Validity of fixed-interval observations for postural assessment in construction work
Introduction
In a review of over 600 epidemiologic studies, the National Institute for Occupational Safety and Health (NIOSH) reported that there was strong evidence for causal relationships between awkward postures and neck/shoulder disorders, a combination of physical ergonomic exposures and upper extremity disorders, and lifting and whole body vibration and back disorders (US Department of Health and Human Services, 1997). In spite of such findings, some remain skeptical about the work-relatedness of musculoskeletal disorders. This is mainly because of conflicting study results and major flaws in many studies (e.g., lack of control for confounders). Poor characterization of exposures is an important factor that contributes to both of these limitations. Valid and precise measurements of numerous independent variables are needed to help clarify the relationships between occupational exposures and musculoskeletal health outcomes, as well as to document changes in exposure associated with the introducing workplace interventions.
Musculoskeletal problems are common among workers in the construction industry. In 1992 the construction industry had the highest frequency of work-related injuries and illness among all economic sectors, with 13.1 cases per 100 full-time workers (US Department of Labor, Bureau of Labor Statistics, 1994a). During this year, there were over 10,000 new cases of non-fatal occupational illness reported in the construction industry, 21% of which were associated with repeated trauma (US Department of Labor, Bureau of Labor Statistics, 1994b). Since this time, there has been a steady decrease in the incidence rate of work-related injuries and illnesses among construction workers, but in 1997 the incidence rate was second only to manufacturing (US Department of Labor, Bureau of Labor Statistics, 1998).
Epidemiologic studies have demonstrated associations between construction work and musculoskeletal disorders of the back and the upper and lower extremities (e.g., Burkhart et al., 1993; Damlund et al., 1982; Holmstrom et al., 1993; Latza et al., 2000; Riihimäki et al., 1990; Stenlund et al., 1993). However, exposure data for most of these studies are limited to trade or job title and provide little information about which work-related risk factors contribute most to the excess risk of musculoskeletal problems in this industry.
Construction workers having the same job title may have different levels of exposure to physical ergonomic stressors because the duration and distribution of job tasks for individual workers may vary among workers and from day to day. The lack of quantitative exposure data is, in part, because construction work generally requires workers to perform multiple tasks with long variable work cycles, making cycle-based ergonomic assessment methods and laboratory simulations of job tasks impractical. Exposure assessment methods designed to quantify the long-term physical requirements are needed to improve epidemiologic and intervention research in construction and other similar types of non-routinized types of work.
Work sampling is a methodology that has been used by industrial engineers for approximately 60 years to quantify the proportion of time that workers or machines devote to different work activities (Pape, 1992). Modified forms of work sampling have been applied to studies of jobs where cycle-based ergonomic exposure assessment methods could not be used easily (e.g., Karhu et al., 1981; Mattila et al., 1993; Ryan, 1989; Wickstrom et al., 1985). For these types of assessments, observations about working postures or manual handling are usually made at fixed intervals throughout a representative work period. Observations are often made on categories of exposure, and the proportion of time recorded for each exposure category is the ratio of the number of observations recorded for the category to the total number of observations. PATH (Posture, Activities, Tools and Handling), a fixed-interval observational approach, has been developed specifically to characterize the proportion of time workers spend in awkward postures, handling loads, and performing manual materials handling (MMH) and other activities within job tasks, as well as the frequency distribution of job tasks during non-routinized types of work (Buchholz et al., 1996).
Observational methods, such as PATH, offer the advantages of providing data on multiple exposures collected simultaneously, over long time periods, and can be used with little disruption to the work. There is, however, some uncertainty about the validity of observations for postural assessment. Some studies have found relatively poor-to-moderate agreement when fixed-interval observational approaches were compared to more sophisticated measurement systems (e.g., Burdorf et al., 1992; De Looze et al., 1994).
The objective of this study was to examine the validity of PATH and a simplified version of PATH for the assessment of trunk, shoulder and knee postures in construction work. Comparisons were made between both observational approaches and discrete reference measurements, between observational and continuous-direct measurements, and between PATH and simplified PATH.
