Abstract
Introduction
The construction industry is one of the most hazardous workplace sectors [1]. A preliminary analysis of American occupational fatalities by the United States Bureau of Labor Statistics revealed that in 2009, more fatal injuries occurred in the construction industry (18.8%; 816/4340) than in any other [2]. Furthermore, construction is a leading cause of work-related fatality in Australia, Canada, China, India, Spain, the United States, and New Zealand, and is therefore an occupational health hazard in both developed and developing nations [3–8].
According to data compiled by the Electrical Safety Foundation International, which initially derived from the US Bureau of Labour Statistics data, the construction industry was responsible for more than half of all occupational electrical fatalities between 2003–2010 [9]. Further, relative to all other industries, electrical fatality risk is four times greater in construction [6]. Although it is documented that the number of electrical fatalities in construction declined in recent years [9], electrical contact continues to be one of the foremost causes of injury and death among construction workers [6, 10–17]. Data from Ontario, Canada, show that the proportion of electricity-related occupational fatalities were electrical-related was second and third highest in the construction industry between, respectively, 2003–2007 and 2008–2012 [18].
Given the scale of these occupational fatalities, there is a pressing need to better understand their causal factors. To this end, a number of studies have identified risk-factors for electricity-related fatalities. For example, Janicak (2009) report that among older construction workers, contact with wiring and electrical transformers is the most common cause of fatality; in contrast, contact with overhead power lines was reported to be a more common cause of fatality among younger construction workers [15]. Ichikawa et al. (2013) show that establishing contact with power lines through body or tool is the most frequent cause of electrical fatality in construction, although they do not stratify their findings by age [19]. With respect to behavioural risk-factors, it has been documented that electrical fatalities are most often caused by “incautious activities”, or those activities undertaken without exercising appropriate safety measures [20]. Similarly, Zhao et al. (2013) report that unsafe worker behaviour and limited effectiveness of current safety programs contribute to electrical fatalities in construction [21]. Failure to properly use protective devices and follow safety procedures, especially by electricians, is also an established risk-factor for electrical fatality [22]. Chi et al. (2012) provide a detailed breakdown of fatal electrocutions in construction, stratified by electrical source, type of contact, and source of injury. Furthermore, seasonality may influence electrical fatality risk, as the majority electrical fatalities occur between June and September [23].
Although informative, the above studies do not contextualize their findings within a Haddon’s Matrix, a widely used and effective paradigm in injury prevention research [24, 25]. Understanding electrical fatalities in construction through use of a Haddon’s Matrix will help us better inform strategies for prevention, given that the Matrix facilitates a multi-dimensional understanding of the causes of injury or death [26].
The present study is part of a larger investigation of 200 construction fatalities in Ontario, Canada that occurred over a 10-year period, ending in 2007. Given the complexity of construction fatalities, our objective was to adapt a Haddon’s Matrix and describe the characteristics of workers, agents, and environmental factors that contribute to fatal electrocutions in the construction industry. This present study informs strategies for prevention that can be applied in Ontario and abroad.
Materials and methods
Design
The present study was a cross-sectional examination of retrospective data.
Source of data
Data were abstracted from Ontario Ministry of Labour (MOL) records on construction-related fatalities that occurred between 1997 to 2007. Records that were part of formal proceedings (coroner’s inquest or legal) during the time of data collection were not included in this study. All available records were screened by Freedom of Information and Privacy personnel prior to release. Data extraction was carried out in a secure MOL location by two members of the Infrastructure Health and Safety Association (IHSA) who were experienced in construction safety and familiar with the data abstraction tool, namely, The Construction Industry Fatality Case Review Tool. This tool allowed records to be evaluated across the following categories: Demographic; Employer; Accident; Project; Site Supervision & Safety. The tool also examines four ‘human factor’ categories: Training Related; Person Related; System Related; and Organization Management. The human factors were based on the work of Hollnagel [27] and Reason [28].
Cause of electricity-related fatality, including electrical contact leading to electrical shock, burn, or explosion, was identified using Ontario [29] and United States [2] codes. However, to assign an immediate cause of electricity-related fatality, the present study adopted a narrative text analysis of the Census of Fatal Occupational Injuries (CFOI) database, using methods similar to those employed by Cawley and Homce [13].
We used a Haddon’s Matrix [24, 25] as a framework to categorize the available variables as host, agent and environmental factors. These matrices have been widely used in injury epidemiology, and were originally designed to evaluate the human, equipment, and environmental factors that cause motor collisions across three temporal dimensions. Our study modified the Haddon’s Matrix to make it context appropriate for occupational electrocutions in the construction industry. Specifically, the ‘human’ and ‘equipment’ categories were adapted into ‘worker’ and ‘electrical agent/equipment’ categories, respectively. We collapsed the three temporal dimensions into a single temporal variable, as the data source lacked post-event information; this made it difficult to determine which factors were in place before the event, as indicated by Runyan [26] and Bondy et al. [30].
