International Journal of Industrial Ergonomics 24 (1999) 431}444

An evaluation of manual materials handling in highway construction work Victor Paquet*, Laura Punnett, Bryan Buchholz Department of Work Environment, 1 University Avenue, University of Massachusetts Lowell, Lowell, MA 01854, USA

Abstract Construction work is generally described as strenuous but few systematic characterizations of physical load have been published to date. PATH (Posture, Activities, Tools and Handling) is an ergonomic assessment method used to measure the frequency of exposure (as a proportion of observations) to manual materials handling (MMH) activities and other exposures for construction and other non-routinized work. Observations of loads handled, body postures, tools and materials handled and hand grasps are available for further description of MMH activities. PATH has been used to characterize manual handling by iron workers (in concrete reinforcement work), carpenters (in form construction) and laborers (in the construction of utilities pits) during highway construction operations. Of these trades, the iron workers were the most frequently observed in heavy MMH activities (involving at least 13.5 kg), and they often lifted in twisted or laterally bent trunk postures. Heavy MMH activities by laborers most frequently involved the handling of boards used to construct the pit walls. Little heavy MMH (at least 22.7 kg) was observed in the carpentry tasks. MMH requirements varied greatly among tasks within each construction operation, demonstrating the importance of a task-based assessment of MMH exposures. Analysis of subsets of data on the iron workers and the carpenters showed that, in some cases, task and worker were important sources of variance. For almost all variables, the error term (interpreted primarily as day-to-day di!erences in exposure) was the largest source of variance. Therefore, an evaluation of multiple workers over long time periods for speci"c tasks is necessary for a reliable characterization of MMH requirements and other ergonomic exposures in highway construction work. Relevance to industry The methods described in this paper allow the systematic evaluation of MMH activities in highway construction and other types of non-routinized work. These can be used to identify job tasks having potentially hazardous MMH exposures, and to provide information needed for the appropriate re-design of tasks or materials to reduce physical loading on the musculoskeletal system.  1999 Elsevier Science B.V. All rights reserved. Keywords: Manual materials handling; Exposure assessment; Posture; Construction; Concrete reinforcement; Laborers; Carpenters

* Corresponding author. Present address: Department of Industrial Engineering, State University of New York at Bu!alo, Bell Hall Rm. 342, Box 602050, Bu!alo, NY 14260-2050, USA. Tel: (716) 645-2357 x. 2118; Fax: (716) 645-3302; e-mail: [email protected]!alo.edu 0169-8141/99/$ - see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 8 1 4 1 ( 9 9 ) 0 0 0 0 9 - 8

432

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

1. Introduction Overexertion injuries related to construction work are a serious problem in the United States. In 1992, construction work had the highest rate of job related illness and injury of any industry in the United States (Bureau of Labor Statistics, 1994). Overexertion injuries were the second leading cause of non-fatal injuries and resulted in the greatest number of days absent from work in 1994 for the construction industry's private sector (Bureau of Labor Statistics, 1996). For this industry, sprains and strains were, by far, the most common injury resulting in absence from work, back problems being reported most frequently (Bureau of Labor Statistics, 1996). There is considerable evidence that many features of manual materials handling (MMH) work a!ect a worker's risk of injury or capacity to perform manual handling. The complex interaction of MMH factors that determine physical load or exposure intensity makes the characterization of these activities challenging in the "eld. Furthermore, it is not su$cient only to evaluate the intensity of MMH demands at a single point in time without considering the temporal pattern of loading and variability in MMH exposure over time, which may be very high in jobs such as construction. Detailed assessment of MMH requirements has been performed with biomechanical, physiological and psychophysical approaches. Biomechanical techniques include the use of two-dimensional sagittal-plane static models (Cha$n, 1969), three-dimensional static models for asymmetric lifting activities, and most recently dynamic biomechanical models (e.g., McGill, 1992). Physiological approaches can be used to estimate the energy expenditure requirements of MMH using direct measurements of heart rate or oxygen consumption, or modeling methods that predict energy expenditure from job related and individual factors (Garg et al., 1978). One application of the psychophysical approach has been to determine the maximum weights and forces acceptable to study subjects (Ciriello and Snook, 1983; Ciriello et al., 1990,1993; Snook and Ciriello, 1991). Because direct measurement methods are designed to quantify factors of a speci"c MMH activity

