- Email: [email protected]
EMG estimated mean, peak, and cumulative spinal compression of workers in five heavy industries
International Journal of Industrial Ergonomics 40 (2010) 448e454
Contents lists available at ScienceDirect
International Journal of Industrial Ergonomics journal homepage: www.elsevier.com/locate/ergon
EMG estimated mean, peak, and cumulative spinal compression of workers in five heavy industries Catherine Trask a, e, *, Kay Teschke b, e, Jim Morrison c, Pete Johnson d, Judy Village e, Mieke Koehoorn b, e a
Centre for Musculoskeletal Research, University of Gävle, SE e 801 76 Gävle, Sweden School of Population and Public Health, University of British Columbia, 5804 Fairview Avenue Vancouver, BC V6T 1Z3 Canada c School of Kinesiology, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6 Canada d Department of Environmental and Occupational Health Sciences, University of Washington Box 357234- 4225 Roosevelt Way NE Suite 100 Seattle, 98195e7234 WA, USA e School of Environmental Health, University of British Columbia 375-2206 East Mall Vancouver, BC V6T 1Z3 Canada b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 June 2009 Received in revised form 12 January 2010 Accepted 26 February 2010 Available online 29 March 2010
The goal of this study was to explore the use of compression-normalized electromyography (CNEMG) to estimate mean, peak, and cumulative loading of the low back in workers of five heavy industries and to compare the estimates to the NIOSH guidelines. Full-shift (5.5e10.3 h) EMG measurements were collected from 105 workers and transformed into units of low back compressive force (Newtons). The mean, peak, and cumulative CNEMG as well as the percentage of work time spent above 3400 N and 6800 N thresholds were calculated. Mean CNEMG (sd) was 1564 N (796), peak was 2721 (1545), and cumulative was 38 MN s (22). Mean time spent above the NIOSH threshold of 3400 N was on average 6.3% of shift, while mean time spent above the 6400 N threshold was around 1%. CNEMG allowed the feasible investigation of tasks and jobs that would be virtually impossible with more advanced biomechanical methods and represents a more objective measure of exposure than observation or self-report. CNEMG is a relatively new method with methodological limitations, however, CNEMG's strength may be as an objective, relative measure of exposure in studies where low back EMG is collected in a relatively systematic and structured manner. Ó 2010 Elsevier B.V. All rights reserved.
Keywords: Electromyography Back injury Exposure assessment Construction Compression-normalized EMG
1. Introduction In the Canadian province of British Columbia, back injuries account for approximately 25% of all workers' compensation claims, 23% of workdays lost, and 20% of total claims costs. Onequarter of back injury claims are made by workers in five heavy industries: construction, forestry, transportation, warehousing and wood products (WorkSafeBC, 2005). Work tasks and activities in heavy industry often involve considerable time in awkward postures and manual materials handling (MMH) (Teschke et al., 2009), and these types of exposures are thought to be related to musculoskeletal injuries (Keyserling, 2000). Multiple lab investigations have demonstrated that nonneutral postures and MMH increase muscle activity, which is in turn related to spinal moments and spinal compression (Marras
* Corresponding author. Centre for Musculoskeletal Research, University of Gävle, SE e 801 76 Gävle, Sweden. Tel.: þ46 604 221 0553. E-mail addresses: [email protected] (C. Trask), [email protected] (K. Teschke), [email protected] (J. Morrison), [email protected] (P. Johnson), [email protected] (J. Village), [email protected] (M. Koehoorn). 0169-8141/$ e see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ergon.2010.02.006
and Sommerich, 1991; Potvin, 2008;). These studies have yielded many excellent biomechanical modeling methods that account for asymmetrical, complex motion and dynamic segment/load inertias (Potvin, 2008). Unfortunately, heavy industry presents extensive measurement challenges to researchers (Trask et al., 2007); the more complex lab-based methods are too cumbersome to be applicable for many work areas or tasks, and too expensive to be used for a large number of workers or a long sampling duration. Although with some loss of precision when compared to labbased methods, portable electromyography (EMG) may provide a practical alternative to complex biomechanical methods because it can be used to estimate the low back compressive forces throughout the diverse activities of a heavy industry work shift. EMG of trunk muscles during work tasks has been proposed as a method for estimating spinal loads using a method called compression-normalized EMG (CNEMG) (Mientjes et al., 1999). CNEMG is a direct measurement method which allows for estimation of spinal compression in workplaces where traditional biomechanical methods, such as video analysis or motion-capture systems, are not feasible. Bao et al. (2009) call measurement of
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
working forces “both important and challenging”, emphasizing that capturing the important forces is a critical component of measurement (Bao et al., 2009). One alternative to measuring only a few forces is to employ a method that allows measurement of a full-shift. EMG has advantages in that it allows workers to move freely throughout their working environments with minimal interruption, and has been used to assess muscle activity in apple pickers working in the field (Earle-Richardson et al., 2008) and estimate all day low back compressive forces in aides working in long term care (Village et al., 2005). A recent review shows mean, peak, and cumulative spinal compression to be related to low back injury (da Costa, 2009). Peak and cumulative spinal loads have been shown to be independently related to back injury, (Village et al., 2005; Norman et al., 1998; Kumar, 1990; Seidler et al., 2001; Kerr et al., 2001; Nelson and Hughes, 2009). Village et al. (2005) used full-shift EMG to estimate peak and cumulative spinal loading in 32 care aides and found a correlation between these measures and musculoskeletal injury rates as well as total tasks observed (Village et al., 2005). For example, a worker could have moderate peak loads that are sustained throughout