Abstract
BACKGROUND:
The paper is a research review focusing on occupational safety in the construction industry.
OBJECTIVE:
The purpose is to present research that highlights the areas of occupational safety and risks and to identify areas where research is lacking.
METHODS:
146 articles from scientific journals, mainly covering the construction industry in Europe, Canada, USA, Australia and Japan have been studied. The findings are presented under 11 categories: accident statistics; individual factors; legislation and regulations; ethical considerations; risk management; leadership, management, organization; competence; safety design; cost-benefit calculations; programs and models; and technical solutions.
RESULTS:
The research is dominated by initiatives from researchers and government authorities, while the construction industry only appears as the object for the research. There is a scarcity of research on integrated systems encompassing subcontractors, as well as a lack of research with sociological perspectives on accidents. Furthermore, only a few studies have applied a gender perspective on safety in construction, i.e. there is a need of further research in this particular area.
CONCLUSIONS:
A range of initiatives have been taken to increase safety in the construction industry and the initiatives are mainly reported to be successful. There are some cultural differences, but basically researchers present similar results regardless of country.
Background
The construction industry has over the last decades shown a positive trend in terms of occupational accidents. At the same time, we note that many workers are still injured at work. In comparison with other industrial activities, the sector has high accident rates. There are different views about the causes of these shortcomings, something that might impede the implementation of effective solutions.
The purpose of this research review is to present research that highlights the areas of occupational safety and risks in the construction industry and to identify areas where research is lacking.
Method and selection of articles
Our research review is based on scientifically significant literature, i.e. such articles that have undergone a review process in internationally recognized scientific journals. We chose to use two bibliographic databases, Web of Science and Scopus, for searchingand specific criteria were designed to capture articles that dealt with health and safety or accidents in the construction industry. This ultimately resulted in the following search key: (“Construction work” OR “Construction industr*”) AND (“health and safety” OR “safety and health” OR accident*) NOT (China OR Taiwan OR “Hong Kong” OR India OR Africa OR Singapore). The search field was limited to “Topic” in Web of Science and “Article or Review” in Scopus.
The material is limited to a Western context and consists mainly of articles from Europe, Canada, USA, Australia and Japan, and a few additional articles with different geographic origin. We concluded that the most relevant research in the area is written after 2000, and that the main findings of earlier research are incorporated into later published material. A test showed that we reached 80 percent of published material, despite the limitation to the year 2000 and later.
The result of the literature search consisted of 737 unique articles. Out of these, 411 articles were excluded due to reasons like geographical areas outside the ones selected for the study; not about health and safety or accidents; not about the building and construction industry as well as language other than English. The remaining 326 articles from the literature search were subject for further reading, and were eventually reduced to 146.
In a first step the articles were categorized based on their main approach and content. This resulted in 11 categories as listed below (23 of the articles were sorted under two categories): Accident statistics - articles that mainly have their focus on different aspects of accident statistics and in different ways compare development over time or between different categories analysing the causes and possible suggestions for improvements. (72 articles) Individual factors - articles that specifically examines person-related factors in accidents, such as age, experience, gender, language, psychosocial aspects, etc. (25 articles) Legislation and regulations - articles that examines the impact of law and regulations on health and safety in the construction industry. (24 articles) Ethical considerations - articles that mainly have their focus on the construction industry’s social responsibility, such as CSR. (8 articles) Risk management - articles that focus on risk management in the construction industry, such as procedures, checklists, risk assessment, etc. (40 articles) Leadership, management, organization - articles where leadership and management systems is examined in relation to accidents and safety. (47 articles) Competence - articles where educational activities, training tools and skills development initiatives are evaluated and analysed. (19 articles) Safety design - articles that in various ways are focusing on the design process before the project starts. (17 articles) Cost-benefit calculations - articles that focus on different aspects of costs in relations to health and safety. (15 articles) Programs and models - articles that examines and evaluates the consequences of major safety programs often initiated by government agencies and industry associations. There are also articles with focus on models, tools or other measures for increased safety. (47 articles) Technical solutions - articles of varying character with its focus on technology and technical solutions for increased safety developed within, or applied to, the construction industry. (21 articles)
∥In our analysis, the eleven categories are further extracted into four larger categories in order to describe the research field from a wider perspective and also identify gaps where more research is needed. These four categories are; accident statistics, safety work- and prevention, technical solutions, and programs, models and methods.