or set of activities, their use has been generally limited to laboratory research or occupational settings that involve simple MMH activities. However, physical demands in many occupations (e.g., maintenance workers, warehouse workers, mechanics, etc.) are usually far more complex, involving the handling of multiple objects that di!er in size, shape, and weight at di!erent frequencies throughout the day, as well as at di!erent frequencies from day to day. Observations of workers, especially in combination with selected direct measurements (e.g., weighing of tools and materials), allows at least partial quanti"cation of MMH requirements. The NIOSH Lifting Equation combines the biomechanical, physiological and psychophysical approaches into a single exposure score to assess sagitally symmetric lifting activities (NIOSH, 1981), and was later revised to include asymmetric lifting as well as other factors (Waters et al., 1993). Checklists have also been used to characterize MMH requirements in occupational settings (e.g., Kemmlert, 1995). These approaches are easy to use, inexpensive and often reliable, but may lack the precision needed to prioritize jobs or job tasks for intervention. Additionally, the duration of observation and number of workers observed needed for a reliable estimate of exposure are rarely speci"ed as part of the observation protocol. In construction, the measurement of MMH requirements is complicated by a constantly changing physical environment, work force and job requirements. The tasks performed by construction workers, the methods and tools used, and the physical characteristics of the construction site are all potential sources of exposure variability. As part of an ongoing research project, we have used an observational approach, PATH (Posture, Activities, Tools and Handling), to evaluate the ergonomic demands on workers in various construction trades during highway construction work. Because of potentially large variability in exposure, an observational work sampling approach was chosen over more detailed methods of assessment so that measurements could be obtained for many di!erent exposures on multiple workers over long periods, rather than only on a limited number of exposures for a few workers over short periods.

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

The objective of this study was to demonstrate how observational work sampling of MMH and other ergonomic factors can be used systematically to evaluate the MMH requirements of non-routinized work. Evaluations of iron workers performing concrete reinforcement work, carpenters building forms and laborers constructing a utilities pit illustrate the method and some general conclusions.

2. Methods 2.1. Highway construction work To facilitate a systematic evaluation of ergonomic exposures in construction, a hierarchical taxonomy was developed to divide the construction process into stages, operations, trades, tasks and activities within each job task (Fig. 1). Information about the stages and operations were obtained from state highway speci"cations, contract speci"cation and other documentation provided by the contractor (e.g., construction schedules). For unionized construction work in di!erent regions of the United States, the tasks of each trade are usu-

433

ally de"ned jurisdictionally, allowing little overlap of responsibilities between trades. For this project, information about the tasks and activities of each operation was obtained primarily by worker interviews and direct observation. Because the taxonomy is based on both engineering speci"cations and worker information, it provides a useful framework for organizing information collected on construction hazards and for communicating this information to the contractors and workers. Additionally, when exposure information is available for multiple job tasks and the distribution of job tasks can be determined for individual workers, unique time-weighted exposures can be constructed for individuals and used in epidemiologic research. Similar classi"cation schemes have been used in building construction (e.g., Niskanen and Saarsalmi, 1983) and could be developed for other types of non-routinized work. 2.2. Study site and population The study took place on a large highway and tunnel construction site in Boston, MA. The entire construction project is a multi-billion dollar e!ort

Fig. 1. Taxonomy of unionized construction work.

434

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

to build a tunnel under Boston Harbor and depress several miles of interstate highway underneath the city. It involves many contractors and subcontractors and thousands of construction workers over the life of the project (12 years anticipated at present). Iron workers, laborers and carpenters were selected for the present study because together they represented approximately 50% of the total work force on the project. Each trade was employed by di!erent contractors and worked on di!erent site locations. Iron workers were responsible for placing and connecting steel rods (rebar) that reinforce concrete structures, while carpenters built and assembled the forms needed for concrete structures. Laborers performed a variety of support tasks which included pouring concrete, erecting sca!olding, housekeeping, stripping forms, and manually excavating and fortifying shafts and tunnels. The evaluation was performed on iron workers reinforcing concrete, laborers building a utilities pit (shaft) and carpenters building and erecting forms. The operations were fairly common and performed repeatedly at di!erent locations on the construction project. Data were collected separately for each trade. 2.3. Assessment of MMH 2.3.1. Data collection PATH was used to characterize MMH exposures for the tasks of each operation. For this, observations are made in real time at "xed, short intervals. At each observation, the observers code the task, body postures, activities, loads handled and tool used (if any) for a single worker, in a check-list type format. Unlike most observational methods that use checklists, evaluations are made repeatedly on crews of workers over periods of several days or weeks. With PATH, tasks can be evaluated for their frequency of MMH activities, body posture categories and loads handled. The observation of multiple workers over long periods permits the evaluation of the variability of exposure among tasks, workers and days (see Buchholz et al. (1996) for a detailed description of the method). MMH activities were de"ned as those that involved handling at least 4.5 kg, excluding activities that involved hand tool or power tool operation (e.g., carrying a jack hammer from one location to