the day, such that the total load on the back could surpass that of a worker with high but infrequent peak loads. Knowledge about cumulative load as an exposure is still developing, and it is not clear how it relates to classical peak loading measures. Understanding the relative contribution of peak and cumulative loads to overall occupational exposure in heavy industries can help target future exposure assessment and intervention strategies. Spinal loading studies have typically assessed exposure in cyclical manufacturing and assembly tasks (Norman et al., 1998; Kerr et al., 2001) or via self-report (Kumar, 1990; Seidler et al., 2001). However, comparatively little has been done on non-cyclical tasks, such as those found in resource industries and construction, perhaps because the most precise and accurate exposure assessment methods are nearly impossible to apply in these industries. Self-report and observation methods can be used to assess exposure, but these methods are more subjective and less precise than direct measurement (Winkel and Mathiassen, 1994; Burdorf, 1992; Burdorf et al., 1997; Wells et al., 1997). EMG represents a compromise between lab-based biomechanical methods and subjective methods; it is an objective, direct measurement exposure assessment method that can be used to attain full-shift estimates of non-cyclical activities in heavy industries. EMG also provides a unique opportunity to estimate mean, peak, and cumulative exposures concurrently, in spinal compression units that can be used as a ‘common metric’ to compare to other research results and to exposure guidelines (Wells et al., 1997). The goal of this study was to explore using full-shift EMG data to calculate CNEMG estimates of mean, peak, and cumulative spinal compression in workers in five heavy industries, and to interpret these estimated exposures in light of existing studies and guidelines. 2. Methods 2.1. Worker recruitment and sampling strategy This analysis was undertaken as part of a larger study of low back injury risk factor measurement in the heavy industrial sectors of construction, forestry, transportation, wood products, and warehousing in the Canadian province of British Columbia. A random sample of workers with a lost-time workers' compensation claim for a back injury in 2001 were contacted by phone and invited to participate; 75% of workers contacted agreed to participate. Employers were then contacted for permission to visit the worksite and make full-shift exposure measurements and to allow
449
recruitment of additional co-workers; 73% of employers gave permission. Prior to measurement, researchers visited each participating worksite to recruit up to 4 co-workers of the original WCB-selected worker, bringing the total sample to 126 individuals. All workers had production jobs without modified duties. The resulting sample represented 50 different worksites and a wide variety of job titles, tasks and task patterns. The work ranged from static, continuous tasks such as delivery truck driving, to dynamic but monotonous tasks such as repeatedly stacking wood off a conveyor belt, to dynamic and varied work such as maintenance work with no repeated tasks within a day. 2.2. EMG data collection and processing Full-shift (5.5e10.3 h, excluding breaks) EMG measurements were made using a portable data collection system with on-board memory (ME3000P4/ME3000P8, Mega Electronics, Finland) and disposable AgeAgCl electrodes (Blue Sensor N-00-S, Ambu, Denmark). Electrodes were placed bilaterally over the erector spinae at approximately the level of L4, with a 20 mm inter-electrode spacing and a ground electrode and preamplifier placed on the posterior aspect of the iliac crest. Signals were collected at 1000 Hz and filtered internally using an 8e500 Hz band pass filter. Rootmean-square values were data-logged at 10 Hz. During work breaks, data from the portable system were downloaded to a laptop computer. After data collection, EMG signals were imported to a customdeveloped LABVIEW program. All shift-long EMG measures were visually inspected for signal drift and noise spikes. Observation records were also consulted to check for notes about worker comments on shifting of electrodes. Data were removed where the observation records or workers comments indicated there were shifting electrodes, observable signal drift or suspicious noise spikes in the data. 2.3. Observation and posture measurements CNEMG methodology has been shown to be less accurate when assessing non-symmetrical tasks i.e. lateral bending or twisting (Mientjes et al., 1999). To investigate, Concurrent observation and trunk inclinometer data were collected to investigate the extent of non-symmetrical postures during measured workshifts. Oneminute sampling observations of trunk postures were made over the whole work shift by trained researchers as described in Village et al. (2009). A portable, data-logging torso-mounted inclinometer (VC-323, Microstrain Inc, Williston, VT, USA) was also used to collect shift-long data as described in Teschke et al. (2009). 2.4. Transformation to spinal compression: compressionnormalized EMG EMG voltage was transformed into units of spinal compression force (Newtons) using a linear calibration equation based on submaximal reference calibration postures performed during work time at the beginning and end of the shift (approximately 1 h for set-up and take-down). The first set of reference contractions involved standing upright with the hands by the sides. The second set of reference contractions involved static 45 forward trunk flexion while holding an 11.5 kg weight directly below the shoulders. The reference contractions were performed twice for 5 s each time at the beginning of each shift. As reported by Jackson et al. (2009) methodological variance due to EMG calibration in the trapezius is less than 5%, and collecting more than two repeated reference contractions over two days yields only marginal improvements in precision. Although the Jackson et al. study was
450
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
