Safety Science 65 (2014) 54–62

Contents lists available at ScienceDirect

Safety Science

journal homepage: www.elsevier .com/locate /ssc i

Construction worker fatalities related to trusses: An analysis of the OSHAfatality and catastrophic incident database

0925-7535/$ - see front matter � 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.ssci.2013.12.016

⇑ Corresponding author. Tel.: +1 850 857 6451.E-mail addresses: agrant1@uwf.edu (A. Grant), hinze@ufl.edu (J. Hinze).

Aneurin Grant a,⇑, Jimmie Hinze b

a Bldg. Const. Program, Dept. of Applied Science, Univ. of West Florida, 11000 University Parkway, Building 70, Pensacola, FL, USAb M.E. Rinker, Sr. School of Bldg. Const., Univ. of Florida, 304 Rinker/P.O. Box 115703, Gainesville, FL, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 6 October 2012Received in revised form 12 November 2013Accepted 17 December 2013Available online 22 January 2014

Keywords:ConstructionSafetyTrussesBracingTemporary structuresFatalities

This study was conducted to gain a better understanding of the risks associated with truss installation inbuilding projects. The Occupational Safety and Health Administration (OSHA) fatality and catastrophicincident database was analyzed for the years inclusive of 1990–2009. The database includes over15,000 incidents, 211 of which pertain to trusses. The incidents were analyzed as to the number of fatal-ities per incident, the type of truss, the truss material, the activity taking place at the time of the accident,the release of the hoisting equipment, the initiation of the accident, the presence of bracing materials, thetype of construction, the length of the trusses, the location of the incident, the type of accident (fall,caught-in/between, struck by, or electrocution), and the year the fatality occurred. Many of the accidentsoccurred at elevation and were initiated in large part by moving or falling objects. The study recommendsthat further research should focus on the stabilization of incomplete roof structures and the implemen-tation of best practices for fall protection while performing truss-related work.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The dangers involved with construction activities are well-known (Swuste et al., 2012). Some activities can be particularlydangerous, even fatal, if proper safety protocols are not followed.One such activity is the setting of trusses on buildings. Trussesare generally heavy and cumbersome. To safely perform trussinstallation work, special safety precautions must be exercised.

Much of the work associated with truss installation is per-formed at elevation. Fall protection is of course a principal concern.However, freedom of movement may also be compromised bytruss spacing and configuration. Some workers are compelled totake unnecessary risks, to perform tasks in awkward positions orin unsafe locations, as in the cutting of truss tails when positionedon ladders. Building design, materials, dimensions and site condi-tions are often unique, which requires adaptation and a learningcurve from site to site. Injuries may occur in a number of waysand at every juncture of the process.

A considerable amount of research has focused on minimizingexposure to risk on construction sites. Weinstein et al. (2005) havesuggested that construction safety issues are sometimes best ad-dressed during the design phase, and that design may have a direct

impact on job site safety. Derr et al. (2001) and Kines (2002) haveattempted to elucidate the causes of falls on construction projects.A study by Kaskutas et al., 2009a based on the St. Louis Audit of FallRisks (SAFR), showed that falls were particularly relevant in thesetting of trusses, and that proper safety protocols were observedin the setting of trusses only 28% of the time. A subsequent studyconducted a comprehensive needs assessment to determine gapsin the school-based apprentice carpenters’ fall prevention training.The study showed that less than one third (33%) of the apprenticesreceived school-based training in home-building tasks that involveworking at heights, such as in the setting of trusses. The research-ers state that it is difficult to demonstrate proper fall protectiontechniques without the use of a full scale model and that, ‘‘securinganchorage sites while setting roof trusses poses unique chal-lenges’’. Moreover, the authors identify truss setting as a ‘‘priorityarea’’ where additional training resources should be directed(Kaskutas et al., 2009b). Truss setting is perceived to be especiallyrisky by apprentice carpenters (Kaskutas et al., 2010; Lipscombet al., 2008). Thirty-six apprentices were surveyed in Lipscombet al. (2008), and perceptions of risk included work done on thetop plate of a framed wall, receiving trusses from a crane, placingpiggybacks, ‘‘riding the ridge’’, and work performed on the topplate of a wall without fall protection. It is logical to conclude thatapprentice carpenters are, or may be perceived to be, at a higherrisk for accidents than more experienced journeymen, in thatjob-specific and specialized knowledge is not acquired instanta-neously or through brief training regimens. Rather, it has been