another was de"ned as MMH but operating a jack hammer was not). There were "ve di!erent MMH activities: lift, lower, carry, push/pull (with arms) and move/place (a "ne adjustment of an object's position). Other variables provided additional information about the types or intensity of the MMH activities: load weight, trunk posture, type of hand grasp, number of hands used, and type of tool or material handled. Before each PATH data collection period, tools and materials commonly used by the workers were weighed and the weights were used to estimate the loads handled by the workers. Trunk posture was coded as neutral, mild #exion (203) #exion(453), severe #exion (*453), twist or lateral bend (*203), #exion(*203) and twist or lateral bend (*203). Type of hand grasp was coded as power, pinch, other or none. The iron workers and carpenters were sampled in a random sequence throughout each sampling period, allowing estimation of the frequency of tasks within each operation. At the start of a day's data collection each observer selected a crew of 4}10 workers to follow for sampling periods of 3}4 h. For the laborers, individual workers were randomly selected and observed for the entire sampling period on multiple days. With both approaches, observations were made at "xed intervals, normally 60 or 45 s. Each operation was observed for 9}13 days over a period of 4}5 weeks. Each approach allowed the frequency of MMH activities and loads handled to be estimated for each of the tasks. Prior to formal data collection, PATH coders spent 2}3 days observing each operation, piloting the PATH data collection and checking inter-observer reliability of PATH codes. 2.3.2. Data management and analysis Data were scanned into a personal computer using optical mark recognition software (Remark O$ce OMR, Principia Products, West Chester, PA). The data were visually reviewed for scanning errors and corrected as necessary by manual data entry. The data were analyzed with the Statistical Analysis System for the PC (SAS Institute, Cary, NC). Missing data (i.e., a speci"c exposure was not recorded for an observation) were excluded from each analysis.

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

For the iron workers and laborers, the loads handled were estimated and coded to the nearest 0.5 kg and were then grouped into six weight categories (0 kg, 0 kg(load(1 kg, 1 kg)load( 4.5 kg, 4.5 kg)load(13.5 kg, 13.5 kg)load( 22.7 kg, load)22.7 kg). Heavy MMH was de"ned as handling at least 13.5 kg. For the carpenters, data were collected with an older version of the method in which loads were coded directly into 5 categories (0 kg, 0 kg(load(2.3 kg, 2.3 kg) load(6.8 kg, 6.8 kg)load(22.7 kg, load* 22.7 kg). MMH activities were de"ned as those that involved handling at least 6.8 kg, and heavy MMH involved at least 22.7 kg. The frequency of exposure (as a percent of observations) to individual MMH activities and loads handled was computed and compared among tasks within each operation with the chi-square statistic. The frequency of individual MMH activities (e.g., lift or carry), non-neutral trunk postures and materials handled were calculated for heavy MMH activities. The sources of variance due to task, worker and day on loads handled and body postures were evaluated for subsets of iron workers and carpenters. Because the original data sets were unbalanced across task, worker and day of observation, subsets were selected to achieve balanced task}worker}day combinations in which worker was completely nested within task, with multiple observation days for each worker. For the iron workers, data on 3 job tasks, 4 di!erent workers per task, and 2 days of observation for each worker were available. For the carpenters, 3 job tasks, 3 workers per task and 2 days of observation for each worker were avail-

435

able. The percent of total variance explained by task, worker and error was calculated for each trade with a nested analysis of variance (ANOVA) (see Ott (1988) for a detailed description of nested ANOVA).