performed on the trapezius in a lab environment, it supports the collection of two repeated reference contractions. For each of the reference positions, spinal compression was estimated using a computer-based link-segment model (Ergowatch 4D WATBAK, Waterloo University, Canada). The estimated spinal compression (in Newtons) and the muscle activity (in uV) for each position were used to generate a linear calibration equation. This process yielded a continuous estimate of “compression-normalized EMG” (CNEMG) for the whole workday. This is similar to the compression normalization methods used by Mientjes et al. (1999) and Village et al. (2005). 2.5. Investigation of potential fatigue effect of testing over a full work shift Given that fatigue is known to impact EMG, the data were analyzed to determine whether the EMG amplitude changed over the worker's shift. Four reference contractions were collected at the beginning and end of each worker's shift/measurement session. The reference contractions included: standing upright, standing with the trunk flexed at 45 , and standing at 45 with an 11.5 kg load and standing at 60 with an 11.5 kg load. There were two repetitions for each reference contraction, totaling 16 reference contractions per subject (2 calibration sessions 4 reference conditions 2 repeats per subject). Pairs of measurements were matched for subject and type of reference contraction, and paired t-tests used to calculate mean differences between pre- and post-shift measurements. Mixed effect models (using subject and measurement session as random effect terms) were used to estimate the components of variance in the EMG measures due to pre- versus post-shift timing, reference task, subject, and measurement session. A ‘measurement session’ refers to the unique electrode/skin interface for each shift measurement per worker. 2.6. Statistical analysis and comparison to guidelines Three EMG exposure metrics were calculated for each individual's work shift data: the arithmetic mean, the 90th percentile, and cumulative exposure. The arithmetic mean is a measure of central tendency in the intensity of exposure. The 90th percentile was included as an estimate of ‘peak’ intensity of spinal load as in previous studies (Mientjes et al., 1999; Jonsson, 1988; Mathiassen et al., 2002; Nordander et al., 2004; Moller et al., 2004). Cumulative exposure represents a combination of the intensity, duration, and frequency of exposure combined into a single ‘daily dose’. It was calculated as the sum of instantaneous values throughout the shift, expressed as MN sec (MegaNetwton-seconds) of cumulative EMG activity. Since shift lengths varied in this sample, cumulative spinal loads were calculated without normalizing to a standard day length to reflect the variation in daily dose. Standardizing cumulative exposure to a constant shift length would generate a rank order of exposures identical to that of the mean exposures. The mean and peak CNEMG levels were compared to levels approximating the NIOSH compression action limit guideline (3400 N), and an upper limit of twice the guideline (6800 N) (Waters et al., 1993). In addition, the percentage of work time spent above each threshold was calculated for each work shift. There are currently no institution-endorsed guidelines for cumulative compression limits analogous to the NIOSH compression guidelines. However, a case-control study (Norman et al., 1998) showed cumulative spinal compression is a significant and independent predictor of back pain. Mean cumulative exposure was 21 MN s in cases with back pain compared to 19.5 MN s in controls, and approximately 50% of workers with cumulative compressions of
23 MN s reported back pain. Although not a guideline per se, a threshold of 21 MN s was used as a benchmark for cumulative compression. 3. Results 3.1. Study participants Full-shift EMG measurements were collected from 103 of the 126 individuals. Repeated measurements were completed for 34% of the 103 included individuals, for a total of 138 worker-days. The interval between workers' first and second measurement days ranged from 1 to 439 days (mean ¼ 93 SD ¼ 64). Demographic data on the sample are summarized in Table 1. 3.2. Observation and posture measurement of on-symmetrical postures Inclinometer results showed lateral bending greater than 12 made up less than 20% of the workday on average. Direct observation results were similar showing a mean of 19.2% (sd ¼ 15%) of the work shift spent laterally bending. Twisting measures are not captured by the inclinometer, so only observational measures are presented. Twisting postures were observed on average 18.0% (sd ¼ 17%) of the work shift. These results are described in more detail elsewhere (Teschke et al., 2009). 3.3. Comparison of pre- and post-shift EMG calibration Paired t-tests showed significant differences between 916 pairs of pre- and post-shift measurements with pre-shift values on average 3.25 mV higher than post-shift values (mean reference contraction voltage was 32 mV). However, pre- and post-shift differences explained only 3.3% of the variability in EMG, This variability was dwarfed by the variability due to the reference task (23%), and due to the subject and measurement session (44.6% combined). 3.4. Compression-normalized EMG in heavy industry Table 2 summarizes the CNEMG measures, stratified by industry and job. Because the numbers of jobs represented in this study are small compared to the numbers in each of these industries, these results represent an initial window on exposure by job. The rank order of mean, peak, and cumulative CNEMG exposures was similar across industries. Mean CNEMG ranged from 1269 N (sd ¼ 654) in transportation to 1771 N (sd ¼ 593) in construction. There was less variation in peak CNEMG, which ranged from 2338 N (1679) in transportation to 3066 N (1065) in construction. Cumulative CNEMG showed a somewhat different ranking, lowest in transportation at 32 MN s (sd ¼ 0.24), and highest in forestry at 44 MN s (sd ¼ 0.20). Transportation work shifts consistently had the lowest measured exposures and construction the highest in all metrics except cumulative exposure. The highest mean and peak compressions were in tree fallers in the forestry industry; the lowest were among heavy equipment operators in transportation.
Table 1 Demographic characteristics of study participants (N ¼ 103). Variable
Value
% Male Height in cm (sd) Weight in kg (sd) Age, in years, on sampling day (sd)
95.3% 178.1 (7.9) 85.2 (16.1) 42.2 (12)
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
451
Table 2 Compression-normalized EMG exposure metrics measured in workers of five heavy industries (N ¼ 138 workshifts).
