were generated for purposes of obtaining a visual comparison between sub-categories.

Fig. 11.

A clustered representation of safety incident occurrence within event categories and event sub-categories.

The visual display generated in the form of a clustered donut indicates that safety incidents occur more frequently within that workplace issues category followed by the work equipment and electricity category. It is also apparent which safety event sub-category within each of these categories, experienced relatively high safety incidents during the execution of the water project.

5.8. Event classification

In the analysis of near-miss reports, falling in the trench and falling from a scaffold emerged as the most significant hazards, accounting for 5.9 % of all reported incidents. Falling in the trench can occur to anyone or anything near the trench, including equipment or materials, such as rocks or stones, which may slip and fall inside the trench. For instance, while welding activities are underway inside the trench, the welding machine is usually placed on plywood pieces near the trench's side. Any unidirectional movement of the machine can result in it falling into the trench, causing injuries to workers and damage to property.

Working inside the trench may also require the use of equipment such as cranes for loading or unloading purposes. The weight of such equipment can cause the trench walls to collapse and fall into the pit. On the other hand, falling from scaffolds is more frequently reported in the reservoir project segment, especially during the dismantling and installation of scaffolds due to the reservoir's shape, which necessitates the frequent use of scaffolds. Accidental falls of materials from the scaffold or reckless behavior of workers to speed up work can result in hitting someone on the ground or causing material damage. Overloading the scaffold platform beyond its structural capacity can cause the scaffold structure to bend, leading to its collapse.

However, many factors can contribute to people or materials falling from the scaffold, such as scaffolds built on soft ground without using adequate sole boards, inadequate tying of scaffolds to the building, insufficient bracing in the scaffold structure, bent, buckled, or corroded standards, incorrect couplers used to join tubes together, scaffold struck by mobile plant, scaffold erected by incompetent workers, or failure to inspect the scaffold before use. All these factors can create a hazardous situation for workers, equipment, and property, as depicted in Fig. 12. It is essential to identify such hazards and implement control measures to prevent incidents and protect workers' safety.

Fig. 12.

Event classification.

The pipeline and reservoir construction project segments showed a higher incidence of poor housekeeping. This can be attributed to several factors, including the remote location of pipeline activities, which makes it difficult to return materials to the storing yard until work is completed. The site's distance from management or less frequent management visits may also contribute to workers keeping equipment untidy and using it for reservoir operations. Additionally, the accumulation of waste in worker camps was observed, resulting from the workers' negligence and failure to keep their living quarters clean and tidy.

It is essential to maintain good housekeeping practices in the workplace for the safety and protection of workers. Poor housekeeping can lead to unsafe conditions, increasing the likelihood of direct cause injuries in the workplace. Slip, trip, and fall hazards are common consequences of poor housekeeping. Improper storage of materials, objects, equipment, and tools can cause workers to trip and fall, resulting in injuries.

Personal Protective Equipment (PPE) is the last line of defense in the hierarchy of safety control measures. While it can protect workers from severe injuries, its use should not be relied upon as the primary safety measure. Proper housekeeping practices should be implemented to prevent the need for PPE use in the first place. Employers should also provide training and education on the importance of good housekeeping practices and encourage their workers to maintain a tidy workplace.

The failure to wear and use the provided Personal Protective Equipment (PPE) has been identified as a major issue, accounting for 5.4 % of the recorded near-miss events. This is a cause for concern as not wearing PPE can have serious consequences, leading to injuries or fatalities. The majority of hazards reported were workers operating in excavated pits or trenches without a helmet. Welding and grinding without wearing safety glasses and a face shield, which can protect the eyes from welding light and radiation, was another type of failure to wear PPE in the workplace.

Working at height is another area where the use of appropriate PPE is critical. Performing a job using a scaffold, for example, should have full protection measures in place against falling from height by using a harness fixed in an appropriate hook point. Failure to wear appropriate PPE in such situations can result in serious injuries or even death, as observed during the wall-shattering process in the NWS operation.

Other observations such as not using proper gloves and safety shoes were also recorded. Failure to wear PPE can be attributed to several factors. For instance, the worker may be unaware of the protection that PPE can provide or may lack proper training on how to use and wear it. Employers have a responsibility to train their employees on the proper use of PPE, including when it is necessary, what PPE is necessary, and how to properly put on, take off, adjust, and wear PPE. They should also provide information on the limitations of PPE and proper care, maintenance, useful life, and disposal of PPE. By providing proper training and ensuring that employees understand and implement the training before starting any job that requires the use of PPE, employers can help reduce the risk of workplace injuries and fatalities.

6. Conclusion

In conclusion, this study focused on analyzing near-miss incidents in-water construction projects and drawing insights from these incidents. The collection of raw near-miss reports from a water construction company provided valuable data for analysis between 2014 and 2020. The projects were classified into several segments, including pipelines, reservoirs, pumping stations, administration building, chlorination room, instrumentation room, tanker filling station, water quality lab, desalination plant, and workshop. The pipeline segment was found to have the highest number of reported incidents due to its complexity, variety of operations, and extensive network of pipes. The reservoir segments followed closely behind.

Understanding the specific operations in which these incidents occur is crucial for enhancing safety measures. The analysis of the near-miss reports revealed that incidents could happen in any process or task, although certain operations carried a higher probability of incidents. Excavation, lifting, loading, and electrical processes emerged with the highest number of incidents. Moreover, more than half of the reported incidents predominantly affected individuals, including workers and residents, particularly in projects executed in residential areas.

A notable finding from the analysis was the decreasing trend in the number of reported incidents over time. This trend signifies the strong commitment of the company's management to health, safety, and the environment, with a clear objective of achieving zero accidents. The identification of focus areas for safety officers based on the analysis of near-miss reports can greatly contribute to better planning, monitoring, and control of construction activities, further improving overall safety performance.

Author contribution statement

Mohammed Al Shuaaili: Mubarak Al Alawi: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Ronald Ekyalimpa: Analyzed and interpreted the data; Wrote the paper.

Bader Al Mawli: Mohammed Al Shahri: Contributed reagents, materials, analysis tools or data.

Abdullah Al-Mamun: Contributed reagents, materials, analysis tools or data; Wrote the paper.

Data availability statement

The authors do not have permission to share data.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.