3. Results 3.1. Ironworkers reinforcing concrete 3.1.1. Tasks performed Iron workers were observed reinforcing concrete during the cement concrete masonry operation of the Structures construction stage. Five major concrete reinforcement job tasks were observed: ground-level rebar construction, wall rebar construction, ventilation rebar construction, preparation work and supervising. Ground-level rebar construction involved the reinforcement of the subbase surface of the highway. Wall rebar construction involved the reinforcement of vertical walls that divided the highway into two parts (1 for each direction of tra$c). For ventilation rebar construction, rebar was assembled for the concrete structures surrounding the tunnel's ventilation ducts. Preparation work included moving materials, guiding crane loads, erecting sca!olding along rebar walls and clearing debris. Common activities for all tasks except supervising included tying and manual handling of the rebar. A total of 2128 observations were made on 17 iron workers performing concrete reinforcement (Table 1). The most common tasks were

Table 1 Number of iron workers observed and observations made on concrete reinforcement tasks during the concrete masonry operation in the Structures stage of highway construction work (Boston, MA, 1994) Job task

Number of workers

Frequency

Percent

Ground-level rebar construction Wall rebar construction Ventilation rebar construction Preparation work Supervising Total

17 15 11 16 5 17


623 516 172 651 166 2128

29.2 24.2 8.1 30.6 7.8 100


Iron Workers were observed performing more than one task. 121 of the 166 observations were made on 1 worker (the foreman).

436

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

preparation work (30.6%), ground level rebar construction (29.2%) and wall rebar construction (24.2%). Ventilation rebar construction and supervising were each observed less than 10% of the time. 3.1.2. Manual materials handling during concrete reinforcement MMH activities were observed frequently in all job tasks except supervising. MMH was most frequent during ground-level rebar construction (19%) and preparation work (14%). The percentage of time that workers performed MMH activities and lifted loads, speci"cally, di!ered signi"cantly among tasks (p"0.001). The frequency of the remaining MMH activities (i.e., lower, carry, push or pull) did not di!er signi"cantly among tasks (p"0.101, 0.154, and 0.259, respectively). When supervising was omitted from the analysis, the percent of time workers devoted to manual handling activities and lifting among the more physically demanding tasks remained signi"cantly di!erent among tasks (p"0.022) (Fig. 2). Loads (tools and materials) were handled more than 40% of the time for each job task except supervising, although most of the loads were light

((4.5 kg). With load categories above 13.6 kg combined because of sparse data, the frequency of the recorded load categories di!ered signi"cantly among tasks, even when supervising was excluded (p"0.001). Overall, loads of at least 13.5 kg were handled in 6% of the observations and were most frequently observed during ground level rebar construction (8%) and wall rebar construction (7%) (Fig. 3). Rebar was handled 93% of the time in heavy MMH; other loads handled included plywood, power saw and cable. Workers were frequently observed in awkward trunk postures during heavy MMH; with twisting or lateral bending observed close to 40% of the time. A power grasp was most frequently used during MMH activities, and 13% of MMH activities were performed with one hand. 3.1.3. Variability in loads handled and trunk postures during concrete reinforcement The amount of variance explained by a subset of tasks (preparation work, ground level rebar construction and wall rebar construction) and workers (4 workers on 2 days) is shown in Table 2. For severely #exed trunk postures and kneeling

Fig. 2. Frequency of MMH activities by task among iron workers reinforcing concrete during the concrete masonry operation. The percent of time observed performing all MMH and lifting di!ered signi"cantly among tasks (chi-square on 4 d.o.f., p"0.001). The percent of time observed performing MMH and lifting di!ered signi"cantly among tasks, excluding supervising (chi-square on 3 d.o.f., p"0.022).

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

437

Fig. 3. Frequency of loads handled by task among iron workers reinforcing concrete during the concrete masonry operation. The percent of time loads were handled (by category) di!ered signi"cantly among tasks (chi-square on 20 d.o.f., p"0.001). The percent of time loads were handled (by category) di!ered signi"cantly among tasks, excluding supervising (chi-square on 16 d.o.f., p"0.001).

Table 2 Relative (%) contribution to total variability in selected exposures for iron workers in 3 concrete reinforcement tasks (preparation work, ground-level rebar construction and wall rebar construction): Nested ANOVA, 4 workers performing each task on 2 di!erent days Source of variance (df ) Exposure

Task (2)

Worker (9)

Error (12)

Handle at least 4.5 kg Non-neutral trunk Slightly #exed trunk Severely #exed trunk Laterally bent or twisted trunk At least one arm above shoulders Kneeling

0.0 22.2 0.0 35.0
6.7 2.0 30.7

39.0 20.6 34.4 0.0 28.4 40.1 17.0

61.0 57.2 65.6 65.0 64.8 57.9 52.3

p(0.1,
p(0.05 for Pr'F.

