Michael Behm Acknowledgements The author wishes to thank Dr. Anthony Veltri and Dr. John Gambatese of Oregon State University. This research study would not have been possible without their support. Abbreviations/Terms
Contents Background Research Methods Results and Discussion Conclusions Recommendations References Figure 1. Design - construction incident investigation model Tables 1. Selection of OSHA accident inspection reports 2. Incidents linked to designing for safety 3. Incidents classified by type of construction project 4. Incidents classified by design element under construction Annexes A. Examples of How the Investigation Model Was Applied B. Existing Design Suggestions Linked to Incidents C. New Design Suggestions Linked to Incidents The author analyzed 450 reports of construction workers’ deaths and disabling injuries to determine whether addressing safety in the project designs could have prevented the incidents. The reports were obtained through the Occupational Safety and Health Administration (OSHA) and the National Institute for Occupational Safety and Health (NIOSH). Using an investigation model developed for this research, the author found that in 151 cases (about one-third of those studied), the hazard that contributed to the incident could have been eliminated or reduced if design-for-safety measures had been implemented. The author aims to demonstrate the value of incorporating design considerations into a systems approach to improving safety on construction sites. Background Designing for safety is defined as the consideration of construction site safety in the preparation of plans and specifications for construction projects. The process often involves specifying permanent features of a project to address construction worker safety. For instance, parapet walls designed to be at least 42 inches high act as a guardrail and can provide fall protection to construction workers (Gambatese 2005; Trethewy and Atkinson 2003). Designing for safety also encompasses communicating about the safety hazards at the construction work site, for instance, noting on contract drawings the location of existing overhead power lines. The concept of designing for safety is consistent with the traditional “hierarchy of controls” approach used by safety professionals. This hierarchy calls for eliminating or minimizing a workplace hazard before relying on personal protective equipment or administrative or temporary controls to protect workers (Manuele 1997). Although construction safety professionals view the design-for-safety concept as a viable means of protecting workers, architects and design engineers (together referred to as “design professionals” in this report) are reluctant to adopt this intervention as part of their standard practice (Gambatese 2005). Design professionals lack motivating forces – legal, contractual, or regulatory – to adopt design-forsafety methods. These professionals may also avoid addressing worker safety out of fear that doing so will open them to lawsuits by an injured construction worker (Gambatese 1998; Coble 1997). Moreover, design professionals’ codes of ethics, such as the code established by the American Institute of Architects (2004), set ethical priorities for ensuring final occupant safety and safety of the finished product, but do not address the safety of the workers performing construction. Regulatory and contractual requirements place the primary responsibility for construction site safety on the constructor. (In this report, “constructor” means the construction firms, contractors, and subcontractors responsible for building a project and employing the construction workers.) For instance, the federal OSHA regulations place the responsibility for worker safety on the constructor as the primary employer. Project owners who make safety a priority also place the responsibility for construction site safety directly on the constructor, by showing preference for pre-qualified contractors who have good safety records, lower insurance rates, and comprehensive safety programs. Research into the root causes of construction accidents has also focused on the role of the constructor. Abdelhamid and Everett (2000) evaluated construction accidents in the United Michael Behm 2 States and developed a model for tracing the root causes of accidents. Their research addressed activities and conditions at the construction site but did not consider potential root causes in the project concept and design phases. The authors attributed unsafe conditions to four main causes: management action/inaction, unsafe acts of workers and co-workers, events not directly humanrelated (such as equipment failure and natural disasters), and unsafe conditions that are a natural part of the construction site (such as uneven terrain and concealed ditches). Abdelhamid and Everett’s approach is consistent with conventional accident root-cause analysis, focusing solely on the actions and inactions of the constructor, rather than adopting a broader view of accident causality that looks at upstream influences, including the design process. One recent study of causal factors in construction accidents looked at the designer’s role. Haslam and others (2003) studied the causes of 100 construction accidents in the United Kingdom, and found that permanent works designers (synonymous with “design professionals” in the United States) could