A. Grant, J. Hinze / Safety Science 65 (2014) 54–62 55

demonstrated that meaningful knowledge and greater understand-ing of one’s role within an organization is acquired through re-peated application, reflection, and discussion with others(Gherardi et al., 1998). Furthermore, perceptions of risk are notlimited to work at elevation. The risks associated with truss instal-lation also include the storage and movement of materials, the re-moval of packaging, material handling, setting the trusses,installing temporary bracing, hoisting, ‘‘rolling the trusses’’ (i.e.setting by hand), bracing and blocking, cutting the truss tails andinstalling fascia. In the analysis of 654 non-fatal construction inci-dents involving framing contractors, truss related activities wereshown to be among the most hazardous, including 138 rooftruss-related incidents and 15 floor truss related incidents (Mitrop-oulos and Guillama, 2010). A special emphasis is placed on thebracing of truss structures, as a part of both the temporary and per-manent structures (Iwicki, 2010; McMartin et al., 1984; Salinaset al., 1985; Galambos and Xykis, 1991; SBCA, 2011). The SBCApublication is particularly comprehensive in this regard, providingdetailed recommendations and diagrams for bracing, handling,installing and restraining trusses.

There are extensive standards of practice related to the settingof trusses. The Occupational Safety and Health Administration(OSHA) Code of Federal Regulations 29, Part 1926 (C.F.R. 29, Part1926), Safety and Health Regulations for Construction has definedsafe work practices pertaining to many aspects of truss installation,including scaffolds (subpart L), fall protection (subpart M), cranes,derricks, hoists, elevators and conveyors (subpart N), steel erection(subpart R), demolition (subpart T), and power transmission anddistribution (subpart V). While the OSHA construction regulationsare often used as the primary protocol for job site safety in the Uni-ted States, references to truss installation are intermittent, and insearching for best practices, the information must be first drawntogether and synthesized. Furthermore, subpart M, 1926.502 (k)is somewhat vague, stating that a ‘‘fall protection plan’’ is anacceptable alternative to personal fall arrest systems (PFAS),guardrails or safety nets, should the employer be able to demon-strate that conventional means of fall protection are infeasible, orconstitute a greater hazard (OSHA Construction Industry Regula-tions). This is particularly relevant to the installation of trusses inresidential settings, in that there is an obvious limitation to theuse of Personal Fall Arrest Systems (PFAS), and other conventionalmeans of fall protection are often perceived to be cost prohibitive.There is also a common perception that the enforcement of theOSHA regulations (1926) is comparatively lax in the residentialsector and that experience and common sense adequately addresspotential safety concerns. It has been shown that work experienceplays an important role in hazard recognition and in the avoidanceof accidents (Jeong, 1998). However, we must acknowledge thatproper safety training and protocol are essential to job site safety.If training methods are absent or ineffective (Kaskutas et al.,2009b), and the dynamics of the dangers are not fully understood(Mitropoulos and Guillama, 2010), further research is certainly jus-tified. Hence, this study was initiated in the interests of furtherclarifying the risks associated with truss installation, and as a basisfor improved safety protocols. It is believed that an analysis oftruss-related incidents in the OSHA fatality and catastrophic inci-dent database will help to achieve this research objective.

2. Method

Data were extracted from the OSHA fatality and catastrophicincident database for the years inclusive of 1990–2009. The data-base includes descriptions of over 15,000 incidents investigatedover a twenty-year period. Truss-related fatalities were noted in

the narrative descriptions of 211 incidents. These incidents wereanalyzed according to the following variables:

� truss characteristics,� number of fatalities per incident,� activity taking place at the time of the accident,� release of the hoisting equipment,� initiating factors,� bracing materials,� location of the incident,� type of accident (fall, caught-in/between, struck by, or

electrocution),� year the fatality occurred.

Narrative descriptions of each incident were coded according toa fixed set of variables. For example, truss materials were coded aswood, steel, or unknown (if the material was not specified).Descriptive statistics were used to categorize the data. It shouldbe noted that the OSHA database rarely includes non-fatal inci-dents. Rather, major injuries are only reported in this database ingroups of three or more, or when coinciding with a fatality. Simi-larly, the number of near misses is not reported. It is thereforeacknowledged that the analysis of such incidents could revealtrends not reported in this study.