postures, more variance in exposure was explained by task than by di!erences among workers within task. For handling loads of at least 4.5 kg and arm

postures above shoulder height, there was more variance among workers within task than among tasks. For all variables, the error term (interpreted primarily as day-to-day di!erences in exposure) was the largest source of variance. 3.2. Laborers constructing a jacking pit for utilities relocation 3.2.1. Tasks performed Laborers were observed while building a jacking pit, a type of shaft used to allow utilities such as drainage pipes and electrical ducts to be installed beneath the ground. They performed four job tasks: top work, manual excavation, construction of pit walls and preparation work. Top work was performed outside of the pit and included cutting wood beams to speci"cations, directing a crane operator and watching the members of the crew who were in the pit. Preparation work involved setting up equipment and materials on site. Construction of the pit walls required burning holes

438

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

into steel supports that lined the pit, retrieving wood beams and attaching the wood beams to the supports. Manual excavation consisted primarily of using a jack hammer and shoveling debris into a crane's bucket. A total of 3222 observations was made on 12 laborers during the construction of two jacking pits (Table 3). Manual excavation was observed most frequently (44.0%), followed by top work (22.0%) and construction of pit walls (17.5%). Preparation

Table 3 Number of laborers observed and observations made on tasks during the jacking pit construction operation in the Utilities Relocation stage of highway construction work (Boston, MA, 1996) Job task

Number of workers

Frequency

Percent

Prepare site Top work Construct pit walls Manual excavation Miscellaneous Total

1 8 9 11 10 12

3 709 563 1416 531 3222

0.0 22.0 17.5 44.0 16.5 100

Laborers were observed performing more than one task.

work was rarely recorded (3 observations) and 16.5% of the observations involved miscellaneous activities that were not related to the tasks. These were not evaluated further. 3.2.2. Manual materials handling during jack pit construction MMH was observed about 5% of the time during the construction of the jacking pits. The percentage of time that laborers performed all MMH activities, lifted, carried and pushed or pulled di!ered signi"cantly among tasks (p" 0.001) and was lowest during manual excavation. Overall, lifting was the most common MMH activity (3.5% of the observations) and was performed most often during the construction of pit walls (Fig. 4). Loads were handled more than 35% of the time for all job tasks; this proportion di!ered signi"cantly among tasks (p"0.001). Load handling was most frequently observed during wall construction (55.4%). While MMH activities were observed less than 2% of the time during manual excavation, loads of at least 4.5 kg were handled over 10% of the time because of heavy tools (e.g., jackhammer operation) (Fig. 5). Workers handled loads of at

Fig. 4. Frequency of MMH activities by task among laborers constructing a utilities jacking pit. The percent of time observed performing MHH, lifting, carrying, and pushing or pulling di!ered signi"cantly among tasks (chi-square on 2 d.o.f., p"0.001).

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

439

Fig. 5. Frequency of loads handled by task among laborers constructing a utilities jacking pit: Highway construction, Boston, MA, 1996. The percent of time loads were handled (by category) di!ered signi"cantly among job tasks (chi-square on 10 d.o.f., p"0.001).

least 13.5 kg in 7.6% of the observations, most often during wall construction. Approximately one-half of the heavy MMH activities involved moving wood beams, while the remainder involved handling tools and materials such as a jack hammer and a chain saw. 3.3. Carpenters constructing forms 3.3.1. Tasks performed Carpenters were observed constructing forms during the cement concrete masonry operation of structures construction. Eight tasks were observed: form building, ventilation form assembly, material moving, form erection, form stripping, sawing/cutting, housekeeping and supervising. Building forms involved connecting sheets and boards together with nails, clamps and bolts. Ventilation form assembly required the connecting parts of ventilation forms together with fasteners and sealing the joints between parts with caulking. Material moving was de"ned as the set of activities performed when forms or form materials were moved from one site location to another. Erecting forms involved working with a crane operator to guide the form into

Table 4 Number of carpenters observed and observations made on form construction tasks during the concrete masonry operation in the Structures stage of highway construction work (Boston, MA, 1994) Task

Number of carpenters

Frequency

Percent

Sawing/Cutting Building Forms Assembling Plastic Forms Material Moving Housekeeping Supervising Erecting Forms Stripping Forms Total

8 8 4

39 187 4

6.0 28.8 0.6

8 6 1 7 5 8

53 23 1 334 8 649

8.2 3.5 0.2 51.5 1.2 100

Carpenters were observed performing more than one task.

place and attaching the completed forms to iron rebar and concrete structures. Six hundred forty-nine observations were made on 8 carpenters building forms on-site (Table 4). The most common tasks observed were form