have reduced the risk associated with the accidents in almost half of the cases. The authors also developed a construction accident causality model that described immediate causes, shaping factors, and originating influences in construction accidents. They concluded that the permanent works design influences the workers’ activities, the site, and the materials and equipment specified for construction. Research Methods The author obtained and analyzed construction accident investigation reports from OSHA State program offices in California, Washington, and Oregon, and from the NIOSH Fatality Assessment Control and Evaluation (FACE) program. These OSHA offices were selected in part because of their proximity to the author at Oregon State University. Also, the reports from these OSHA offices and the FACE data were publicly available at low or no cost. Statistical methods were used to randomly select reports and to analyze the findings. But this study does not claim to be an accurate statistical sampling of all the available data on construction injuries and deaths. The author initially conducted a pilot review of 25 Oregon OSHA construction accident inspection reports and determined that enough information was available in these reports to link the accident to the design-for-safety concept. For instance, the reports contained detailed notes about the work site hazards and conditions that contributed the incident. A similar review of the FACE data showed that these reports also contained enough information to determine whether there was an association between designing for safety and the construction accidents. Selection of OSHA Inspection Reports Oregon and Washington maintain their OSHA inspection reports in a central office, and reports from all districts in these states were included in the analysis. In California, inspection records are maintained individually by each of the 22 California OSHA (Cal/OSHA) field offices in the district where the inspections occurred. Given time and funding limitations, the author chose to obtain records from four Cal/OSHA field offices located near one another (in Torrance, Los Angeles, Anaheim, and Van Nuys). The following criteria were used to conduct a database search of the OSHA inspection reports from the field offices:
Following from this formula, at least 48 cases would be needed from each of the six OSHA offices (286/6 = 48). Fifty cases from each office (except for Torrance, California) were randomly selected from among the cases meeting the search criteria listed above, using the random case selection function of the Statistical Package for the Social Sciences (SPSS). The Torrance office had only 35 case reports meeting the search criteria and all 35 were reviewed. In all, 226 OSHA reports were chosen for the study (see table 1). Cases were rejected because either they lacked information, the accident was deemed not work-related by the OSHA inspector (for instance, death resulting from heart attack), or the actual file was not available. Of the 226 incidents analyzed, 37 had resulted in deaths and 189 had resulted in disabling injuries to workers. Table 1. Selection of OSHA accident inspection reports
Selection of FACE Data The FACE program studies deaths resulting from occupational injuries. The program’s goal is to prevent work-related deaths by investigating work situations posing a high injury risk to workers and then formulating and disseminating guidance on prevention strategies (NIOSH 2003). The FACE program selects construction deaths for study based on direct requests from OSHA as well as on FACE’s special-emphasis programs. For instance, recent FACE investigations in the construction industry have targeted Hispanic workers, young workers, steel erectors, and roofers (Virgil Casini, NIOSH, personal communication, Oct. 23, 2003). The complete investigation reports for all industries, including construction, are publicly available on the NIOSH website (http://www.cdc.gov/niosh/face/). In all, 224 randomly selected FACE reports, obtained from a database of 450 construction accident investigations, were selected for the analysis. FACE reports from 1990 on were used for this study because they contained more details than earlier reports. Each report evaluated in this study was unique and the author verified that there was no overlap between the incidents described in the OSHA inspection reports and those evaluated by the FACE program. Investigation Model To evaluate the reports selected for review, the author developed a model that aims to provide an objective analysis of possible links between construction incidents and designing for safety (figure 1). In creating the model the author used design-for-safety suggestions developed by Gambatese (1996), and by Gambatese, Hinze, and Haas (1997). The model addressed three questions:
Figure 1. Design - construction incident investigation model Here are two examples of how the author applied the investigation model shown in figure 1:
Hypothesis Testing After applying the model and identifying incidents linked to the design-for-safety concept, the author conducted hypothesis testing to determine how the design process or other factors might have affected the incident. For instance, the research sought to determine whether design-linked incidents were related to a specific type of project (residential, commercial, engineering, industrial) or design element under construction at the time of the incident (such as electrical, masonry, and thermal/moisture protection). The aim of this analysis was to identify the designrelated efforts that would be most effective in improving worker safety. Results and Discussion Table 2 summarizes the results of applying the investigation model (figure 1) to the OSHA and FACE reports. The research findings linking the incidents to specific design suggestions are presented in annexes B and C. Detailed results of the statistical analysis can be obtained by contacting the author at: behmm@mail.ecu.edu. Also, Behm (2005) contains a more detailed description of the results summarized here. Table 2: Incidents linked to designing for safety
As table 2 shows, the author linked 48 (21.2%) of the 226 OSHA-investigated incidents to the design-for-safety concept and categorized 9 (4.0%) as “maybe” linked. A greater percentage of FACE cases were found linked to the design: 88 (39.3%) of the 224 incidents were linked and an additional 6 (2.7%) were designated as “maybe” linked to the design. The author attributes these differences between the OSHA and FACE results mainly to the amount of relevant information found in the reports, which in turn reflects differences in these organizations’ functions. OSHA’s focus in conducting inspections is to determine compliance with regulations and the inspection reports tend to focus on conditions associated with compliance. One of the main goals of the FACE program is to develop injury prevention strategies. The FACE reports typically contained more detailed information on causal factors, and sometimes included mention of the design process, information that was particularly relevant to this study. Among the OSHA reports, more deaths (15 of 37) than disabling injuries (42 of 189) were linked to the design-for-safety concept, a result that could be attributed at least partly to the greater level of detail about work site conditions and overall information in the inspection reports of workers’ deaths. In nearly half (70) of the total 151 incidents found to be linked (includes “maybe” linked cases) with designing for safety, the deaths and disabling injuries were due to falls. In these cases, the author determined that the incident may have been prevented by implementing permanent fall protection such as the following measure: design special attachments or holes in members at elevated work areas to provide permanent, stable connections for supports, lifelines, guardrails, and scaffolding (Gambatese 1996; Gambatese, Hinze, and Haas 1997). Electrical hazards were a factor in 18 incidents linked to the design (an incident may be linked to more than one design suggestion). In these cases, the author determined that implementing the following design suggestion may have reduced or eliminated the hazard: disconnect, reduce voltage, or re-route power lines around the project before it begins. In 15 cases, the author found that the following measure could have prevented the incident: locate on contract drawings the existence of overhead power lines and their location in relation to the new structure. Additional study findings are summarized below. New design modifications and suggestions: Thirty new design-for-safety suggestions have been developed as a result of the analysis (summarized in annex C). For instance, one new suggestion calls for designing fall protection anchor points on storage tank interiors to be used during tank entry for construction and maintenance purposes. Type of construction project: 210 OSHA reports and 215 FACE reports contained sufficient information to be classified according to type of construction project (see table 3 below). Of these, the author linked 52 OSHA-investigated incidents and 92 FACE cases to the design. Analysis of the OSHA data showed that design-linked construction deaths and disabling injuries were not significantly related to a single type of construction project (p = 0.059, Cramer’s V = 0.188). Analysis of the FACE data also found no significant association between design-linked construction deaths and a single type of construction project (p = 0.214, Cramer’s V = 0.144). These results suggest that the design-for-safety concept can be applied in all types of construction projects. Table 3. Incidents classified by type of construction project
Design element: As table 4 shows, 208 OSHA reports and all 224 FACE reports contained sufficient information to be categorized according to design element being constructed at the time of the incident. Of these, 151 cases were found to be linked to the design-for-safety concept. Among the OSHA cases, the analysis found that 57 (27.4%) design-linked construction deaths and disabling injuries were related to the design element being constructed at the time of the accident (p = 0.001, Cramer’s V = 0.33). The category of thermal/moisture protection and doors and windows contained significantly more design-linked cases than expected. The author attributed this finding to the relatively large number of deaths and disabling injuries (17) caused by falls from and through roofs and through skylights, where anchorage points, guardrails, and other forms of fall protection, if designed into the permanent structure and used by the constructor, would have prevented the incident. Table 4. Incidents classified by design element under construction
