Identifying and Reducing Workforce Fatigue in Rapid Renewal Projects (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

38 C h a p t e r 5 The information collected and analyzed in the team’s field research and reviews of fatigue research from other work domains needs to be viewed in the context of the highway construction industry. This industry has a set of established practices and operational constraints that will influence how fatigue risk management is carried out. This chapter discusses what the team views as critical issues and risks associated with implementing fatigue risk management in this industry. Critical Issues and risks Fatigue Is a Safety Issue not “Owned” by a Specific Stakeholder Occupational and public safety related to fatigue resulting from work schedules generally falls under the purview of a regulatory agency that oversees the industry of concern. This is true for aviation, maritime commerce, interstate trucking, rail- roads, and nuclear power plants. All of these industries are overseen by specific regulatory agencies, and each industry must comply with a set of work-hour rules that are designed to reduce fatigue problems among personnel. Individual organi- zations in each of these industries are responsible for comply- ing with the work-hour limits, or demonstrating that exceptions do not increase risk. Although the work-hour rules are not a panacea for fatigue problems, they have motivated consider- able research and innovation among affected industries to implement fatigue management programs. Additionally, in most of these industries, there have been high-profile incidents or accidents that can be directly linked to work performed at times when fatigue is highly likely (e.g., early morning hours), and in several industries there are active union committees charged with ensuring sufficient rest for their members. Highway construction contractors are distinct from the individual operating entities in the industries mentioned above in that their safety performance is not the purview of an independent federal regulatory agency. OSHA establishes broad operating parameters for rest breaks and overtime pay, and more detailed safety oversight occurs at the state level. OSHA and state safety agencies simply require injury or acci- dent reporting to be documented; they do not investigate for root causes such as fatigue. Fatigue thus falls into an “institutional no-man’s land,” with insufficient motivation in any organization to systemati- cally address the issue. Contractors currently handle worker fatigue in two ways: (1) generally limiting work to 55-h weeks, with the exception of closure periods, and (2) implementing informal countermeasures such as caffeine and rest breaks as necessary. Neither approach represents a systematic means of fatigue reduction; the pressures of job performance seem to outweigh concern with adequacy of worker sleep. The primary risk related to this issue is that the problem will continue to be acknowledged as a work-related risk, but one that does not warrant attention at a level sufficient to address it properly. The field research data suggest that the problem is indeed recognized, yet there are inconsistent attitudes regard- ing training received and the general need to “tough it out.” The team was also told explicitly by one contractor that “when it matters to OSHA, it matters to us.” Fatigue Risk Management Programs May Be Over-Sophisticated for the Current State of the Industry Given the lack of external or internal motivators described in the previous section, and the team’s field observations of con- struction contracting companies, the team is concerned that the concept and particulars implied by “fatigue risk manage- ment systems” (FRMS) may be overly complex for most con- struction companies. The few published analyses of FRMS implementation portray industries of the type described above: regulated operations, usually within the transport sec- tor (Fourie et al. 2010; Gander et al. 2011). It is likely that the safety functions in regulated industries are more developed Critical Issues for Fatigue Risk Management in the Highway Construction Industry

39 than in construction companies. The largest project the team investigated is a $350 million, 3-year program, based on a joint venture between two large companies; this project was served by one full-time safety officer. In smaller firms or projects, the safety function appeared to be served by a foreman on a very part-time basis. The team’s field research did not penetrate the “upstream” portion of the organizations, which may have more extensive safety staff than the team observed. Given these fairly minimal staffing levels (and that available resources are prob- ably small) it would be a considerable challenge to implement FRMS in the manner typically described in the literature. Implementation of FRMS implies a high degree of organiza- tional development into which specific policies and practices concerning fatigue mitigation would be introduced. Available data from process evaluations (Fourie et al. 2010) suggest that even the regulated transport