All industries Construction Construction carpenter Construction labourer Construction supervisor Construction trades Otherd Forestry Boomman Faller Heavy Equip.mechanic Log-Machine Operators Otherd Transportation Airport ramp attendant Automotive mechanic Heavy equip op Truck driver Otherd Warehousing Forklift operator Warehouse person Wood products Cabinet maker Forklift operator Lumber puller grader Papermaking Otherd a b c d
na
Meanb (sd)
Cumulativec (sd)
10th %ileb (sd)
90 th %ileb (sd)
% time above 3400 N (sd)
% time above 6800 N (sd)
138 25 7 6 4 3 5 30 10 4 4 4 8 33 6 6 2 6 13 23 21 2 27 6 9 5 5 2
1564 1771 1591 2076 1901 1650 1664 1727 1482 2738 2159 1351 1496 1269 1193 1062 863 1084 1452 1444 1465 1223 1653 1988 1623 1473 1433 1787
38 37 33 40 39 35 38 44 37 69 52 40 39 32 23 20 27 31 39 36 36 26 39 38 45 32 36 46
641 731 710 823 666 702 808 667 659 770 840 574 592 533 451 430 500 506 581 602 610 521 691 797 667 693 621 664
2721 3066 2739 3626 3306 2882 2816 2911 2555 4789 3484 2143 2350 2911 2218 1907 1515 1939 2746 2531 2573 2092 2821 3284 2825 2391 2441 3439
6.34 5.06 4.00 2.04 2.3 7.09 8.00 7.51 5.00 23.7 13.8 4.05 0.87 3.66 2.92 0.53 0.31 2.13 6.92 3.66 6.29 1.59 10.3 14.2 12.3 4.03 5.00 32.5
0.9 0.16 0.06 0.16 0.03 0.49 0.21 0.65 0.55 0.58 0.76 1.12 0.08 0.84 0.22 0.0 0.0 0.13 1.31 0.84 0.78 0.02 2.00 3.33 3.54 0.06 0.1 5.06
(796) (593) (218) (564) (1012) (695) (693) (760) (601) (1154) (709) (310) (525) (654) (272) (344) (176) (327) (533) (755) (787) (293) (1076) (1590) (1232) (441) (556) (1483)
(22) (17) (7) (4) (26) (2) (1.4) (20) (13) (38) (10) (15) (16) (23) (13) (8) (20) (10) (18) (27) (29) (5) (21) (18) (28) (8) (14) (47)
(244) (217) (147) (216) (86) (389) (393) (225) (269) (325) (228) (94) (129) (216) (90) (93) (5) (126) (193) (193) (200) (77) (311) (457) (368) (193) (118) (298)
(1545) (1065) (308) (1103) (2056) (994) (1423) (1247) (1025) (1738) (744) (566) (694) (1247) (489) (731) (573) (658) (1197) (1574) (1640) (511) (1945) (2606) (2358) (736) (1046) (3124)
(10.9) (0.7) (2.23) (0.7) (2.13) (8.44) (12) (0.77) (3.83) (0.77) (9.05) (3.31) (0.62) (6.83) (2.18) (0.68) (0.37) (1.31) (6.47) (0.7) (8.71) (1.69) (2.0) (29.8) (24.2) (2.24) (5.74) (2.0)
(2.99) (0.04) (0.071) (0.04) (0.043) (0.78) (0.37) (0.07) (0.5) (0.07) (0.28) (1.52) (0.11) (3.4) (0.24) (0.01) (0) (0.1) (2.22) (0.30) (1.56) (0.03) (0.50) (7.41) (7.29) (0.03) (0.14) (0.5)
Sample Size. Newtons of Spinal compression (N). MegaNewton-seconds (MN s). Jobs with few measurements combined to prevent identification of individuals.