Results
In the following section, we summarize the results of the literature review. The presentation follows the categorization in 11 groups presented above.
Accident statistics
The major part of the high-quality research comes from England [1], USA [2–9], Australia [10] and to some extent Spain [11–14].
Almost all the research on accidents use statistics in one way or another, at least to show that the research is justified (e.g. accident frequencies). There are a number of large general quantitative studies on accidents in different countries that are basically designed in similar ways, but any comparison of statistics from different countries must be made with caution because there are no uniform methods for the collection and classification of data [1]. With this reservation in mind, we can see that: The accident rate is high in the construction industry, regardless of the country being studied. Mining and agriculture are other industries with high accident rates. Fall accidents peak the statistics for serious accidents. Minor accidents without absence from work has other reasons (cuts or damage to the lumbar spine, eyes or shoulder/arms/hands) compared to fatalities, which usually are caused by falls from height, electric shocks, crushing, or that the worker is hit by an object. Generally, the studies show a decrease in the number of fatalities.
The question of where, when and how is central to almost all research based on accident statistics. The risk for accidents with serious consequences is higher in small companies [11, 15]. The risk for accidents is higher among subcontractors [16]. Slips and trips accidents constitute a large part of the number of accidents [17]. Environmental conditions like dust, snow, ice, mud and weather are identified as important contributing factors for accidents [7, 11].
We can see some similarities and differences between countries when it comes to accident occurrence during the course of a work day. In Sweden most accidents occur between 10 a.m. and 12 p.m., which is roughly the same as in the Spanish construction industry (between 10 a.m. and 11 a.m.). The probability of an accident having serious consequences, however, is greater after the lunch break in Spain, which takes place between 2 p.m. and 3 p.m. This may be related to what workers consume at these occasions (e.g. alcohol), [in cases of a genuine siesta or rest break being observed, it could also be drowsiness due to nascent awakeness] although there is a need of further studies on the underlying problems [11].
Individual factors
A number of studies are based on individual factors, i.e. the link between individual characteristics and the risk of accidents, alternatively the accident rate. Gender, age and language skills are the most common factors discussed [10, 18–25].
A recurring result is that men are at greater risk of injury than women [10, 26–29] but all of these studies neglect to take into account what type of tasks women and men performed. Several studies show that women have administrative tasks to a much higher degree than men [26–28]. We also note that there is no real gender analysis in the sense of gender as a social construction in the construction industry.
The impact of age varies slightly between studies. In most countries, the younger age groups are over-represented in the general accident statistics [10, 30]. However, there is a consensus that the severity of accidents increases with age [10, 30]. The increased experience that comes with age cannot fully compensate for the age factor’s impact.
Workers with poor language skills tend to run into more and more serious accidents than others [21]. There are also studies showing that psychosocial factors [31] and problems such as lack of sleep [32] contribute to fatigue, which in turn can have significant effects on individuals’ well-being, performance and safety [33].
Legislation and regulations
As a result of the EU’s Directive 92/57/EEC - temporary or mobile construction sites there are a number of studies that illustrate how the directive has been implemented in different countries and its effects [34–39].
Martínez-Aires et al. [37] shows in a study that in 2005, i.e. between 5 and 10 years after the introduction of the EU Directive, the accident rate had dropped in 11 of the 15 countries studied. Greece and Belgium showed the most marked reductions with 65,7 percent and 42 percent respectively. In a follow-up case study, Martínez-Aires et al. [39] explores how the directive was introduced in two specific countries, Spain and the United Kingdom, and how this interacts with the established practices in “Prevention through Design” (PTD). The results show that PTD works better in the British context, since the regulatory demands for coordination already at the planning and design stag are clearer than in the Spanish.
Oversight and inspection is problematic in a sector with temporary workplaces that constantly change. According to Weil [40], US authorities have traditionally focused a large part of its inspection resources on larger construction companies, which meant that the really problematic areas - usually the work carried out by smaller subcontractors in smaller construction projects - ended up in the background. One conclusion from the study is that the government authorities should reallocate its resources to some extent, to also cover these more problematic sectors of the industry. Similar conclusions are also drawn in both a Danish [15] and an Australian [41] study.