440

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

erection (51% of observations) and form building (29%). Ventilation form assembly, supervising and form stripping each made up less than 2% of the observations. Six tasks had fewer than 60 observations, limiting our ability to characterize the exposures in these tasks. 3.3.2. Manual materials handling during form work MMH activities were observed less than 10% of the time in all tasks except material moving (Fig. 6). Light loads (less than 6.8 kg) were handled often in all job tasks (Fig. 7). Di!erences in the frequency of all MMH activities could not be tested due to sparse data. 3.3.3. Sources of variability for loads handled and trunk postures Sources of variability in exposure frequency were explored in a subset of the data for 3 carpentry tasks (form erection, building forms on site and building forms in a shop) that each included 3 workers on 2 days (Table 5). For handling loads of at least 6.8 kg and laterally bent or twisted trunk postures, task explained more of the variance than di!erences among workers perform-

ing the same task. For kneeling, di!erences among workers were more important than di!erences among tasks. For all exposures but kneeling, the greatest source of variance was `errora interpreted primarily as day-to-day variability of exposure among workers.

4. Discussion 4.1. MMH in highway construction work The frequency of MMH and heavy MMH activities di!ered considerably among the trades and construction operations studied here. MMH was most often observed during concrete reinforcement work performed by iron workers. Ground-level rebar construction, one of the more common concrete reinforcement tasks, had the highest frequency of MMH activities and heavy MMH activities. Common characteristics of heavy MMH activities during concrete reinforcement were the handling of rebar and work in laterally bent or twisted trunk postures. These results suggest that controls designed to reduce physical load should be

Fig. 6. Frequency of MMH activities by job task among carpenters in form work during the concrete masonry operation.

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

441

Fig. 7. Frequency of loads handled by task among carpenters in form work during the concrete masonry operation.

Table 5 Relative (%) contribution to total variability in selected exposures for 3 carpentry job tasks (form erection, building forms on site and building forms in a shop): Nested ANOVA, 3 workers performing each task on 2 di!erent days (modi"ed from Punnett and Paquet, 1996) Source of variance (df) Exposure

Task (2)

Worker (6)

Error (9)

Handle at least 6.8 kg Non-neutral trunk Slightly #exed trunk Severely #exed trunk Laterally bent or twisted trunk Flexed and twisted trunk At least one arm above shoulders Kneeling

17.5 24.6 0.0 16.4 49.5
18.5 23.0 0.0

0.0 3.6 21.2 0.0 0.0 0.0 20.8 85.9

82.5 71.8 78.8 83.6 50.5 81.5 56.2 14.1

p(0.1,
p(0.05 for Pr'F.

directed at reducing the frequency of heavy rebar handling and use of awkward trunk postures, particularly during ground level rebar construction. More sophisticated measurement techniques (e.g.,

biomechanical modeling using data from electrogoniometry, force transducers and electromyography) could be used for more detailed task analysis and the evaluation of controls. For laborers performing jacking pit construction, heavy MMH was most frequently observed during the construction of the pit walls and usually involved the handling of wooden boards. The boards were approximately 25 cm wide and could not be lifted with a power grasp. Laterally bent or twisted trunk postures were not observed during heavy MMH in the jacking pit. For this operation, improvements could be directed toward improving the design of the boards. For example, lighter materials or a better hand-to-board coupling (e.g., handles or a groove in the boards) could reduce the physical loading. Heavy MMH was rarely observed during the carpentry work, but six of the eight construction tasks were observed infrequently, and heavy MMH activities in these tasks may have been missed. For the two tasks with over 180 observations (building and erecting forms) MMH activities were observed less than 5% of the time, and MMH activities