3.5. Comparisons to guidelines The mean CNEMG-estimated compression value for all industries combined (mean ¼ 1564 N) was only 42% of the NIOSH action limit guideline; each of the industry means was well below the NIOSH limit as well. Mean peak CNEMG measurements were much closer to the NIOSH values, but even the highest means (in construction) represented only 89% of the NIOSH limit. However, in every industry some measurements were above the 3400 N and 6800 N thresholds. Mean time spent above the 3400 N threshold was 6.3% for all workers (Table 2); wood products had the greatest time above this threshold at 10%, while transportation had the least at 3.7%. Wood products also had the most time spent above the 6800 N threshold at 2%; construction had the least at 0.16%. Although construction had the highest mean and peak CNEMG values, it had the lowest variability (in terms of standard deviations), and did not include the worst case jobs for time spent above NIOSH recommended limits. Compared to the cumulative compression levels from previous studies suggesting a likelihood of back pain, the cumulative compression measures were high. The mean CNEMG-estimated cumulative compression in every industry and almost every job exceeded the 21 MN s value associated with back pain reporting (Norman et al., 1998). 4. Discussion 4.1. Exposure intensity The average mean and peak CNEMG-estimated compression values are below the NIOSH-based 3400 N and 6800 N thresholds for all industries. It should be noted that the ‘peak’ values reported in this paper were 90th percentile values, so 10% of the exposure
values in a day were higher than these levels. All industries had workshifts that exceeded these levels at some point during the day, albeit for as little as 0.16% of the work shift. Clearly exposure intensity is problematic in this industrial population, as evidenced by the percentage of time spent above NIOSH thresholds. Although the percentage of time spent above a NIOSH guideline has not been explicitly compared to risk for back injury, any exceedance of the NIOSH guidelines is considered to introduce elevated risk for some workers (Waters et al., 1993) and presents an opportunity for improvement via appropriate controls. 4.2. Cumulative exposure Cumulative exposure has been shown to be a significant and independent predictor of back pain (Village et al., 2005; Norman et al., 1998/12; Kumar, 1990; Seidler et al., 2001). The mean cumulative compression estimates were substantially above a level shown to be related to back pain reporting, in both autoworkers (Norman et al., 1998) and compensated injury rates in care aides in a long term care facility(Village et al., 2005). This would seem to suggest that mean or peak exposure intensity, as well as the duration and frequency of these exposures, may be problematic in these heavy industries. An example can be seen in the higher cumulative compression values for forestry compared to construction; the difference might result from longer shift lengths in forestry where the average measured shift length (excluding breaks and the approximately 1-h study set-up period) was 7.4 h versus 5.8 h in construction. 4.3. Implications for back injury risk Workers in some jobs had high exposures in this sample. For example, the amount of time spent above the NIOSH guideline
452
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
among fallers (23.7%, sd ¼ 0.77%), heavy equipment mechanics (13.8% sd ¼ 9.05), cabinet makers (14.2% sd ¼ 29.8) and forklift operators (12.3% sd ¼ 24.2) suggests that even if these workers were young and very fit, they would be at increased risk of back injury. Since these jobs were observed by researchers during data collection to spend a large proportion of time manual materials handling and bending over in a sustained way (fallers and heavy equipment mechanics) or bending repetitively while lifting (forklift operators and cabinet makers), intervention efforts might begin with strategies to control these characteristics of the work tasks. Finding compression-normalized EMG values that exceed NIOSH guidelines in a heavy industrial working population is not surprising, especially since these workers spend substantial time performing MMH activities and in flexion postures (Teschke et al., 2009) and such exposures are known to be related to both spinal compression (Potvin, 2008) and back injury (Keyserling, 2000; Nelson and Hughes, 2009). The NIOSH action limit was exceeded about 5% of working time. The guidelines that form the NIOSH lifting index were developed based on combined biomechanical, physiological, and psychophysical criteria and were developed to be protective for the majority of the working population (Waters et al., 1993). The data presented in this paper represent primarily biomechanical exposures, and the physiological demands of the work are also likely to contribute to risk in a way undetected by these comparisons to NIOSH guidelines. Even more telling than the high prevalence of risk factors are the high rates of back injuries in heavy industry. Back injury claims in these industries are elevated above other industries (WorkSafeBC, 2005). The majority (77%) of the workers in this study, including those who were not selected based on a prior workers compensation claim, reported low back pain in the last 6 months (Trask et al., 2008). Together, these facts make a compelling argument for controlling exposures in these industries. 4.4. Implications for sampling strategies The differences between mean, peak, and cumulative measures of CNEMG speak to the importance of exposure metric selection and the sampling strategies employed in ergonomics. If only mean and peak exposure were consulted, the work tasks might be deemed ‘acceptable’ because they did not exceed the NIOSH guidelines. Similarly, if the full variability in tasks is not captured, exposure misclassification results. Measuring for multiple full shifts can be costly, but task-based sampling requires advance understanding about the type of work tasks and the scheduling of these tasks, so that each task may be targeted for measurement. Worstcase sampling requires an additional judgment about which tasks are likely to have the highest exposures. EMG is a valuable exposure assessment tool in that it is possible to efficiently measure many different exposure metrics over long periods with one data collection instrument. This makes it possible to investigate the effects of multiple risk factors in epidemiological studies. This may ultimately help pinpoint the aspects of exposure that are important to back injury and prioritize the aspects of work that should be modified using control strategies. Given that the literature on cumulative exposure is relatively young and the exposures in heavy industry seem to be high, this is an area that merits further study. 4.5. Limitations 4.5.1. Single equivalent muscle and asymmetrical movement The method described in this paper used only one muscle location for estimating spinal compression. Although measurements from the erector spinae muscles were the most practical for