Section snippets
Study site and subjects
The study took place at a laborers’ training grounds that closely resembled a typical construction site, equipped with a construction trench, utilities pit, truck, asphalt and dirt surfaces, and a variety of building materials (e.g., bricks, blocks, boards). Five male college students were brought to the study site to simulate 3 of 6 construction job tasks. The order in which subjects performed each of the 3 tasks was randomly assigned.
Job tasks
The six construction job tasks were designed to facilitate
Shoulder postures
There were 463 electronic inclinometer, 461 simplified PATH and 457 PATH fixed-interval recordings made on shoulder postures. There were no significant differences in the frequency of exposure to the three shoulder posture categories among the three methods (chi square on 4 d.o.f.=3.1, p=0.54). The frequency of the fixed-interval measurements also closely approximated the continuous measurements of shoulder postures (Fig. 1).
Agreement in the frequency of arm-posture categories was extremely
Validity of fixed-interval observations
In approximately 225 min of work and six simulated construction tasks, the inter-method agreement among the fixed-interval reference and fixed-interval observational measurements of posture was generally high. The strongest agreement among the three methods was found in the categorization of shoulder postures. Analysis of the shoulder-posture frequency among methods demonstrated that the observational data closely approximated the reference measurements, with the largest discrepancies for the
Conclusions
Fixed-interval observations made in real-time provided frequency estimates of shoulder, trunk and some leg posture categories closely approximated measurements obtained with electronic instruments or with video analysis. The largest discrepancies among the reference and observational methods were in the evaluation of knee flexion and trunk lateral bending and twisting. For the observational approaches used in this study, it appears that valid estimates of exposure frequency to arm postures at
Acknowledgements
This work was supported by a grant provided by the National Institute for Occupational Safety and Health through the Center to Protect Workers’ Rights (Grant# U02/CCU312014). Parts of the data collection were also supported by the Department of Transportation and the Laborers’ Health and Safety Fund of North America (Grant# DTFH61-93-X-00024). Gary Davis, Ted DesMarais, Martin Forde, Niki Howard, Jung-Soon Park, and Trevor Schell assisted with data collection. The authors also acknowledge the
References (34)
- et al.
PATH: a work sampling-based approach to ergonomic job analysis for construction and other non-repetitive work
Applied Ergon.
(1996) - et al.
Measurement of trunk bending during work by direct observation and continuous measurement
Applied Ergon.
(1992) - et al.
High agreement but low kappaI
the problems of two paradoxes. J. Clin. Epidemiol.
(1990) - et al.
A portable ergonomic observation method (PEO) for computerized on-line recording of postures and manual handling
Appl. Ergon.
(1995) - et al.
Correcting working postures in industry: a practical method for analysis
Appl. Ergon.
(1977) - et al.
Validation of the Portable Ergonomic Observation (PEO) method using optoelectronic and video recordings
Appl. Ergon.
(1997) - et al.
Accuracy of a three-dimensional lumbar motion monitor for recording dynamic trunk motion characteristics
Int. J. Ind. Ergon.
(1992) - et al.
Analysis of working postures in hammering tasks on building construction sites using the computerized OWAS method
Appl. Ergon.
(1993) - et al.
Working postures and activities of lorry drivers: a reliability study of on-site observation and recording on a pocket computer
Appl. Ergon.
(1992) - et al.
Job tasks, potential exposures, and health risks of laborers employed in the construction industry
Am. J. Ind. Med.
(1993)
Low back pain and early retirement among Danish semi-skilled construction workers
Scand. J. Work, Environ. Health
The validity of visual observation to assess posture in a laboratory-simulated, manual material handling task
Ergonomics
Healthy lower backs in the construction industry in Sweden
Work & Stress
Observing working postures in industry: examples of OWAS application
Appl. Ergon.
Cited by (75)
Sensor-based computational approach to preventing back injuries in construction workers
2021, Automation in ConstructionAutomated postural ergonomic risk assessment using vision-based posture classification
2021, Automation in Construction