3. Results

3.1. Summary findings

3.1.1. Type of constructionEighty-nine incidents specified the type of construction project.

Fifty-nine (66%) of these incidents occurred in a residential con-struction setting, twenty-five (28%) occurred in commercial con-struction, and five (6%) in industrial construction. Thisdistribution does not seem to reflect the number of employees ineach sector of the construction industry. According to the UnitedStates Department of Labor, an estimated 1.6 million people areemployed in the construction industry, with approximately halfemployed in residential construction specifically (USDOL, 2013).There are perceived differences in the implementation of job sitesafety requirements from sector to sector. Safety protocols in thecommercial construction sector are often stricter than they are inthe residential sector. Similarly, it is often suggested that residen-tial firms have limited resources for safety, and generally speaking,residential construction sites present distinct hazards from thosepresent on commercial sites.

3.1.2. Number of fatalitiesOf the 211 incidents examined in the OSHA database, there were

205 incidents of single fatalities, four with two fatalities per acci-dent, one accident involving three fatalities, and one with four fatal-ities. While most fatality incidents pertaining to truss workinvolved a single fatality, the data indicate some potential for multi-ple fatality incidents. For example, the four-fatality incident in thestudy involved the collapse of 34 trusses, each 69 feet (�21 m) inlength. The incident description does not explicitly refer to any kindof temporary bracing. However, it can be concluded that consider-able quantities of heavy material were involved, and that the mate-rials were not fully stabilized. The three-fatality incident reportdescribes the temporary and loose fastening of truss and structuralroof material. Two separate (single fatality) incidents describe a‘‘domino’’ effect where temporary bracing was found to be inade-quate. These incidents identify temporary bracing as a common ele-ment and one of the most important safety control measures intruss installation. Temporary bracing is especially important in

56 A. Grant, J. Hinze / Safety Science 65 (2014) 54–62

the initial stages of truss installation, where sheathing materialsand other stabilizing structures have yet to be installed.

3.2. Truss characteristics

3.2.1. Truss typeOne-hundred and eighteen incidents described a specific truss

type. One-hundred (85%) of these were described as roof trusses.An additional ten incidents (8%) described other types of trussesthat are common for roof installations, including four (3%) webtrusses, two (�2%) scissor trusses, two (�2%) gable trusses, one(�1%) outrigger truss, and one (�1%) jack truss. In total, roof typetrusses comprised ninety-three percent of the sample dataset. Se-ven incidents (6%) referred to floor trusses. One incident (�1%) re-ferred to a plate truss. Each of the floor and plate truss incidentsdescribed the work as being performed at elevation.

3.2.2. Truss materialSeventy-one incidents described a specific truss material. Forty-

one (58%) incidents involved wood trusses, and thirty (42%) in-volved steel trusses. It should be noted that there are importantdifferences between structural steel and wood materials. Steel isfar denser than wood, and depending on the size of the individualmembers, steel trusses tend to be much heavier. The hoisting andstabilization of heavier objects – those that could potentially spangreater distances – may prove more difficult and requires addedvigilance. Beyond this basic distinction, there is no evidence thatwood trusses present a greater hazard than steel trusses, or viceversa. Rather, the greater number of wood trusses in the samplemay simply indicate the prevalence of wood trusses in the con-struction industry.

3.2.3. Truss lengthTwenty-seven incidents specified the length of the truss. Seven-

teen of these incidents (63%) involved trusses that were forty feet(12.2 m) in length or longer, six (22%) had lengths between twentyand forty feet (6.1–12.2 m), and four (15%) were twenty feet(6.1 m) in length or less. While these incidents represent only13% of the entire 211 incident sample, the distribution of frequen-cies should not be ignored. The hazards posed by larger trussesshould be given additional consideration, particularly as they

Fig. 1. Principal activity performed

may pertain to the bracing of temporary structures. This recom-mendation is echoed in the Structural Building Components Asso-ciation (SBCA) publication, where bracing requirements intensifyin accordance with increases in truss span. There are also implica-tions for the hoisting and movement of larger, heavier trusses, asthe requirements for rigging, the size of equipment and the place-ment of workers necessitate added precaution.