organizations have some difficulty implementing FRMS, due to resource limita- tions, unclear guidance from the regulator, and a tendency to “cut and paste” regulatory guidance templates into a policy document and feed it back to the regulator. Other observa- tions include uneven implementation across organizations, a tendency of smaller operators to lack resources or interest, unfamiliarity or discomfort with fatigue modeling software, increased administrative workload, conflict with organized labor, and inconsistent application. The principal risk in developing FRMS implementation guidance is that it would simply go beyond the capabilities of most contracting firms. The team believes that FRMS as cur- rently conceived in the literature, coupled with low motiva- tion and lack of belief in the economic impact of fatigue, will be too complex for most firms to undertake. This problem also applies to specific aspects of risk management tool devel- opment and project selection. The team’s field research sug- gests that fatigue and work scheduling currently play no role in risk assessment and contract award selection, and that attempting to introduce such approaches without a “push” from project oversight agencies, or at least a clear business value proposition, will be problematic. Safety Data Are Extremely Limited In the literature review the team discussed the limitations of data concerning the linkage of safety problems on the job to fatigue and scheduling parameters. This is the case in the highway construction sector, but also more generally. The estimated range of fatigue-related safety problems is quite wide—from 4% to 33%, using data from various transpor- tation sectors (Gander et al. 2011; Horne and Reyner 1995; McCallum et al. 1996). It is reasonable to assume that the same range is applicable in the highway construction sector. The lack of data concerning accidents and injuries in highway construction in relation to fatigue and scheduling, however, is another factor contributing to reduced motivation to address the problem. Since state agencies do not require collection of prior sleep-wake history in accident reports, there is no data- base being developed to help understand the problem. This problem is inherent in most accident investigation procedures. Unless prior sleep-wake data are collected as a routine matter in accident investigation protocols, only the most severe acci- dents where fatigue is already suspected will address worker sleep history. As described in the fatigue countermeasures review, data are limited regarding the effectiveness of interventions such as training and more sophisticated fatigue risk management sys- tems. Without demonstrated causal relationships between fatigue and accident and injury risk in the highway construc- tion sector, as well as solutions that have been shown to reduce the problem, operators are likely to conclude that they are doing the best they can and are operating with acceptable risk. A Credible Dissemination Pathway for Fatigue Risk Management Tools Is Needed Development of various products, such as fatigue training and countermeasure tools, assumes effective ways to distribute the material. The most important issue to be addressed concerns the mechanisms and responsibilities for an online repository, that is, a website that is routinely maintained and can provide access to the various materials. Experience, however, has shown that a website is insufficient for widespread outreach. In addition, the website needs to be made broadly available through other sources that are routinely contacted by high- way construction safety personnel. Principal organizations in this regard would be the American Association of State and Highway Transportation Officials (AASHTO), the American Road and Transportation Builders Association (ARTBA), the Associated General Contractors of America (AGC), and possibly the OSHA/FHWA Work Zone Safety website. Rapid Renewal Market Growth, Fatigue Safety Impacts, and Training Needs Overview This section discusses the potential size of the rapid renewal highway construction market and the corresponding workforce needs over the next 10 years, and it evaluates the impact this renewal effort will have on the overall workforce, including potential worker training needs and potential impacts on safety. The basis for this analysis was data contained in several sources, including the FHWA Conditions and Performance report to Congress (e.g., U.S. DOT 2008), the Bureau of Labor Statistics (BLS), and state-level data concerning specific historical lane