Ethical considerations
An interesting proposal is to incorporate safety issues as part of a company’s corporate social responsibility (CSR) program [42, 43]. Here, the work is at the very beginning and may require, among others things, a mapping of ethical considerations vis-á-vis different stakeholders [44]. It is important to include the developer, the general contractors and the general project planner to achieve a well-functioning safety management system [45].
Risk management
There is a long tradition of incident analysis in operations where the consequences could be large, such as aviation and the nuclear industry. In the construction industry, we cannot find the same interest, although several studies show the benefits that come with having a well-developed system of incident reporting and risk analysis. A well-functioning system requires procedures that are integrated into the safety culture and for the results to be distributed throughout the organization in question there must be a clear division of roles and responsibilities [46]. Probst et al. [47] show in a major study that in companies with a poor safety culture it was more common with underreporting of injuries with as much as 81 percent of the injuries not being reported. In companies with a more positive safety culture, the figure for underreporting was only 47 percent.
There is much research focusing on models and methods for risk management, risk analysis and risk assessment [48–52]. The studies usually have an applied nature where different models are tested in existing construction projects and then evaluated by health and safety experts. Risk assessments are often based on simple tools, like checklists, which can be practical to use but at the same time could lead to a limited and insufficient basis for deciding on measures [53]. Fung et al. [54] stresses that the responsible actors in the construction industry tend to rely on previous experience when assessing risks, leading to subjective assessments that not always capture the whole problem. Risk perception may also vary depending on what role the person has in the construction process [55]. In order to avoid this problem, there are new risk assessment models based on the mathematical theory, known as “fuzzy sets” techniques [56]. Another common approach is to focus on risk assessments already in the early stages of construction projects, i.e. the planning and design phase [51, 59].
A seemingly controversial issue is whether the safety measures can be prioritized based on their economic outcomes. Sousa et al. [60] have developed a quantitative risk assessment model, “Occupational Safety and Health Potential Risk Model” (OSH-PRM), with a cost-benefit analysis incorporated in the risk assessment. The authors underline that this is a sensitive subject because health and safety should not be reduced to only their economic outcomes. Due to companies oftentimes having limited resources, however, it is important to include these dimensions as a way to improve the management and allocation of available resources.
Leadership, management, organization
Extensive research shows that management’s motivation and commitment to safety issues are important when it comes to achieving safe workplaces [61–68]. Törner and Pousette [69] argues that the key to a high level of safety is to develop relationships based on trust, both between individuals and between different functions on different levels of the organization. Törner and Pousette list four interrelated aspects to consider: the nature of the construction project, organization and structure, values, norms and behaviours, individual attitudes, knowledge and skills.
Knowledge and motivation with regard to safety are essential prerequisites for personal safety behaviour, i.e. specific behaviours aimed at personal protection through the use of safe work equipment and choosing safe working methods [70]. Safety behaviour in general has also been shown to be causally related to safety climate [71, 72]. Safety climate is, however, not necessarily uniform in a company and may vary between different working teams. Overall, managers prioritizing safety is an important aspect of safety climate [73] as well as safety culture [74, 75].
Competence
Education and various forms of training in occupational health and safety is often put forward as key measures to prevent illness and accidents [76–78], in the sense that it leads to better understanding of the individual’s own personal responsibility for safety [9, 74]. One problem may be the lack of motivation for education, partly because of a lack of adaptation to the individual and personal needs of the workers [79]. In response, new types of technologies have been developed where both employees and managers can train their skills in virtual environments. Numerous studies use virtual reality (VR) environments where participants have the opportunity to practice different work situations but also practice on identifying risks [80–83]. The VR-environments offer a high degree of interactivity, in which both teachers and colleagues can learn together. One example is Zhao and Lucas [84], which describes the development of a VR-based training program for electrical safety. Other ways to make use of the new computer technology are explored by Li et al. [85] that describe and evaluate a data processing system in which knowledge and experience from incidents and accidents can be entered, distributed and reused for learning. The system is described as a social network and can be compared with something like a wiki, i.e. a kind of knowledge bank that takes advantage of the common knowledge and experience. Esmi and Ennals [86] describes the problems of a more mobile and flexible workforce where skills are often in short supply, and point to solutions like Knowledge Management (KM) which are based on systems where knowledge and experiences can be spread through computers. However, Flynn and Sampson [45] argues that knowledge about safety is not enough and that other factors such as organization, group dynamics, culture and climate also play important roles when it comes to safety performance and awareness.