workers performing frequent bending, co-contraction of the paraspinal muscles, such as the rectus abdominus, obliques, or latissimus dorsi, is an important source of spinal compression and is not accounted for in the exposures presented. Co-contraction may be one reason why CNEMG was shown to be less accurate during non-symmetrical lateral bending or twisting or combined movement tasks (Mientjes et al., 1999). In addition, the linear model used to transform EMG volts to Newtons of spinal compression was based on a static biomechanical model and the reference postures used for calibration were static postures in the sagittal plane. Lateral bending and twisting comprised 19.2% and 18% of the workday, respectively. Although these movements should have a small effect on the daily spinal compressive dose estimates, it still represents a non-trivial amount of time of over the whole workday (roughly 86 min). 4.5.2. The length-tension and velocity-tension relationships There are several overlapping issues relating to joint position or posture and EMG that can affect the CNEMG estimates, including the muscle length and velocity of change in muscle length. Tension in a muscle is developed from passive structures, such as tendons and fascia, as well as the active muscle fibres. Therefore the tension in a muscle is due to not only the activation of muscle fibres as measured by EMG but also the passive tensions in the soft tissue structures which increases with increasing muscle length (NIOSH, 1992). Muscles develop more tension while lengthening than while contracting; the faster an eccentric contraction is, the more force can be produced. In contrast, faster concentric contractions produce less force as the overlap of sliding filaments increases (NIOSH, 1992). These muscle length, tension, and velocity relationships are important in heavy industrial work; our examination of posture in heavy industry work shifts (Teschke et al., 2009) showed the tasks to involve both a wide range of postures (and therefore muscle lengths) as well as varying movement speeds (and therefore velocities). However, the CNEMG calibration procedure in this study (and in other CNEMG studies) uses only two static calibration reference postures. It is likely that muscle length and tension interact with muscle activation (as measured by EMG) in a very complex manner to produce the compression forces in the spine. Unfortunately, there is no way to partition and account for the changes in CNEMG amplitude due to muscle length or velocity. As described by Kothiyal and Kayis (2001), these limitations necessitate cautious interpretation of EMG (or in this case, CNEMG) results. 4.5.3. Fatigue Under heavy to moderate forces, when muscles fatigue and metabolites accumulate, EMG amplitude usually increases and the conduction velocities in the muscle decrease, resulting in a shift in the mean power frequency of the EMG signal to lower frequencies (NIOSH, 1992). Although power spectral data could not be derived from the RMS EMG data, the study did show a 3.5 mV increase in EMG amplitude between the pre-shift and post-shift reference measures. Subject, measurement session, and task condition accounted for the vast majority of explained EMG variability in the reference measures, while pre- and post-shift measurement accounted for only 3.3%. 4.5.4. Shear forces Although not as widely-studied or well-understood as compressive forces, shear forces have been shown to be an important risk factor for back injury (Potvin, 2008; Kerr et al., 2001). Shear forces act perpendicular to the compressive forces and may contribute to destabilization and structural damage to the spine (Fathallah et al., 1998). One limitation of the CNEMG
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
methodology is that it does not provide estimates of shear forces, and therefore a potentially important risk factor is not accounted for. 4.5.5. Pain Pain has been shown to have an impact on muscle recruitment and EMG, where those with pain activate their erector spinae muscles for longer than those without. This, in addition to other differences in spinal mechanics, lead to higher spinal loads in the pain group (Marras et al., 2004). Back pain is very common, 70e85% lifetime prevalence (Andersson, 1999) and known to be higher among workers in heavy industry, who have elevated rates of back strain claims (WorkSafeBC, 2005). In the heavy industrial population studied in this paper, 77% reported having back pain in the last 6 months; working with pain is a fact of life for these workers. It would be very difficult to find workers without back pain, and such workers would certainly not be a representative sample of their industries. However, given that the CNEMG methodology was developed using healthy participants, this discrepancy must be considered when interpreting results of the current study. 