442

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

involving at least 22.7 kg were not observed. It appears that the carpentry tasks in this operation require very little heavy MMH. Instead, these tasks required frequent handling of light loads. 4.2. Advantages and limitations of PATH In order best to understand the physical requirements of manual handling activities and to provide information aiding the appropriate design of controls, the important characteristics of MMH should be measured simultaneously. While no one evaluation tool can characterize all attributes in great detail, PATH does provide some quantitative information about the percent of time devoted to material handling, in addition to the weight of the object handled, body postures, number of hands and grasp type used during MMH. Additionally, information about the individual (e.g., weight, height and age) and the construction environment (e.g., temperature, weather) is recorded in the form of "eld notes. While this additional information was not used in the evaluations described in this paper because of small sample sizes, it could be used to provide a more complete evaluation of the physical workload in construction. By using PATH, we have traded some precision in our measurement of a single individual's exposure for the ability to evaluate multiple exposures to multiple workers over long periods. The collection of such data is necessary to obtain a reliable estimates of exposures when they vary greatly among workers and days. In these cases, precise measurements made on a limited number of people over a short duration may not accurately re#ect the true long term exposure. Observational methods have several limitations. First, exposure information is relatively crude, usually allowing only broad categories of loads handled, posture, and grasp type to be measured reliably. When weights are unknown to the observers, loads handled by the workers must be estimated. Work sampling does not allow the actual frequency of MMH activities or other exposures to be recorded (as motions or exertions per unit time). Because observations are made intermittently rather than continuously, the sequence of

activities is lost and rare events such as extremely heavy manual handling may be missed. When information about the dynamics of a manual handling activities (e.g., whether a `jerka was used during the lift) more detailed methods of exposure assessment must be used. A reasonable approach for a complete evaluation of MMH exposures in construction might involve PATH for the gross characterization of MMH and more detailed methods (e.g., use of bioinstrumentation) to quantify speci"c exposures more precisely for selected tasks. 4.3. Considerations for exposure assessment While great e!ort was made to obtain reliable estimates of long term MMH exposures for this study group by randomly sampling as many workers as possible on multiple days spanning periods of several weeks, the exact frequency of MMH activities, loads handled and postures may not replicate the population exposures for the tasks described in this paper. This is possible because the true worker-to-worker and day-to-day variance of exposure within each of the tasks was not known at the time of data collection. Burdorf (1995) demonstrated how the variability of exposures within and between workers can a!ect the reliability of exposure measures and attenuate the true odds ratio for exposure and health outcome in epidemiologic studies. As the variability in exposure among workers and days increases, the number of workers and days required for the evaluation increases for a reliable estimate of exposure. To improve exposure assessment, Burdorf (1995) suggested piloting the exposure assessment e!ort to estimate sources of variability, as well as balancing the number of workers studied and the number of repeated measurements for each worker. At least two earlier studies have examined sources of variance in the frequency of trunk postures in other industries (Burdorf et al., 1994; van der Beek, 1995). Both studies found that variability of trunk postures within-shift for individual workers was greater than the inter-worker and inter-day variability. Although these results may not apply to jobs with di!erent temporal patterns of exposure, especially for work without cycles or with cycles longer

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

than a day, they do demonstrate the importance of sampling throughout as much of a worker's 8-h shift as possible to obtain a reliable estimation of daily back loading. Additionally, di!erent types of job tasks performed during a shift may also explain at least some of their observed large within-shift exposure variance. In this study, the relative importance of task, worker and day on the variance of loads handled and postures was examined for some concrete reinforcement and carpentry tasks. Observations were made over a large percentage of the each 8-h shift. Task was an important source of variance in exposures to some non-neutral trunk postures in both trades. Task was also an important source of variance for kneeling postures among iron workers and for load handling among carpenters. In other cases, di!erences among workers were an important source of exposure variance. These results suggest that, for at least some ergonomic exposures, task is an important predictor of exposure, but its share of the variance di!ers among the construction operations and ergonomic exposures. Understanding how exposures vary between tasks and individuals within a task has important implications for the appropriate design of controls. When exposures vary greatly between tasks and little among workers performing the same task, controls can be directed at reducing exposures for all workers performing a potentially hazardous task. When exposures vary little between tasks but are largely explained by di!erences among workers, work practices and other individual factors (e.g., age or anthropometry) might be considered when developing controls. For highway construction, the between-worker and within-worker exposure variability depended on the construction job tasks studied and the exposure of interest. Therefore, many workers performing the same tasks on many days may be needed for reliable estimates of MMH and other ergonomic exposures. In the future, determining e$cient sampling strategies (minimizing the number of required workers and observation time needed) that provide reliable group estimates of exposure within tasks will be an important topic for the assessment of ergonomic exposures in construction and other non-routinized work.

443

5. Conclusions The methods described in this paper have permitted a systematic objective assessment of MMH requirements for highway construction work. The frequency of MMH exposures di!ered signi"cantly among job tasks for each trade that was evaluated in this study, demonstrating the importance of task-based exposure measures for evaluating ergonomic exposures. Long-term measurement of exposure on multiple workers is desired for a reliable estimate of exposure for construction tasks. After the long-term assessments are made, more detailed evaluation methods may be used to characterize exposure for high exposure tasks more precisely. More research is warranted to develop e!ective exposure assessment strategies for construction and other non-routinized work.