The analysis of the FACE data also found that the design-linked construction deaths (94 of 224) were related to the design element being constructed at the time of the accident (p = 0.001, Cramer’s V = 0.33). The metals category contributed highly to the significant result. The author attributed this finding to the relatively large number of deaths (20) caused by falls from structural steel framing and buildings, where permanent anchor points, lifeline systems, and other forms of permanent fall protection could have been designed into the permanent features of the constructed building. Size of firm: Analysis of the FACE reports showed that design-linked construction deaths were not significantly related to the affected company’s size (based on number of employees) or Standard Industrial Classification (SIC). The OSHA-investigated incidents do not report this information consistently and were not included in this analysis. Conclusions
Follow-up research related to the current project should include a second reviewer to evaluate reports, in order to increase the reliability and validity of the results. The author also suggests eliminating “maybe” as a potential response and using only a “yes”/“no” option when applying the model. A Delphi panel (panel of experts) consisting of construction industry professionals could be established to examine the validity of the responses and to determine the feasibility of implementing each linked design suggestion. Suggestions for future research on designing for safety are as follows:
Abdelhamid, T., and J. Everett. 2000. Identifying Root Causes of Construction Accidents. Journal of Construction Engineering and Management, 126(1): 52-60. American Institute of Architects. 2004. 2004 Code of Ethics and Professional Conduct. Available: http://www.aia.org/SiteObjects/files/codeofethics.pdf Behm, M. 2005. Linking Construction Fatalities to the Design for Construction Safety Concept. Safety Science, 43(2005):589-611. Coble, R. 1997. Knowing Your Role in Construction Safety to Avoid Litigation. Journal of the American Institute of Constructors, 21(3): 25-28. Cohen, J. 1988. Statistical Power for the Behavioral Sciences (Second ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Gambatese, J.A. 1996. Addressing Construction Worker Safety in the Project Design. Unpublished Doctor of Philosophy Dissertation, University of Washington.
Haslam, R., S. Hide, A. Gibb, D. Gyi, S. Atkinson, T. Pavitt, R. Duff, and A. Suraji. 2003. Causal Factors in Construction Accidents (Research Report RR 156). Health and Safety Executive (United Kingdom). http://www.hse.gov.uk/research/rrpdf/rr156.pdf HHS (U.S. Department of Health and Human Services). 2000. Healthy People 2010: Understanding and Improving Health. Washington, D.C.: U.S. Government Printing Office. Hinze, J.W. 2001. Construction Contracts. New York: McGraw Hill Companies. Manuele, F.A. 1997. On the Practice of Safety. New York: John Wiley and Sons, Inc. NIOSH (National Institute for Occupational Safety and Health). 2003. NIOSH Fatality Assessment and Control Evaluation Program, NIOSH Publication No. 2003-146 (PB 2003-106-834).
Annex A. Examples of How the Investigation Model Was Applied This section presents examples of how the author applied the investigation model (shown in figure 1 of the report) in deciding whether the incidents described in the OSHA and FACE reports were linked to the design-for-construction-safety concept. Structural Failure (Question 1) In the example that follows, question 1 was answered “yes” because the structure itself failed during construction activities and its collapse could be linked to the design process. An employee was working in a crane next to a brick wall, removing roof sections attached to the brick wall. The brick wall was not designed to be free-standing. When the roof section was removed, the brick wall collapsed onto the crane and crushed the worker. This incident review led the author to propose the following new design-for-construction-safety modification: before demolishing and renovating any structure, ensure that an engineering survey is performed by a competent person to determine the condition of the structure, evaluate the possibility of unplanned collapse, and plan for potential hazards. Question 1 was answered “yes” in the following incident review because the physical aspects of the project prohibited the constructor from implementing effective temporary safety measures. Two employees were walking on the roof of an industrial building to access electrical equipment needing repair when they fell through the roof. The roof had contained many corrugated fiberglass panels (non-load bearing) that were indistinguishable from panels designed as a walking surface and the employees fell through these panels. This design prohibited the constructor from recognizing a hazard and from implementing a temporary safety measure or utilizing a different technique to perform the work. The author concluded that a designer using a design-for-safety approach would have taken into account the construction and maintenance work on the roof and would have provided for an appropriate walking surface for such activities. Question 1 was answered “no” in this example: The FACE program investigation team determined that one of the main causes of a structural collapse leading to the workers’ deaths was the