3.3. Activity

Two-hundred and four incidents described the activity takingplace at the time of the accident. Seventy-eight of these incidents(38%) described an adjacent activity, such as the installation ofsheathing and fascia board, painting, mechanical installations, orwork on lower floors. Incidents were classified in this mannerwhen workers had no direct involvement with truss installationwork. Seventy-one (35%) incidents described employees directlyengaged in the setting of trusses. These incidents describedemployees positioning or re-positioning trusses as they were set,or connecting the trusses to the existing structure. Twenty-fourincidents (12%) were a result of hoisting operations. Hoisting fatal-ities were attributable to the improper selection of hoisting equip-ment, deficient rigging practices, absence of taglines, or theunnecessary exposure of employees to moving loads. Fourteenincidents (7%) occurred during the stacking, loading or unloadingof materials. These incidents occurred at both ground level andat elevation. Demolition activities resulted in eight incidents(4%). In these cases, poor choices were made and inadequate eval-uations of the remaining structure resulted in a collapse. Four inci-dents (2%) were the result of transportation, where large loadsstruck employees standing adjacent to the path of travel. Threeincidents (1%) involved the mounting or dismounting of an adja-cent work platform. Two incidents (1%) were attributable to cut-ting or making truss modifications prior to installation. Theresults of this analysis are presented in Fig. 1.

3.4. Release of crane or hoisting equipment

Forty incidents described varying stages of crane or hoistingequipment engagement or release. Thirty-one fatalities (78%) tookplace while the hoisting equipment was still engaged. These

at time of accident (N = 204).

Fig. 2. Truss fatalities by initiating factor (N = 171).

A. Grant, J. Hinze / Safety Science 65 (2014) 54–62 57

incidents involved truss members striking exposed employees,resulting in direct trauma, or a loss of balance and a subsequentfall. Nine incidents (22%) described situations where the hoistingequipment had been released. In these cases, the trusses shouldhave been connected at both ends, and properly braced to preventany type of movement or overturning.

Table 1Recommended spacing of Top Chord Temporary Lateral Restraint (TCTLR) accordingto truss span.

Truss span Top Chord Temporary Lateral

3.5. Initiation

3.5.1. Primary initiating factorsOne-hundred and seventy-one incidents identified a clear initi-

ating factor in the accident. One-hundred and five (61%) of theseincidents were initiated by falling or moving objects, which in-cluded examples of structural collapse, movement or shifting dur-ing hoisting, and unsafe transportation practices. The weight andbulk of the trusses certainly played significant roles in the causationof the majority of these accidents. Large trusses have considerablemomentum when moving, and they can be difficult to stop. Severalincidents described crushing type fatalities. Other incidents de-scribed workers being ‘‘nudged’’ from their work positions at eleva-tion. Fatalities also occurred during the transportation of trussmaterials, due to oversized and unstable loads, slippage, and dom-ino-effect scenarios. Forty-nine incidents (29%) were initiated whenworkers lost their balance. Many of these incidents describedemployees working at elevation, e.g. on the top plate of a framedwall without fall protection. Other incidents describe contact withequipment or hoisted materials causing a loss of balance and subse-quent fall. Thirteen incidents (8%) were the result of electrocutions.Twelve of these thirteen incidents involved contact with overheadpower lines, either through hoisting operations or the movementof materials. Evidently, the blanketing of power lines or the tempo-rary suspension of electrical current was not requested on theseprojects. None of the electrocution incidents mentioned a dedicatedspotter. Three incidents (�2%) were the result of adjacent hazards,identified as holes in elevated floor structures or unsafe work areas.One incident (�1%) was the result of an employee passing out whileworking at elevation, resulting in a fatal fall. The results of this anal-ysis are presented in Fig. 2.

Restraint (TCTLR) spacing

Up to 300 (9.1 m) 100 (3.05 m) o.c. max300 (9.1 m) to 450 (13.7) 80 (2.44 m) o.c. max450 (13.7 m) to 600 (18.3 m) 60 (1.83 m) o.c. max600 (18.3 m) to 800 (24.4 m)a 40 (1.22 m) o.c. max

a Consult a registered design professional for trusses longer than 600 (18.3 m).

3.5.2. Co-initiating factorsSeveral incidents had more than one initiating factor, or a chain of

events that preceded the incident. A set of co-initiating mechanismswere therefore identified for forty-one incidents. Eight incidents(19%) referred to the trusses being improperly or inadequately set.