40 mile construction for different pavement types in order to esti- mate labor requirements for rapid renewal projects. The team’s general finding was that there is considerable uncertainty in making these types of forecasts, because of changes in economic outlook and the level of granularity available in the data. For highway construction, budgets at the federal and state levels, for example, tend to be categorized very broadly in terms of new construction, maintenance, and rehabilitation. Rapid renewal projects fall within the latest category, but there is no information provided within bud- gets that allows determination of the rapid renewal compo- nent. Similarly, the data concerning safety impacts in highway construction are very sparse, thus restricting the team’s ability to estimate the fatigue-related component of safety impacts in rapid renewal. Notwithstanding these limitations, this section describes an analytic process and its application to data that the team deems to be representative program estimates for rapid renewal construction. Rapid Renewal Program Size Estimates Throughout the United States, governments spent $161.1 bil- lion on highways in 2006. About $78.7 billion (48.8%) of this total was spent on capital projects. Of the $78.7 billion of capi- tal spending in 2006, $40.4 billion was spent for rehabilitating the existing system; $16.2 billion was used to construct new roads and bridges; $13.8 billion was used for widening exist- ing facilities; and $8.2 billion supported system enhancements such as safety, operational, and environmental enhancements. The portion of total capital outlay funded by the federal gov- ernment rose from 41.6 to 44.0% between 1997 and 2006; while state and local capital investment increased from $28.3 billion to $44.1 billion. These and related outlays and percentages are illustrated in Table 5.1 (U.S. DOT 2008). A contrasting profile is illustrated by data from California, which shows that rehabilitation work in the 2006 time period comprised more than 50% of the highway construction work, as shown in Table 5.2. This table also lists an important parameter in estimating workforce requirements, namely, the number of lane miles to be rehabilitated. Historical data concerning lane mile construction trends are difficult to obtain. The team reviewed data sources from three states, including California, Texas and Washington, and concludes that Washington state (1) provides the most useful sources of historical construction data in terms of lane miles per year and (2) is more likely “average” in terms of its road network than larger states such as California and Texas. Workforce Estimates The BLS maintains data on current and projected employ- ment in a variety of sectors. This information is available for the specific sector of highway construction. The team’s SME in highway construction engineering developed a compre- hensive list of job classifications for typical projects and used this as a basis for accessing the BLS tool for projecting workforces in specific industrial sectors. The BLS tool pro- jections “are developed in a series of six interrelated steps, each of which is based on a different procedure or model and related assumptions: labor force, aggregate economy, final demand (GDP) by consuming sector and product, industry output, employment by industry, and employment by occupation. The results produced by each step are key inputs to following steps, and the sequence may be repeated multiple times to allow feedback and to insure consistency” (BLS 2012). Table 5.3 provides the output of the BLS tool for the key occupations involved in highway construction. It shows the overall anticipated increase in employment in the 10-year period between 2008 and 2018 is 16%. In order to determine the extent of this workforce that may be involved in rapid renewal construction, the team modeled the work requirements of specific types of lane mile construction—rigid and flexible using the Construction Analysis for Pavement Rehabilitation Strategies program Table 5.1. Highway Expenditures in the U.S. by Category in 2006 Activity Amount (B$) Percentage Rehabilitation (renewal) 40.4 25.1% New roads and bridges 16.2 10.1% Widening existing facility 13.8 8.6% Support system enhancements 8.2 5.1% Maintenance & operations 40.4 25.1% Safety 14.5 9.0% Administration costs 13.2 8.2% Interest and bond retirement 14.2 8.8% Sum 160.9 100.0% Table 5.2. Budget Figures for Highway Construction in California, 2006 (Caltrans 2008) Expenditure Amount Lane Miles Capital Preventive Maintenance $229 M 1,358 Preventive Maintenance $43 M 957 Base Maintenance $28 M 381 Rehabilitation $367 M 654

41 (CA4PRS; http://www.fhwa.dot.gov/research/deployment/ ca4prs.cfm). CA4PRS, FHWA’s market-ready technology product, incorporates three interactive analytical modules: a Schedule module that estimates project duration, a Traf- fic module that quantifies the delay impact of work zone lane closures, and a Cost module that compares project cost among alternatives. Typically, highway pavement renewal projects consist of the following major activities during construction: • Pavement demolition activities: milling, saw cut, excava- tion, loading, and hauling to dumping yards. • Paving activities: material delivery and supply from plants, base paving, asphalt paving, concrete (including precast) paving, compaction, cooling, curing, and finishing. • General activities: traffic control, lane marking, quality control, lighting, clean up, and field management. First, construction productivity (in terms of lane mile per closure) of typical pavement renewals is estimated, using CA4PRS schedule analysis. Based on the productiv- ity, the