4.6. Interpretation and application of the CNEMG estimates of spinal compression Despite the limitations of EMG, it has been used successfully as a measure of working exposures in the upper limbs and shoulder (Earle-Richardson et al., 2008; Jones and Kumar, 2007a,b; Kothiyal and Kayis, 2001; Jin et al., 2009; Akesson et al., 1997; Østensvik et al., 2008; Sporrong et al., 1999; Straker and Mekhora, 2000; Capodaglio et al., 1996), the trunk (Earle-Richardson et al., 2008; Village et al., 2005; Keir and MacDonell, 2004; Lavender et al., 2007a,b, Lavender et al., 2007c), as well as the hip and lower limbs (Earle-Richardson et al., 2008; Jin et al., 2009). However, CNEMG is still a young method in terms of the number of published studies that have employed it; notably a lab-based study simulated lifting tasks (Mientjes et al., 1999), and an occupational field study of long-term care nurses (Village et al., 2005). Nonetheless, the Village et al. study demonstrated a relationship between CNEMG and injury reporting, indicating that CNEMG may have some predictive validity for back injury. However, the many limitations of CNEMG methodology mean that the CNEMG-estimated spinal compression values presented must be considered estimates of working exposures. To give an idea of the potential discrepancies, Mientjes et al. (1999) found difference between CNEMG and measures from a lab-based biomechanical model of 12e25% for simple tasks, and 25e45% for dynamic tasks or axial twisting. Although it is illustrative to compare the spinal compression estimates reported in the current study to those of previous studies and to NIOSH guidelines, biases due to the limitations of CNEMG estimates are unknown and as such, CNEMG may be best used as a relative measure for comparison within populations measured in the same way or for comparisons before and after an intervention. Assessing exposures in the working environments of heavy industry presents necessary choices and tradeoffs. What is lost in the accuracy of each individual estimate is gained in the ability to perform full-shift measurements and low cost (relative to labbased biomechanical methods). These benefits allowed collection of 135 full shifts of data in our study. CNEMG is imperfect and caution is required when interpreting its estimates, but not more so than other exposure assessment methodologies. For example, observation is often used to assess physical working exposures without great cost savings yet more subjectivity in the results (Trask et al., 2007). Similarly, self-reports and even job titles have been used as exposure assessment techniques to examine
453
relationships with back injury (Nelson and Hughes, 2009). According to Schultz and Andersson (1981), “.even a rough estimate of the loads generated by an activity will usually suffice for the solution of practical problems.” Given the widespread use of far less precise methods of exposure assessment, a direct measurement method such as CNEMG has much to contribute. Direct measurement of exposure to low back risk factors in heavy industry is challenging, with inclement, dusty, and vibrating environments, very high mobility, extreme postures, confined spaces, and unpredictable task schedules and work methods (Trask et al., 2007). However, the current study was able to capture a tremendous variety of tasks, jobs, and work conditions beyond stationary, repetitive assembly work. Although it may represent a compromise in accuracy compared to more advanced biomechanical methods, CNEMG allows for investigation of tasks and jobs that would be impossible to measure using video-based or motioncapture methods, and yields better objectivity and precision than observation or self-report. In this respect, CNEMG represents a quantitative ‘middle ground’ in the spectrum of exposure assessment tools available in ergonomics today. 5. Conclusion CNEMG provided a feasible, quantitative alternative for measuring risk factors for back injury in the construction, forestry, transportation, warehousing, and wood products industries. Mean and 90th percentile CNEMG-estimated compression was highest in construction and cumulative compression was highest in forestry. Transportation jobs had the lowest CNEMG measures across all metrics, likely because of the more sedentary driving tasks seen in this industry. When comparing CNEMG levels to NIOSH guidelines, there were non-trivial excursions above exposure thresholds. Overall, estimated low back compressive forces in heavy industry were in a range that would be considered ‘high risk’ when compared to NIOSH guidelines, although there was variability across industries and jobs. Even in this exploratory study, the percentage of time exceeding guideline thresholds (particularly in wood products and forestry) suggests that controlling exposure through engineering or administrative interventions is warranted. Acknowledgements The authors would like to thank WorkSafeBC, the Michael Smith Foundation for Health Research, and the Canadian Institutes for “Health Research Strategic Training Program: “Bridging Public Health, Engineering, and Policy Research” for financial and in kind support;” Yat Chow, Nancy Luong, Kevin Hong, James Cooper and Melissa Knott for data collection; and especially the workers and employers who participated in this study. References Akesson, I., Hansson, G.A., Balogh, I., Moritz, U., Skerfving, S., 1997. Quantifying work load in neck, shoulders and wrists in female dentists. Int. Arch. Occup. Environ. Health 69 (6), 461e474. Andersson, G.B., 1999 Aug 14. Epidemiological features of chronic low-back pain. Lancet 354 (9178), 581e585. Bao, S., Spielholz, P., Howard, N., Silverstein, B., 2009. Force measurement in field ergonomics research and application. Int. J. Ind. Ergon. 39, 333e340. Burdorf, A., 1992 Feb. Exposure assessment of risk factors for disorders of the back in occupational epidemiology. Scand. J. Work Environ. Health 18 (1), 1e9. Burdorf, A., Rossignol, M., Fathallah, F.A., Snook, S.H., Herrick, R.F., 1997 Aug. Challenges in assessing risk factors in epidemiologic studies on back disorders. Am. J. Ind. Med. 32 (2), 142e152. Capodaglio, P., Jensen, C., Christensen, H., 1996 JuleAug. Quantification of muscular activity in the shoulder region during monotonous repetitive work. Med. Lav 87 (4), 305e313.
454
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