Case selection
Two IHSA codes identify accident causes: Type Accident code (1–6), and Accident Minor IHSA Code (0–72). The latter code contains more detailed information on the accident event. Further, two variables, ‘major cause of injury’ and ‘minor cause of injury’, were used to identify electrocutions. Major Code 4 identifies electrocution, while Minor Codes 16 and 17 specify contact with electrical current and electrical flash/explosion, respectively.
Variables measured
Host (Worker): Age, occupation type, rate group, human factor (cause of injury)
Age: After mean age was calculated, age variables were grouped by 10-year age categories. There was no gender variable in the dataset. However, epidemiological studies on construction fatalities indicate that virtually all the construction fatalities were among males.
Occupation type: The IHSA code for type of occupation was coded in the dataset. Occupations were categorized as electrician and non-electrician for comparative purposes.
Rate group: Rate group was coded in the dataset. Workers with similar expected occupational risks and rates of injury and illness are placed in the same rate group. This study used rate group information as a proxy for job requirements of workers.
Human factor: The dataset contained coded variables indicating human factors that contributed to the injury.
Agent (electrical agent/equipment): Voltage, contact type and electrocution pattern
The voltage, contact type, intermediary vector type and electrocution type were quantified through a narrative text analysis conducted by two researchers experienced in work-related injury, with assistance from a master electrician.
Voltage: Thresholds for high voltage vary by country. For example, in Canada, voltages greater than 750 volts are considered high voltage according to Occupational Health and Safety regulations in Canada [31], whereas in the United States, 600 volts or greater is considered high voltage [32]. For the purposes of this study, we used a more conservative threshold of 600 volts to define high voltage.
Contact type: Direct contact with electrical source versus indirect contact by the intermediary of a vector (e.g., aluminum ladders), which allow electric current to flow to the body [33].
Electrocution pattern: These variables were developed through a narrative text analysis. For this analysis, categories of occupational electrical incidents identified from Cawley and Homce [13] were utilized as a reference.
Environment: Location of site, construction site, days of week, season
Location of site: Indoor/outdoor, above/below ground were coded in the dataset.
Construction site: Geographical variations were coded as remote, rural and urban. The remote and rural areas were collapsed into a single variable.
Days of week and Season: These temporal variables were analyzed descriptively to identify the days of the week and seasons when electricity-related occupational fatalities were highest.
Table 1 provides a summary of the variables measured.
Analyses
Descriptive and univariate analyses were performed. ANOVA and Chi-Square/Fisher’s exact tests were used to compare factors (see Table 1) by occupational groupings. In addition to quantitative analysis, we adopted a qualitative approach and performed a narrative text analysis to acquire more comprehensive information on the circumstances surrounding the accident event. This approach has been valuable in injury epidemiology and facilitates understanding of the accident event in detail [16, 34]. More specifically, Williamson et al. [35] stated that narrative analyses are a promising alternative for investigating mechanisms of injury and death. There are few published reports on fatality outcomes in the construction sector [30, 34] and occupational electrocution [13, 16].
Results
Background data
The dataset documented 200 cases of construction-related occupational fatality in Ontario between 1997–2007. Table 2 lists mechanism of injury, in addition to the mean age of workers at fatality. Work-related electrocutions comprised 15% of all construction fatalities.
Host (Worker) characteristics
Twenty-nine construction workers suffered electricity-related fatalities during our window of observation. The mean age of fatally injured workers was 35.7 years (range = 19–57). Electrocution occurred mostly in those aged 25 to 34 years, with approximately 60% of the electrocutions occurring among workers younger than 35 years of age. Within the construction industry, electricians were most often implicated with electricity-related fatality (n = 11; 37.9%); however, as a collective, non-electricians comprised more than 60% of electricity-related occupational fatalities in the construction industry. Accordingly, the rate group was also found to be greater for non-electrical services. The most hazardous human factors that contributed to electricity-related occupational fatality were unsafe action and failure to shut down power (n = 17, 58%). See Table 3 for a summary of host characteristics associated with electricity-related occupational fatality.
Agent characteristics
The majority of cases involved high voltage contact (n = 21; 72%); see Table 4. Among all cases, 81% involved contact with overhead power lines (n = 17). Among indirect contacts, the most dangerous intermediary vector was the aluminum ladder (n = 7; 41%), which was most often handled by younger non-electricians (aged 13–19), followed by boom and dump boxes. Table 5 provides case scenarios that exemplify various types of electricity-related fatalities.