total size of human resource (contractor crew and agency staff) in terms of their numbers and duration is cal- culated. CA4PRS schedule analysis output indicates the configuration of typical major equipment resources, such as hauling and delivery trucks, paving machine, and pro- duction plants with their usage (operations) hours per clo- sure. Contractors’ crew such as equipment operators and laborers are derived from these major equipment operation activities outputs. The team applied CA4PRS analysis to the anticipated pave- ment renewal expected in Washington state for the next 10 years to provide a baseline for scaling up to the United States. The following assumptions were incorporated: • There are 110 working days per year on average (May to October). • Rapid renewal will be approximately 40% of the total renewal market. • Washington state total renewal over 10 years will be 8,580 lane miles. Using these inputs, CA4PRS calculates that the rapid renewal contractor workforce will require approximately 500 workers per year. Scaled up to the United States this workforce would currently entail 25,000 workers; with the 16% growth expected in highway construction employment by the year 2018, this workforce would require 29,000 workers. Table 5.3. Estimated Workforce in Highway Construction Jobs, 2008–2018 Occupation 2008 2018 Change Employment (in thousands) Percent of Industry Employment (in thousands) Percent of Industry Number (in thousands) Percent Total, all occupations 328.9 100.00 380.4 100.00 51.5 15.7 Construction laborers 83 25.25 104.9 27.57 21.8 26.3 Operating engineers and other construction equipment operators 47.2 14.35 52.9 13.91 5.7 12.1 Paving, surfacing, and tamping equipment operators 17.8 5.41 19.9 5.24 2.1 12.0 Cement masons and concrete finishers 12.8 3.88 14 3.68 1.2 9.8 Construction managers 7.6 2.33 9.4 2.47 1.7 22.6 Excavating and loading machine and dragline operators 3.9 1.18 4.5 1.18 0.6 15.7 Highway maintenance workers 2.6 0.78 2.9 0.78 0.4 14.8 Maintenance and repair workers, general 2.3 0.70 2.7 0.70 0.4 15.7 Reinforcing iron and rebar workers 1.8 0.53 2 0.52 0.2 12.2 Mixing and blending machine setters, operators, and tenders 0.3 0.09 0.4 0.10 0.1 34.6 Construction and building inspectors 0.3 0.08 0.3 0.08 0 10.2 Crushing, grinding, and polishing machine setters, operators, and tenders 0.1 0.04 0.1 0.04 0 12.3 Source: BLS, n.d.

42 Training Needs As indicated in the preceding section, the team estimates that 25,000 to 29,000 highway construction personnel will be engaged in rapid renewal work at some point over the next 10 years, and thus that is the number of workers that the team estimates could benefit from fatigue-oriented training. The working group members who have responded to the project’s request for comments suggest these estimates may be low, given the increasing backlog of maintenance work to be done. This needs to be balanced against the prospect of continuing economic problems and their adverse effect on highway construction. Notwithstanding these factors, the training needs for this workforce are readily apparent on the basis of the team’s field work and assessment of the current state of fatigue counter- measure implementation. The basic content of fundamental training should be the same for labor and management and includes the following content areas: • In the long run, there is no substitute for sleep. • Fatigue is based on physiological mechanisms and cannot be overcome by motivation or willpower. • Self-assessment can be unreliable and potentially biased by work circumstances, but can be useful with attention to specific circumstances. • Individuals vary in sleep need and responses to sleep loss, and it is difficult to predict on a case-by-case basis. • There is no “one-size-fits-all” solution. • There are ways to prevent and mitigate fatigue, but they must be properly employed. • Fatigue has safety, well-being, and economic consequences. Additional content areas that might be stressed for managers would include the following: • Managers are not immune to fatigue and should not use the mantle of responsibility to ignore basic sleep needs. • Rapid renewal schedule work practices have differential effects upon fatigue and can be cumulative. • Specific rapid renewal manifestations, such as lack of a day off when switching to night shifts or many continuous days of work following a closure, need to be addressed with spe- cific countermeasures. • Fatigue-proofing strategies can be developed and employed when project requirements preclude fatigue-reducing schedules. These later two points emphasize the importance of develop- ing countermeasures and interventions that are specific to rapid renewal work. Additionally, the fundamental training material will need to be contextualized to the highway construction environment by providing appropriate illustrative and opera- tionally relevant examples. Although the team had initially considered developing training content that would be specific to state DOT officials, the field work suggests that they are subject to many of the same fatigue risk factors as contractors and can