On five occasions, this led to a primary initiating factor of falling ormoving objects. Twenty-two incidents (54%) described the move-ment of equipment as a co-initiating factor. In thirteen of these inci-dents, this led to the falling or movement of materials. Equipmenttypes included forklifts, personnel lifts and cranes. The incidents de-scribed machinery tipping over, equipment striking employees, ormaterials otherwise poorly controlled by the equipment operator.Eleven incidents (27%) described weather as a co-initiating mecha-nism. Logically, all construction work becomes more hazardous withinclement weather, wind, wet materials and slippery workplatforms.

3.6. Temporary bracing

Thirty-six incidents made specific references to the use of tem-porary bracing. Fifteen of these incidents (42%) stated that therewas no bracing between the trusses. The remaining twenty-oneincidents (58%) described scenarios where bracing had been in-stalled. However, five of these twenty-one incidents describedthe bracing as ‘‘inadequate’’ and two described a roof structure col-lapse following the removal of temporary truss bracing. Guidelinesare available for the temporary bracing of truss structures,although these are not codified in any manner. The StructuralBuilding Components Association and the Truss Plate Institutepublication, Building Component Safety Information: Guide to GoodPractice for Good Handling, Installing, Restraining and Bracing of Me-tal Plate Connected Wood Trusses (SBCA) offers comprehensive dia-grams and tables for the temporary bracing of truss structures, asshown in Table 1 and Figs. 3–7.

As stated previously and shown in Table 1, temporary bracingshould be spaced at more frequent intervals as truss spans increase.Generally, larger truss spans portend a commensurate increase inmass, and additional bracing largely negates any potential

Fig. 3. BCSI: Bracing of initial gable truss – Section View.

Fig. 4. BCSI: Bracing of initial gable truss and installation of Top Chord Temporary Lateral Restraint (TCTLR).

Fig. 5. BCSI: Recommended Spacing for Top Chord Diagonal Braces.

58 A. Grant, J. Hinze / Safety Science 65 (2014) 54–62

instability in these structures. As shown in Fig. 3, temporary struc-tures and ground bracing are often used to initiate the installationof trusses. According to the SBCA, this type of temporary structureis ideally located at one of the gable ends of the structure. The SBCAoffers alternative bracing configurations for hip roof and piggybacktruss scenarios, with the general premise remaining the same; brac-ing should be initiated as quickly and effectively as possible. Theseconcepts are reinforced in Fig. 4, which shows exterior ground

braces in alignment with the Top Chord Temporary Lateral Re-straint (TCLR). Alternatively, ground braces may be installed towardthe interior of the building if space is constrained. Fig. 5 illustratesthe requirements for diagonal bracing along the top chord of thetrusses, with a maximum spacing of one diagonal brace for everyfour trusses. Fig. 6 shows the spacing of bracing along the web ofthe trusses. The SBCA states that the installation of web diagonalbracing is intended to provide triangular support between the top

Fig. 6. BCSI: Recommended Spacing for Diagonal Web Bracing.

Fig. 7. BCSI: Recommended Spacing of Temporary Bracing Along the Underside or Bottom Chord of Trusses.

A. Grant, J. Hinze / Safety Science 65 (2014) 54–62 59

and bottom chord temporary lateral restraints, perpendicular to theplane of the trusses. Such bracing is intended to create additionallateral stability. The SBCA suggests a spacing of one brace per 10trusses, or 20 feet (approximately 6.1 m) maximum. Additionally,web spacing should range from 10 to 15 feet (3.05–4.57 m) maxi-mum. Bracing for the bottom chord (Fig. 7) should be spaced inaccordance with the web diagonal bracing, with lateral bracingspaced no more than 15 feet (4.57 m) apart, and diagonal horizontaltruss braces every ten spaces, or 20 feet (approximately 6.1 m)maximum.

3.7. Location

One hundred and ninety-six incidents described a specific loca-tion. One-hundred and sixty-six incidents (85%) occurred at eleva-tion; ninety-four of these incidents (48%) were on upper storiesand 72 (37%) were on roofs. Twenty-six (13%) incidents took placeat ground level, with employees working beneath hoisted loads orthe collapse of unstable structures. The remaining four incidents(2%) took place on an adjacent platform.

3.8. Type of accident

One-hundred and thirty-eight incidents (65%) were classified asfalls, 28 (13%) as caught-in/between, 22 (10%) as struck-by acci-dents, and 14 (7%) as electrocutions. An additional nine incidents(�4%) described indiscernible or multiple mechanisms of injuryor fatality and were therefore categorized as other, as depicted inFig. 8.