Environmental characteristics
The majority of electrocutions occurred in urban areas and outdoors. While most workers were working for employers covered by the Ontario Workplace Safety Insurance and Board (WSIB), 17% of electrocution accidents were not covered under WSIB mandatory coverage; see Table 6. In terms of seasonal variations, electrocution occurred most frequently in autumn (35%), notably October. Amongst weekdays, the highest frequency of events occurred on Fridays (24%). However, analyses of fatalities by season and day of the week did not reach statistical significance (p = 0.324, p = 0.434, respectively).
Electrician vs. non-electrician
Non-electricians were younger than electricians, although this difference did not reach statistical significance (p = 0.362). Unsafe action and inattention to activity (n = 13, 72%) accounted for most fatalities among non-electricians, while failure to shut power down and lack of use of personal protective equipment were the leading cause of occupational-fatality in construction among electricians. It is also important to note that high voltage and indirect contact were significantly associated with fatalities amongst non-electricians. Furthermore, in most cases electricians were fatally injured while working alone, whereas non-electricians had a higher tendency to be injured when with co-workers; see Table 7.
Discussion
Our data show that approximately one in seven construction fatalities were due to electrical contact. Fatal electrocutions among electricians were significantly related to lower voltage, indirect contact and outdoor activities.
This study adopted a unique methodological approach by combining coded data analysis and narrative data analysis in order to provide more comprehensive information on electricity-related construction fatalities in construction. Through narrative text analysis, we were able to identify additional quantifiable variables not previously examined, such as level of voltage, contact type, electrocution patterns, type of vectors and composition of working team. These variables were important in order to understand the nature of electricity-related deaths in construction.
Rossignol and Pineault [33] attempted to classify fatal occupational electrocution across all industries in Quebec. They found two categories of workers at highest risk. First, they found that workers assigned to indoor electrical tasks were more likely to be electrocuted through direct contact with low voltage sources. Second, they report that workers assigned to outdoor non-electrical tasks were more likely to be electrocuted by an intermediary that was in contact with a high voltage source. Our findings similarly show differences in electrical-related fatality between workers assigned to indoor and outdoor tasks.
Our results showed that about 17% of fatal cases were not under WSIB insurance coverage, suggesting the need for more comprehensive health and safety coverage. Being disconnected from the WSIB could reflect a lack of electrical safety training, as has been reported by Lombardi et al. [16] as a risk factor for electrocution at the organization level.
Regarding risk factors for electricians, occupational fatality due to electricity was associated with indoor tasks, direct contact with low voltage and not wearing personal protective equipment (PPE). Non-electricians, however, were more likely to be electrocuted during outdoor tasks within non-electrical services and with indirect contact (involving aluminum ladders) with high voltage and less use of PPE.
A greater number of electrocution deaths occurred in the under 35 years of age group. This was similar to a National Institute for Occupational Health and Safety (NIOSH) [36] report on occupational electrocution death, where the age group of 20–24 years had the highest fatality rate.
Although electricity-related occupational fatalities in construction have been shown to peak in August, a summer month in the US [36], our data indicate a peak in fatalities/non-fatal injuries in autumn, specifically October. This difference may be attributable to pressures to finish work before the winter season when work decreases due to weather. Also, inclement weather and shorter work-days may also be contributing factors. Overall, this trend needs to be investigated further [37–39].
The present study only used selected variables which had the least amount of missing data. This study was restricted to these variables, meaning that other factors of interest require further exploration with a better quality data in the future.
Prevention of injuries in general, and fatalities in particular, continues to be a major challenge for both industry and regulators. According to National Census of Fatal Occupational Injuries in 2010 [40], construction compromises about 8% of U.S. workers, but accounts for 17% of the fatalities in the US workforce.
Conclusion
This study details electricity-related construction fatalities using coded data and quantified narrative data from the Ontario Ministry of Labour.
As indicated in Cawley and Homce’s study [13], there was no single ‘silver bullet’ method to prevent occupational electrocution and traumatic injury. Our findings suggest a broad range of approaches tailored to workers from a range of occupational categories are required. Further, there needs to be more research to address the temporal aspects of these injuries in the Canadian context.
Conflict of interest
The authors have no conflicts of interest to declare.
Footnotes
Acknowledgments
This study was funded by the Ontario Neurotrauma Foundation and the Ontario Ministry of Labour. Dr. Colantonio received funding from the Saunderson Family Chair, Toronto Rehabilitation Institute-UHN and a CIHR Research Chair in Gender Work and Health (#CGW-126580).