benefit from similar training content. The original NASA fatigue countermeasures training took several days at the NASA-Ames field site and was presented by experts in fatigue research. Over time, the material from that program has been disseminated widely and adopted by numerous operational organizations—primarily airlines. More recent material of a similar nature has been developed for Transport Canada and contains an emphasis on fatigue risk assessment at the individual level by PSWM. These prior efforts tend to generate material that is made publicly avail- able, with the expectation that it will be accessed and adapted by users with training needs. The team’s impressions of the highway construction indus- try suggests that “turn-key” material is most likely to be used, and it should not be expected that individual contractor safety officers will adapt material from publicly available documents. This latter approach has been attempted, and the quality and utility of the resulting adaptations is unclear. The team’s assumption is that safety officers confronted with detailed web-based documents would most likely abandon the effort to develop training material, or down-select material in such a way as to reduce effectiveness. Safety Impacts Numerous studies have demonstrated that impaired neuro- behavioral performance and elevated risks for injuries and accidents are associated with extended schedules and shift work in manufacturing (Folkard and Tucker 2003; Folkard and Åkerstedt 2004; Folkard and Lombardi 2006; Folkard et al. 2006), transportation (Hursch et al. 2006), and medi- cine (Landrigan et al. 2004; Ayas et al. 2006; Barger et al. 2005; Gander et al. 2008; Sharpe et al. 2010). The U.S. construction industry is among the most hazardous for workers. In spite of the prominence of construction in occupational injury and death, however, few studies have evaluated the contribution of schedule or fatigue to this very high risk among these workers, and none evaluates these risk factors among high- way construction workers; studies of accidents among high- way construction workers focus instead on immediate causes (Pegula 2004; Mohan and Zech 2005; Center for Construc- tion Research and Training 2008). Below is a standard risk assessment model used in public health research to guide the estimation of the numbers of persons at risk for any health- or safety-related outcome: M P B DR E= × × ×

43 where • M is the number of deaths (or injuries, etc.) attributable to a cause; • P is the population at risk (in this case, the number of high- way construction workers); • B is the baseline mortality rate (or injury rate, etc.); • DR is the dose response (i.e., the size of the effect of increased exposure on the outcome); and • E is the level of exposure to the causal factor. If all required data elements were available, this model could be used to estimate the numbers of accidents or injuries (M) that could be expected based on different work sched- ules, such as schedules commonly used in rapid renewal high- way construction relative to those used in more traditional highway construction. The main presumption in this analysis would be that fatigue is the intervening factor between work schedule and the increased accident or injury rates. The field research provides some evidence that work sched- ule affects fatigue levels in highway construction workers; how- ever, it cannot provide the data elements necessary to perform the estimation above, primarily due to limits of sample size and study scope. Analyses of risk factors in accidents typically require very large samples and the less frequent the accident, the larger the sample required. In addition, to link accidents reliably with schedule-related factors would require reports to be collected as soon as possible after the accidents, and these reports would need to include detailed information on recent work and sleep schedules. At best, three of the necessary data elements are available in the public domain, and all have significant limitations. First, the population of highway construction workers (P) is esti- mated by the BLS, which also provides projections of this population over the next 10 years. Estimates by occupational category within highway construction are also provided. But BLS figures do not provide a breakdown by type of activity, such as new construction versus renewal. Here, the team uses estimates of the rapid renewal workforce described above. These estimates have their own limitations, including extrapo- lation from operations in a single state to represent operations throughout the country. However, this approach has the poten- tial for estimating workforce needs by occupational group or task, so that specific fatigue profiles could be applied, if known. Second, the literature on construction worker accidents related to schedule provides statistics for some schedule charac- teristics analogous to the dose response (DR) element required, at least for occupational injuries. Two studies using the same national survey dataset (the National Longitudinal Survey of Youth) found links between workplace injuries and