3.9. Falls and initiators

Forty-four of the 101 fall incidents (44%) were initiated by fall-ing or moving objects. An additional forty-nine incidents (49%)were initiated by a loss of balance. The remaining eight incidentswere the result of other initiating factors, including four incidents(4%) due to a deficient working platform, three incidents (3%) dueto an adjacent hazard, and one incident (�1%) due to an employeepassing out.

3.10. Struck by and caught-in/between

Fifty incidents were categorized as either struck by or caught-in-between. In 28 (56%) of these 50 incidents the employees wereworking beneath unsecured materials. In 20 incidents (40%), theemployees were working beneath secured materials or in otherlocations. In two incidents (4%), the position of the workers wasnot known. Of the 28 incidents where employees were working be-neath unsecured materials, 14 involved non-essential employees.Conversely, 12 of the 28 incidents involved employees who wereessential to the principal activity taking place. In the remaining 2incidents, the role of the worker was not clear. These results arepresented in Fig. 9.

3.11. Year of fatality

Seventy-nine incidents identified the year the fatality tookplace. Of these 79 incidents, a minimum of two fatalities, and amaximum of seven fatalities occurred each year. There was no

Fig. 8. Fatalities by accident type (N = 138).

Fig. 9. Worker location in struck by and caught in/between incidents (N = 50).

60 A. Grant, J. Hinze / Safety Science 65 (2014) 54–62

observable downward trend in fatalities during the study period(1990–2009). On the contrary, forty-six of the incidents (58%) oc-curred in the latter 10 years of the study period, as compared with33 incidents (42%) in the first 10 years, as depicted in Fig. 10.

4. Discussion

Several patterns emerged in the analysis of these data. For in-stance, 138 of the 211 incidents were categorized as falls. This, ofcourse, is not an unexpected discovery. Logically, there is certainrisk of falling with any construction work performed at elevation.However, the circumstances of worker exposure were unantici-pated. In many instances, worker exposure occurred in contraven-tion to established safety guidelines. In other cases, it seems thatbest practices may not adequately address potential exposure, orthat safety guidelines may have been misinterpreted. There are alsoquestions about temporary bracing. Twenty-two incidents reporteddeficient bracing; 15 incidents described no bracing, 5 describedinadequate bracing and 2 described the removal of bracing. Whilethe SBCA publication offers comprehensive guidelines for the tem-porary bracing of trusses, there is no indication, at least from theanalysis presented in this study, that standard industry practice

adheres to these guidelines. Fifty incidents were classified as struckby or caught-in/between. In part, these incidents can be explainedby the nature of truss installation work. Trusses are heavy andmay move with considerable momentum. However, this studyshows an unexpected exposure. Twenty-eight fatalities occurreddue to employees working beneath truss installations, fourteen ofwhom were non-essential to the principal activity.

4.1. Falls

The data revealed that falls are a major hazard in the setting oftrusses. Most of the 138 fall incidents in this study do not mentionthe use of fall protection. In part, this is due to the fact that mostconventional means of fall protection are ill-suited to truss instal-lation work. Personal Fall Arrest Systems (PFAS) are of limited usesince these devices are to be connected to an anchor above the em-ployee, with pull-out a capacity of 5000 lb. Safety nets have similarstructural requirements and may prove costly. Scaffolds provide analternative to PFAS and safety nets, however these systems are alsorelatively expensive. Subpart M, Appendix E of the OSHA standardoffers a fall protection plan as an alternative to conventional meansof fall protection, should the employer be able to demonstrate that

Fig. 10. Truss fatalities by year (N = 79).

A. Grant, J. Hinze / Safety Science 65 (2014) 54–62 61

conventional means of fall protection are infeasible, or constitute agreater hazard. There are also apparent differences in the enforce-ment of fall protection regulations, e.g. Title 8 of the CaliforniaCode or Regulations 1669(a) and 1670 (California Code ofRegulations).

Truss installation work is therefore unique with respect to fallprotection, and so are the mechanisms and initiators of the associ-ated fatalities and injuries. Many of the incidents in the OSHA data-base described employees walking or resting on the top plate of aframed wall, a surface generally between 3.5 in. (89 mm) and5.5 in. (140 mm) wide. Perhaps, it is not surprising that forty-nineof the 138 falls categorized in this study occurred due to a basic lossof balance. In addition, forty-four of the fall incidents in this studywere initiated by falling or moving objects. It seems plausible thatthis latter group of falls may have been averted with appropriatelysized working platforms or proper fall protection in place.