Only a few articles highlight the importance of integrating safety programs in the formal education system. There are some examples, such as Le et al. [83] that describe and evaluate a virtual learning environment intended to be used in upper secondary school, and Petersen et al. [87] focusing on a safety training programme for prospective construction engineers. In a Finnish survey study, Kaskutas et al. [88, 89] evaluate an apprentices training for avoiding fall from heights. The results show that this training came in too late and many apprentices had already been exposed to this risk before they received the knowledge of how the risk should be managed.
Safety design
Safety design is an area that is receiving increasing attention in research [1, 91–98]. Simply put, safety design is about incorporating safety considerations already at the planning and design phase of a given building project. Research shows that it is possible to derive workplace accidents from how building components are designed and constructed [57, 100]. One problem has been that the norms and standards that exist around how constructions and components should be designed have mainly revolved around the final user’s health and safety and have thus neglected the work environment aspects during the building stage. The researchers in the field agree that the professionals involved in the early planning and design phases have to take a greater responsibility for health and safety on the construction sites.
Cost-benefit calculations
Many studies address the issue of whether or not preventive safety work creates profitability and all show some sort of positive effect [51, 101–108]. Hallowell’s study [102] stands out in this regard and shows that the benefits in the studied case exceed the costs with a ratio of 3:1. Several studies show that bonus schemes that reward avoiding accidents has a positive impact in the short term, but the effect diminishes with time and that programs consequently need to be continually updated to have an effect [109–111]. Taking a different perspective, Musonda and Pretorius [112] focuses on the impact economic incentives may have on the developer’s work with health and safety. The results show that it makes it more likely that different proactive safety measures will be implemented at the construction site.
Programs and models
The efforts construction industries place on accident prevention is often manifested in more comprehensive programmes for health and safety [113–116]. The initiatives for these programmes are often taken by government authorities and industry trade associations. Common for these programs is that they usually contain a number of activities, involving people from several levels of the organization and that they last for a long time. Rarely are they linked to various sanctions programs. Research on these programs and models is essentially about evaluations; usually what effect they have given to the accidental frequency or health insurance outcomes.
The criticism that researchers have directed towards these types of programs is that they are not generally theory-based and that they hardly give any assurance about the results. Methodological problems involve both risk exposure level and the lack of control groups, which implies difficulties in deciding the overall significance of these types of safety interventions [117, 115].
Some success factors reported: Programs and models for safety should last for a long time, otherwise you risk “resetting” or “restoring” effects, i.e. to return to previous positions with regard to safety [116]. Programs that provide technologies for changed behaviour seems to be more successful than programs that merely provide information and training [116]. Programs that contain various forms of rewards or punishments have greater effect on accident statistics than those that do not contain these measures [116, 118]. Programs that are used by the whole company are more stable and persistent over time, and even safety-unaware companies can improve their safety work [114, 119]. It is important that management is safety-oriented and that subcontractors are involved in the programs [75, 118–120]. Programs that contain a combination of several strategies are more successful than those that only have a few [120]. In order for the programs to be successful, it is important that the authorities, unions and industry collaborate on these issues [121–123]. A positive social control within the industry sector is important since there often lies a certain status among participating companies to be a part of a program [122, 124]. It is important that a given programs content, purpose and idea is understood by everyone involved [115]. In studies performed in the US in particular, anti-drug programs are highlighted as a successful preventive measure [125–127].
Evaluation of programs aimed at small and medium-sized enterprises point to the difficulty for them to develop adequate safety work and implement the appropriate measures [128]. The employer who is responsible for the work environment lacks experience and training in the field. If you do not have any experience of accidents, it is difficult to work preventively to ensure health and safety, and the employees’ close relationship with the employer may hamper opportunities for giving constructive criticism [129].
Technical solutions
Research on new technical solutions to increase safety in construction work mainly focuses on various forms of computer tools and primarily real-time based ones [85, 130–134]. The articles deal mainly with two different types of real-time systems where machines and personnel are equipped with so-called RFI-tags. In the first case, it is about creating systems that in real time can provide information about different risk situations that may arise, for example, in the form of collisions of machines, materials and people [135, 136]. The main task of these systems is to improve communication between the different actors in the workplace and warn for the risks that may arise. In the second case it is about sensors being placed on staff to monitor the individual’s work situation, e.g. exposure to hazardous substances or radiation, but it is also possible to monitor physical overload and poor working postures [137].