work schedule among construction workers. In the first, those working more than 8 h per day (odds ratio = 1.02, p < .01), more than 50 h per week (OR 1.98, p = .03), or starting work before 7:00 a.m. (OR 1.28, p < .01) were all found to be more likely than others to report a work-related injury when adjusting for other risk factors (Dong 2005). Significant find- ings from the second study also included a positive associa- tion between overtime work and likelihood of injury relative to not working overtime (hazard ratio 1.48, p < .05), as well as increased risk of injury related to working the evening shift compared with day shift (hazard ratio 2.86, p < .05), after adjustment for other risk factors (Dembe et al. 2008). A haz- ard ratio is the ratio of two hazard rates and is similar to a relative risk. A few other studies suggest relationships between schedule or fatigue and injuries (Lowery et al. 1998; Arditi et al. 2007; Powell and Copping 2010); however, effects appropriate for risk assessment based on specific schedules are typically not reported. Assuming that effects reported for construction workers generally are adequate proxies for highway construction workers specifically, statistics derived in these studies could be used to estimate numbers of injuries based on working more than 50 h a week compared with working less; start- ing work before 7:00 a.m. compared with starting work later; working overtime hours compared with no overtime; and working evening shift compared with working day shift. The data source used to derive these estimates is publicly avail- able and could be used to make other comparisons based on a number of relatively simple schedule differences. To the extent that these comparisons “look like” the differences between rapid renewal scheduling practices and traditional scheduling practices, the dose response estimates may be useful. However, there are additional limitations to these particular survey-based estimates. The data source relies on self-reporting of hours worked and injuries, and it requires the respondent to recall details over a relatively long period of time. Also, work schedules may be highly variable over time, and information about the schedule being worked at the time of the injury may not be available. Finally, these studies can consider only injuries, not fatalities or other inci- dents of potential interest (e.g., near misses). Third, the same data source that provides limited dose response values can be used to estimate a baseline injury rate (B), that is, the rate of occupational injuries that occur with “traditional” schedules. In order to calculate either the num- ber or the rate of injuries attributable to rapid renewal sched- uling scenarios, the team must first know the portion of the total rate that occurs under the traditional schedule scenario. For example, in one study, 10.4% of construction workers working 7 to 8 h per day reported a work-related injury in the prior year (Dong 2005); this figure could be considered the baseline injury rate. However, all the same limitations apply here as apply to the dose response estimates derived from the same source.

44 The baseline rates for injuries, fatalities, or other incidents are not available from other national sources. Injury and fatal- ity rates reported by the BLS, by OSHA, and others are overall rates, and sufficient information to decompose these rates into partial rates by any work schedule element (e.g., nights versus days, overtime versus non-overtime) is not available. Even the detailed accident reports generated by occupational fatalities do not usually mention things like work schedule or sleep obtained. One state DOT SME respondent when asked why he thought such details were not made part of the official record in either individual reports or in official statistics answered that there was “no interest” in doing so. Some sources provide estimates for the proportion of injuries due to fatigue (Gander et al. 2011; Horne and Reyner 1995; McCallum et al. 1996); however, such estimates are not useful as elements of the risk assessment model because the estimates, as typically reported, do not offer the degree of granularity required to differentiate the fatigue component in different groups, and may have little relevance outside the domain of origin due to substantial dif- ferences in work schedules and task requirements. The final element, level of exposure to the causal factor (E), is not available. In this case, exposure would be operational- ized as the proportion of workers exposed to an “at-risk” schedule (i.e., a schedule used more frequently on rapid renewal projects, such as night and extended shifts) and for how long. The estimate of 25,000 to 29,000 highway con- struction personnel in rapid renewal, about 40% of the over- all highway workforce, is not adequate for estimating level of exposure to risky schedules. There is wide variability in sched- ules in both rapid renewal and traditional projects; extended schedules and night work may be used in both but to varying degrees, and even in rapid renewal environments some sched- uling practices (e.g., nights) are not used consistently. A com- prehensive approach would evaluate, first, the relative risks associated with exposure to different schedules, and second, the relative risks associated with the perhaps increased pro- pensity to use higher-risk schedules over lower-risk schedules on rapid renewal projects. No extant data source provides this information. Elements like exposure, and perhaps the baseline rate and dose response, may be most reliably estimated when data can be obtained directly from sources such as contractor and DOT payroll records and accident reporting mechanisms. Significant bureaucratic obstacles exist, however, in acquiring these kinds of data. Additionally, even if these data were made available from among the projects recruited for the study, there would remain significant methodological challenges to ensuring that the data obtained were sufficiently representa- tive to yield estimates that could be generalized to a larger population. While the team is reasonably confident that ours and others’ findings support the claim that increasing use of rapid renewal practices is likely to result in both higher rates and larger numbers of occupational injuries, it is not possible at this time to quantify these effects with precision. Although a formal risk assessment of the impact of fatigue on occupational safety for highway construction workers can- not be made, the team can make rough, preliminary estimates of the number of fatigue-related injuries in the rapid renewal workforce using two approaches. The first approach combines injury and fatality statistics for the construction industry published annually by the BLS (here, treated as the baseline rate) with published estimates for the proportion of accidents due to fatigue (here, treated as the dose response) and estimates of the size of the rapid renewal workforce (the population), with the assumption that expo- sure is uniform for all members of the rapid renewal work- force. The BLS reported 4.3 nonfatal injuries per 100 full time equivalents (FTEs) in 2009 within the construction industry as a whole [North American Industry Classification System (NAICS) code 23], and 4.6 per 100 FTEs in highway, street, and bridge construction specifically (NAICS 2373) in the same period (BLS, 2010). Estimates for the rate of fatigue- related accidents vary widely, from 4% to 33% (Gander et al. 2011; Horne and Reyner 1995; McCallum et al. 1996); the midpoint of this range is approximately 20%. Applying the midpoint of the fatigue estimate range (20%) to the highway construction injury rate (.046) and the expected number of FTEs based on the projected personnel in rapid renewal high- way construction over a 10-year period (see above), the team estimates at least 230 to 267 nonfatal injuries per year due to fatigue on rapid renewal projects. In the construction indus- try overall, 9.9 fatalities per 100,000 FTEs were reported in 2009 (BLS 2010). When the same procedure is applied to this rate, the team estimates about one-half a death per year in the rapid renewal workforce. The limitations to this approach are clear. The baseline rates used are not true baseline rates because they correspond to the population as a whole, not just to those working “traditional,” lower-risk schedules. Also, the dose response used is not specific to any activity or schedule, and exposure to high-risk versus low-risk schedules is treated as uniform within the rapid renewal workforce. A second approach combines the same population esti- mates as above with estimates for construction workers overall reported by Dong (2005). This study estimated an unadjusted incidence of an occupational injury among workers who reported working more than 8 h per day of 15.0%, compared to only 10.4% (the baseline) among those working a 7-h or an 8-h day. The proportion of the incidence attributable to the longer work shift (analogous to the dose response) is 15.0% - 10.4% = 4.6%. If the projected rapid renewal work- force estimate is treated as the number of individual workers, then between 1,150 (.046 * 25,000) and 1,334 (.046 * 29,000) workers each year in the rapid renewal workforce would be expected to sustain an injury attributable to fatigue as a result

45 of their longer schedules. The number of expected injured due to fatigue is much higher using this approach. A compli- cation with this method is that the incidence rates reported in Dong are based on individual workers rather than FTEs and are therefore not adjusted for exposure. If respondents who worked longer hours have a higher incidence rate than those working shorter hours, this higher rate is partially explained by their increased exposure to the risk injury. This method also requires use of the same problematic uniformity-of-exposure assumption as the first approach. In conclusion, only very tenuous estimates of the safety impact of work schedule and fatigue can be made at this time; the above estimates are simulations only, requiring major assumptions. Data with more granularity are required to make progress on this front, and these data are not likely to become available soon.