4.2. Temporary bracing

Twenty-two (22) of the incidents examined referred specificallyto deficient bracing. It should be noted however that the approachto coding these data was decidedly conservative; the incidentswere only coded as having deficient bracing if the incident descrip-tion specifically stated so. Many more incidents suggested thatdeficient bracing practices were employed, amongst these, a por-tion of the one-hundred and five (105) incidents that referred tofalling or moving objects as the chief initiating factor. While no di-rect relationship between these two categories was establishedduring the course of this study, it is important to consider the po-tential impacts of deficient bracing. Certainly, this is a topic thatshould be investigated in future research.

4.3. Limited access zones

Several incidents described employees working underneathunsecured materials, either as the materials were being hoisted,or being set into place. Fifty incidents were categorized as eitherstruck by or caught-in/between. Twenty-eight of these incidents

(56%) described employees working underneath unsecured materi-als. On fourteen occasions, these employees were non-essential.

There are of course specific dangers associated with the hoistingand installation of heavy materials. For example, it is particularlyhazardous to hoist materials above active work areas or employ-ees. In the setting of trusses, there is a clear distinction betweenessential and non-essential workers. The truss cannot be set with-out connectors, and when hoisting, the load may need to be stabi-lized with taglines. A spotter is appropriate in certaincircumstances. All other activities should be considered non-essen-tial and performed independently, away from the principal area ofactivity. The analysis of these data reveals the involvement of non-essential workers in several incidents.

5. Conclusions

There are limitations to the OSHA fatality and catastrophic inci-dent database (1990–2009). Some of the incident reports are min-imally descriptive, and in looking for a categorical breakdown ofcertain variables, certain information has been left out. Even so,several themes have emerged from this analysis.

Fall protection is a major concern. Sixty-eight percent of theaccidents described in this database resulted in falls. As trussinstallation work is primarily conducted at elevation, this findingis perfectly plausible. Yet, when the predecessors of these fallsare examined, the absence of fall protection becomes conspicuous.Many of the incidents describe fall protection as inadequate andnon-compliant. It is believed that fall protection standards maybe misinterpreted with respect to truss installation. There are safeways to install trusses, with adequate means of fall protection.Sawhorse scaffolds may be used as working platforms at lower ele-vations. Personnel lifts may be used as work platforms at higherelevations. In contrast, the use of the top plate as a ‘‘working plat-form’’ is a far riskier proposition. Although, it should also be notedthat standards vary as to the threshold elevation for fall protection;C.F.R. 29, 1926 requires fall protection above six feet (1.83 m),whereas Title 8 California Code or Regulations 1669(a) mentionsan exception for truss installation at fifteen feet (4.57 m) above

62 A. Grant, J. Hinze / Safety Science 65 (2014) 54–62

the finished floor. C.F.R. 29, 1926 also refers to an alternative toconventional fall protection, the fall protection plan.

There are risks associated with the movement and hoisting ofmaterials. One-hundred and five incidents were initiated by fallingor moving objects. These incidents are explained in part by unsafehoisting practices. Improper bracing techniques offer an alterna-tive explanation, as explicitly described in 22 incidents. There issome evidence that deficient bracing techniques were responsiblefor a greater number of incidents. However, as the data in thisstudy were coded using a conservative approach, any conclusionsin this regard seem premature in the absence of an accurate mea-sure of magnitude.

6. Future research

Further research is needed to examine the actual practice of fallprotection during the installation of trusses. This study indicatesthat fall protection is a factor in many truss related fatalities. How-ever, the data are insufficiently descriptive at times. A more com-plete survey, designed specifically for the evaluation of trusssafety might provide a basis for re-evaluating current safety guide-lines and standards.

It is unclear why the residential construction sector has morefatal truss incidents than the commercial and industrial sectorscombined. Certainly, this difference warrants further investigation.A greater scale is implied with commercial and industrial construc-tion. Furthermore, there are perceived differences in the prioritiza-tion of safety from sector to sector.

The topic of temporary bracing warrants further investigation.More specifically, future studies should focus on the differencesbetween industry recommendations and actual practice. The dataused in this study suggest that temporary bracing is not installedin accordance with best practices. A more focused study isappropriate.