constructor’s failure to adhere to design specifications. The incident occurred while iron workers were fabricating a steel skeleton for an automotive repair shop. The building plans and procedures specified that ¾-inch steel sway bracing rods be installed and kept in place immediately after hoisting the beams into place for the purposes of maintaining structural stability. The author determined that this incident was not directly design-related; rather the constructor failed to comply with the design specification. This case highlights the critical importance of the constructor’s role in implementing designing-for-safety specifications. Existing Design Suggestions (Question 2) The following example demonstrates how a previously developed design suggestion (see complete list in annex B) was linked to a construction incident. An employee laying insulation on a flat roof stepped backwards and fell off the roof edge 30 feet to the ground. The author linked the following two design suggestions to this incident:
The author developed 30 new design suggestions based on the research; these are summarized in annex C. The section below discusses the development of this new suggestion: design anchor points on the interior of the tank for construction and maintenance purposes. The author developed this suggestion after extracting the following information from a FACE report (90-16): A small construction company, which had been contracted to sandblast and paint the interior of a tank, planned to use suspension scaffolding to reach all sections of the 48-foothigh and 30-foot-diameter tank. Before the suspension scaffold was raised into position, the victim used a ladder to weld steel brackets to the walls at the top of the tank. The brackets were intended to anchor a horizontal steel cable serving as a fall protection anchor cable. The nylon suspension ropes, which were used in the suspension scaffold system, were lying on the floor of the tank while the brackets were being welded. After the welding, the owner inspected the suspension ropes but did not notice any damage; however, the suspension rope broke when the employee pulled it to move the scaffold higher. The scaffold failed and the employee fell 40 feet to the bottom of the tank. The FACE study found numerous causes for this accident. One cause was the employee’s failure to follow the OSHA regulation prohibiting welding, burning, riveting, or open flame work on staging suspended by means of fiber or synthetic rope [29 CFR 1926.451(a)(18)]. An OSHA investigation after the incident determined that the rope had broken at a point where it had been burned, probably because the rope was on the ground below where the worker had been welding the anchor points. However, the author also determined that another factor could be linked to the design-for-safety concept: If there had already been anchor points designed-in to the tank’s interior, the employee would not have been required to field-fabricate the anchor points, thus eliminating one of the main hazards associated with the incident. The FACE review states:
Annex B. Existing Design Suggestions Linked to Incidents The design suggestions presented in this section were originally developed by Gambatese (1996). The concept of designing for construction safety includes modifications to the permanent features of the project and preparation of plans and specifications for construction in such a way that construction site safety is considered. It also includes hazard control and communication of risks regarding the design in relation to the site and the work to be performed. The research findings linking the incidents described in the FACE reports and OSHA reports to these design suggestions are presented below.
Annex C. New Design Suggestions Linked to Incidents The findings linking the incidents described in the FACE reports and OSHA reports to new design suggestions are presented below. The author developed the new suggestions while applying the investigation model (in figure 1) to the reviewed incidents.
1 Cohen (1988) recommends a sample size of 214 cases for a chi-square test of independence (alpha = 0.05, power = 0.80, effect size = 0.30, and df = 16). This paper appears in the eLCOSH website with the permission of the author and/or copyright holder and may not be reproduced without their consent. eLCOSH is an information clearinghouse. eLCOSH and its sponsors are not responsible for the accuracy of information provided on this web site, nor for its use or misuse. © 2006, CPWR – Center for Construction Research and Training. All rights reserved. This research was made possible by CPWR – Center for Construction Research and Training (CPWR) as part of a cooperative agreement with the National Institute for Occupational Safety and Health, NIOSH (NIOSH grant CCU317202). The research is solely the responsibility of the authors and does not necessarily represent the official views of NIOSH. CPWR—a research, development, and training arm of the Building and Construction Trades Department, AFL-CIO—is uniquely situated to serve workers, contractors, and the scientific community. A major CPWR activity is to improve safety and health in the U.S. construction industry. CPWR, Suite 1000, 8484 Georgia Ave., Silver Spring, MD 20910, 301-578-8500, www.cpwr.com. eLCOSH | CDC | NIOSH | Site Map | Search | Links | Help | Contact Us | Privacy Policy |