Acknowledgements This work was funded by GrantC U02/CCU312014-02 from the National Institute for Occupational Safety and Health through the Center to Protect Workers' Rights and by ContractC 200-94-2861 from the National Institute for Occupational Safety and Health. Susan Moir assisted with site access and the development of the taxonomy for highway construction work. Michael Grasso, Joel Garrett, Diane Lee, William Rodwell, Trevor Schell and Hellen Wellman collected data. The authors wish to thank the members of the International Association of Bridge, Structural and Ornamental Iron Workers (Local 7), the United Brotherhood of Carpenters and Joiners of North America (Local 218) and the Laborers' International Union of North America (Local 223) who participated in the study.

References Buchholz, B., Paquet, V., Punnett, L., Lee, D., Moir, S., 1996. PATH:: a work sampling-based approach to ergonomic job analysis for construction and other non-repetitive work. Applied Ergonomics 27 (3), 177}187. Burdorf, A., 1995. Reducing random measurement error in assessing postural load on the back in epidemiologic surveys.

444

V. Paquet et al. / International Journal of Industrial Ergonomics 24 (1999) 431}444

Scandinavian Journal of Work, Environment and Health 21 (1), 15}23. Burdorf, A., Verburgh, A., Elders, L., 1994. Time-dependent variation in back load of workers in a dairy factory. Occupational Hygiene 1, 199}206. Bureau of Labor Statistics, U.S. Department of Labor, 1994. Survey of Occupational Injuries and Illnesses, 1992. Summary 94}3. Bureau of Labor Statistics, U.S. Department of Labor, 1996. Characteristics of Injuries and Illnesses Resulting from absence from work, 1994. May 8, USDL-96-163. Cha$n, D., 1969. A computerized biomechanical model: development of and use in studying gross body actions. Journal of Biomechanics 2, 429}441. Ciriello, V., Snook, S., 1983. A study of size, distance, height, and frequency e!ects on manual handling tasks. Human Factors 25 (5), 473}483. Ciriello, V., Snook, S., Blick, A., Wilkinson, P., 1990. The e!ects of task duration on psychophysically-determined maximum acceptable weights and forces. Ergonomics 33 (2), 187}200. Ciriello, V., Snook, S., Hughes, G., 1993. Further studies of psychophysically determined maximum acceptable weights and forces. Human Factors 35 (1), 175}186. Garg, A., Cha$n, D., Herrin, G., 1978. Prediction of metabolic rates for manual materials handling jobs. American Industrial Hygiene Association Journal 39, 661}664. Kemmlert, K., 1995. A method assigned for the identi"cation of ergonomic hazards*PLIBEL. Applied Ergonomics 26 (3), 199}211.

McGill, S., 1992. A myoelectrically based dynamic three-dimensional model to predict loads on the lumbar spine tissues during lateral bending. Journal of Biomechanics 25 (4), 395}414. NIOSH, 1981. Work practices guide for manual lifting. Department of Health and Human Services (DHHS), National Institute for Occupational Safety and Health (NIOSH), Publication No. 81}122. Niskanen, T., Saarsalmi, O., 1983. Accident analysis in the construction of buildings. Journal of Occupational Accidents 5, 89}98. Ott, L., 1988. An Introduction to Statistical Methods and Data Analysis, third ed. PWS-Kent Publishing, Boston, MA, pp. 781}790. Punnett, L., Paquet, V., 1996. Ergonomic exposures to carpenters and carpentry laborers in tunnel construction. In: Mital, A., et al. (Eds.), Advances in Occupational Ergonomics and Safety I. Proceedings of the Eleventh International Occupational Ergonomics and Safety Conference, IOS Press, Amsterdam (revised paper available from authors). Snook, S., Ciriello, V., 1991. The design of manual handling tasks: revised tables of maximum acceptable weights and forces. Ergonomics 21, 963}985. van der Beek, A., Kuiper, J., Dawson, M., Burdorf, A, Bongers, P., Frings-Dresen, M., 1995. Sources of variance in exposure to non-neutral trunk postures in varying work situations. Scandinavian Journal of Work, Environment and Health 21, 215}222. Waters, T., Putz-Anderson, V., Garg, A., Fine, L., 1993. Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics 36, 749}776.