Additional context would be provided in an analysis of non-fatal incidents, such as lost-time injuries and near misses. It is cer-tainly possible that additional observations would emerge fromsuch an analysis. Likewise, an international perspective on thesafety of truss installation could also provide insight. Perhaps thereare differences in the incidence of fatalities or injuries in interna-tional settings, such that the research could pinpoint specific indi-cators of safety as they pertain to the installation of trusses.

Ultimately, this study provides a foundation for further study.Continued research will provide a basis for truss-specific guide-lines, best practices and training principles. Such studies may evenlead to the future codification of relevant standards.

Acknowledgements

The authors would like to thank Nathan Sink, Junji Brown, Mar-shall Ewing and Shelby Johnson for their creative input to thispaper.

References

29 C.F.R., 1926, 2012. OSHA Construction Industry Regulations. U.S. Department ofLabor.

Derr, J., Forst, L., Chen, H.Y., Conroy, L., 2001. Fatal falls in the US constructionindustry, 1990–1999. J. Occup. Environ. Med. 43 (10).

Galambos, T., Xykis, C., 1991. The effect of lateral bracing on the stability of steeltrusses. J. Constr. Steel Res. 20 (4).

Gherardi, S., Nicolini, D., Odella, F., 1998. Toward a social understanding of howpeople learn in organizations: the notion of situated curriculum. Manage. Learn.29 (3), 273–297.

Iwicki, P., 2010. Sensitivity analysis of critical forces of trusses with side bracing. J.Constr. Steel Res. 66 (7).

Jeong, B.Y., 1998. Occupational deaths and injuries in the construction industry.Appl. Ergon. 29 (5), 355–360.

Kaskutas, V., Dale, A.M., Nolan, J., Patterson, D., Lipscomb, H.J., Evanoff, B., 2009a.Fall hazard control observed on residential construction sites. Am. J. Ind. Med.52 (6).

Kaskutas, V., Dale, A.M., Lipscomb, H., Gaal, J., Fuchs, M., Evanoff, B., The Carpenters’Joint Apprenticeship Program Instructor Team, 2009b. Changes in fallprevention training for apprentice carpenters based on a comprehensiveneeds assessment. J. Saf. Res. 41 (3).

Kaskutas, V., Dale, A.M., Lipscomb, H., Gaal, J., Fuchs, M., Evanoff, B., 2010. Fallprevention among apprentice carpenters. Scand. J. Work Environ. Health 36 (3).

Kines, P., 2002. Construction workers’ falls through roofs: fatal versus seriousinjuries. J. Saf. Res. 33 (2).

Lipscomb, H.J., Dale, A.M., Kaskutas, V., Sherman-Voellinger, R., Evanoff, B., 2008.Challenges in Residential Fall Prevention: Insight From Apprentice Carpenters.Am. J. Ind. Med. 51 (1).

McMartin, K.C., Quaile, A.T., Keenan, F.J., 1984. Strength and structural safety oflong-span light wood roof trusses. Can. J. Civ. Eng. 11 (4).

Mitropoulos, P., Guillama, V., 2010. Analysis of residential framing accidents,activities, and task demands. J. Constr. Eng. Manage. 136 (2).

Salinas, J.J., Gillard, R.G., McMartin, K.C., 1985. Strength and structural safety oflong-span light wood roof trusses. Reliability analysis using safety index. Can. J.Civ. Eng. 12 (1).

Swuste, P., Frijters, A., Guldenmund, F., 2012. Is it possible to influence safety in thebuilding sector? A literature review extending from 1980 until the present. Saf.Sci. 50 (5), 1333–1343.

The Structural Building Component Association (SBCA), 2011. Building ComponentSafety Information: Guide to Good Practice for Good Handling, Installing,Restraining and Bracing of Metal Plate Connected Wood Trusses. IndustryGuide, The Structural Building Component Association and the Truss PlateInstitute.

Title 8 California Code or Regulations, Subchapter 4, 2012. Construction SafetyOrders Article 24. Fall Protection, 1669(a).

United States Department of Labor (USDOL), 2013. Residential ConstructionIndustry (accessed 10.29.13).

Weinstein, M., Gambatese, J., Hecker, S., 2005. Can design improve constructionsafety? Assessing the impact of a collaborative safety-in-design process. J.Constr. Eng. Manage. 131 (10), 1125–1134.