Augmented Reality (AR) is another area in the research on technical solutions to improve the safety work. AR provides tools for safety analysis and design based on a combination of virtual and real environment [58].
A third area is more advanced simulation of the workplace to proactively identify various possible risk factors and dangerous situations. A simple example is to simulate the cranes movement patterns; a more complicated one is to simulate how a work task should be performed in narrow spaces [138, 139].
Overall, technical solutions should always be developed in accordance with the perspectives and needs of the workers. This could include, for example, individual characteristics such as age [140], experience [141] as well as the ability for safety responses [142]. Furthermore, the technology should also be developed and adapted to comply with formal rules and regulations of different kinds [143].
Discussion
This review consists of 146 articles concerning occupational safety in the construction industry from a broad perspective. Initially the literature was categorized into eleven descriptive categories showing the width of areas of focus in research on risk and safety without going too much into detail. In the following analysis, the content of the articles was extracted from the eleven categories into four larger categories in order to describe the research field from a wide perspective and also identify gaps where more research is needed (see Fig. 1). It is difficult to draw a sharp line between the different categories; rather they complement each other in more complex patterns. But for the purpose of providing a broader description of the research field the chosen categories represent different areas of focus in research.

Factors affecting accidents and safety performance.
Research concerning risks and research on safety usually have the accident in center as its starting point. Starting from the accident the four main categories identified are; accident statistics, safety work- and prevention, technical solutions, and programs, models and methods. These categories are both directly and indirectly connected to the initiatives originating from research, from the industry itself and from other stakeholders such as legislators and unions.
In
In intervention studies researchers implement their
Our immediate observation is that a range of initiatives are taken to create a safer work environment in the construction industry and that the initiatives mainly are reported to be successful. The results are not so dependent of the context. There are some cultural differences, but basically researchers present similar results regardless of country.
This literature review shows that the work to minimize risks and increasing safety is a complex matter. The four-field model shows that preventive initiatives have to be taken from many angles by both the industry itself but also government authorities, industry trade organizations and researchers. In the construction industry the way work is organized with temporary work sites further complicates the work with safety. Most of the research reviewed here is quantitatively oriented, which is not surprising since the evaluations usually are expressed in different types of accident frequencies. What is needed is more qualitative oriented studies. There is also no universal solution ready for implementation. What emerges, rather, is the complexity of the problem and the need for more integrated solutions. Integrated actions including accident statistics, safety work- and prevention and technological solutions offer new possibilities for increasing safety in the construction industry.
An important question is in which fields there is a lack of research. There is a scarcity of research on integrated systems encompassing subcontractors in the construction industry. Few studies have, furthermore, applied a gender perspective on the issue of safety, which may be a consequence of the low number of women employed in the construction industry overall. There is also a lack of research on wage dumping with the help of undeclared labour, and how it affects health and safety. The question of undeclared work has been addressed on both national and wider European levels and has been defined as: “Any paid activities that are lawful as regards their nature, but not declared to public authorities” [144]. This means work that can be labelled as “black”, such as not paying taxes or social fees, and work that can be considered as “grey”, for example subcontractors that work in an employment-like situation without actually being employed. A concrete issue in these circumstances is that these workers run the risk of not receiving necessary safety education and other benefits often afforded individuals under more secure employment conditions.
Finally, focus should more succinctly be placed on issues related to multi-employer worksites and inter-organizational complexity and the consequences of this for safety. A number of recent studies [e.g. 145, 146] have highlighted that those involved in multi-employer arrangements, regardless of industry sector, share common challenges regarding, for example, breakdown in communication and ambiguity in the division of safety responsibilities. Included in this is the matter of hierarchically structured subcontracting chains and how safety is practiced at different levels, as well as how safety-related communication and information sharing is managed across organizational boundaries. Given that construction projects often involve small- to medium-sized companies, some of which may lack the necessary resources or know-how to prioritize safety, smaller construction firms active as subcontractors deserve particular attention in focused case studies. These companies may, for example, have difficulties in adopting technological solutions that fundamentally require financial resources and opportunities for research and development. Moreover, subcontractors may be especially challenging to include in wider safety programs initiated by government authorities and industry trade organizations given the temporary and transitory nature of construction projects. A focus on these issues is paramount since (paraphrasing the old proverb) a subcontracting chain is only as strong as its weakest link.
Conflict of interest
None to report.