da Costa, B.R., Vieira, E.R., 2009 Sep 14. Risk factors for work-related musculoskeletal disorders: a systematic review of recent longitudinal studies. Am. J. Ind. Med. 14, 14. Earle-Richardson, G., Jenkins, P., Fulmer, S., Mason, C., Burdick, P., May, J., 2008 May. An ergonomic intervention to reduce back strain among apple harvest workers in New York State. Appl. Ergon. 36 (3), 327e334. Fathallah, F.A., Marras, W.S., Parnianpour, M., 1998 Mar 15. An assessment of complex spinal loads during dynamic lifting tasks. Spine (Phila Pa 1976) 23 (6), 706e716. Jackson, J.A., Mathiassen, S.E., Dempsey, P.G., 2009 Jun. Methodological variance associated with normalization of occupational upper trapezius EMG using submaximal reference contractions. J. Electromyogr. Kinesiol. 19 (3), 416e427. Jin, S., McCulloch, R., Mirka, G.A., 2009. Biomechanical evaluation of postures assumed when harvesting bush crops. Int. Arch. Occup. Environ. Health 39, 347e352. Jones, T., Kumar, S., 2007a. Assessment of physical demands and comparison of multiple exposure definitions in a repetitive sawmill job: board edger operator. Ergonomics 50 (5), 676e693. Jones, T., Kumar, T., 2007b. Assessment of physical exposures and comparison of exposure definitions in a repetitive sawmill occupation: trim-saw operator. Work 28 (2), 183e196. Jonsson, B., 1988. The static load component in muscle work. Eur. J. Appl. Physiol. Occup. Physiol. 57, 305e310. Keir, P.J., MacDonell, C.W., 2004 Feb 26. Muscle activity during patient transfers: a preliminary study on the influence of lift assists and experience. Ergonomics 47 (3), 296e306. Kerr, M.S., Frank, J.W., Shannon, H.S., Norman, R.W., Wells, R.P., Neumann, W.P., et al., 2001 Jul. Biomechanical and psychosocial risk factors for low back pain at work. Am. J. Public Health 91 (7), 1069e1075. Keyserling, W.M., 2000. Workplace risk factors and occupational musculoskeletal disorders, part 1: a review of biomechanical and psychophysical research on risk factors associated with low-back pain. AIHAJ 61 (1), 39e50. Kothiyal, K., Kayis, B., 2001. Workplace layout for seated manual handling tasks: an electromyography study. Int. J. Ind. Ergon. 27 (1), 19e32. Kumar, S., 1990 Dec. Cumulative load as a risk factor for back pain. Spine 15 (12), 1311e1316. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Gacki-Smith, J., Kohok, A.K., 2007a. Designing ergonomic interventions for EMS workers, Part I: transporting patients down the stairs. Appl. Ergon. 38 (1), 71e81. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Kohok, A.K., Gacki-Smith, J., 2007b. Designing ergonomic interventions for emergency medical services workersepart III: bed to stairchair transfers. Appl. Ergon. 38 (5), 581e589. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Kohok, A.K., 2007c. Gacki-Smith J. Designing ergonomic interventions for EMS workers e part II: lateral transfers. Appl. Ergon. Mar 38 (2), 227e236. Marras, W.S., Ferguson, S.A., Burr, D., Davis, K.G., Gupta, P., 2004 JaneFeb. Spine loading in patients with low back pain during asymmetric lifting exertions. Spine J. 4 (1), 64e75. Marras, W.S., Sommerich, C.M., 1991 Apr. A three-dimensional motion model of loads on the lumbar spine: II. Model validation. Hum. Factors 33 (2), 139e149. Mathiassen, S.E., Burdorf, A., van der Beek, A.J., 2002 Feb. Statistical power and measurement allocation in ergonomic intervention studies assessing upper trapezius EMG amplitude. A case study of assembly work. J. Electromyogr. Kinesiol. 12 (1), 45e57. Mientjes, M.I., Norman, R.W., Wells, R.P., McGill, S.M., 1999 Jun. Assessment of an EMG-based method for continuous estimates of low back compression during asymmetrical occupational tasks. Ergonomics 42 (6), 868e879. Moller, T., Mathiassen, S.E., Franzon, H., Kihlberg, S., 2004 Jan 15. Job enlargement and mechanical exposure variability in cyclic assembly work. Ergonomics 47 (1), 19e40. Nelson, N., Hughes, R.E., 2009. Quantifying relationships between selected workrelated risk factors and back pain: a systematic review of objective
biomechanical measures and cost-related health outcomes. Int. J. Ind. Ergon. 39, 202e210. NIOSH, 1992 March. Selected Topics in Surface Electromyography for Use in the Occupational Setting: Expert Perspective. Report No.: Publication No. DHHS (NIOSH). 91e100. Nordander, C., Balogh, I., Mathiassen, S.E., Ohlsson, K., Unge, J., Skerfving, S., et al., 2004 Aug. Precision of measurements of physical workload during standardised manual handling. Part I: surface electromyography of m. trapezius, m. infraspinatus and the forearm extensors. J. Electromyogr. Kinesiol. 14 (4), 443e454. Norman, R., Wells, R., Neumann, P., Frank, J., Shannon, H., Kerr, M., 1998/12. A comparison of peak vs cumulative physical work exposure risk factors for the reporting of low back pain in the automotive industry. Clin. Biomech. 13 (8), 561e573. Østensvik, T., Veiersted, K.B., Cuchet, E., Nilsen, P., Hanse, J.J., Carlzon, C., et al., 2008 NoveDec. A search for risk factors of upper extremity disorders among forest machine operators: a comparison between France and Norway. Int. J. Ind. Ergon. 38 (11e12), 1017e1027. Potvin, J.R., 2008 Dec. Occupational spine biomechanics: a journey to the spinal frontier. J. Electromyogr. Kinesiol. 18 (6), 891e899. Schultz, A.B., Andersson, G.B., 1981 JaneFeb. Analysis of loads on the lumbar spine. Spine 6 (1), 76e82. Seidler, A., Bolm-Audorff, U., Heiskel, H., Henkel, N., Roth-Kuver, B., Kaiser, U., et al., 2001 Nov. The role of cumulative physical work load in lumbar spine disease: risk factors for lumbar osteochondrosis and spondylosis associated with chronic complaints. Occup. Environ. Med. 58 (11), 735e746. Sporrong, H., Sandsjo, L., Kadefors, R., Herberts, P., 1999 Dec. Assessment of workload and arm position during different work sequences: a study with portable devices on construction workers. Appl. Ergon. 30 (6), 495e503. Straker, L., Mekhora, K., 2000 Sep. An evaluation of visual display unit placement by electromyography, posture, discomfort and preference. Int. J. Ind. Ergon. 26 (3), 389e398. Teschke, K., Trask, C., Village, J., Johnson, P., Koehoorn, M., 2009. Measuring posture for epidemiology: comparing inclinometry, observations, and self-reports. Ergonomics 52 (9), 1067e1078. Trask, C., Teschke, K., Village, J., Chow, Y., Johnson, P., Luong, N., et al., 2007. Measuring low back injury risk factors in challenging work environments: an evaluation of cost and feasibility. Am. J. Ind. Med. 50 (9), 687e696. Trask C.M., Teschke K., Koehoorn M., 2008. Back Pain and Disability Reporting by Workers in Heavy Industry: Implications for Case-control Studies. Association of Canadian Ergonomists 39th Annual Conference October 5e8 Alymer, Quebec, Canada. Village, J., Frazer, M., Cohen, M., Leyland, A., Park, I., Yassi, A., 2005 Sep. Electromyography as a measure of peak and cumulative workload in intermediate care and its relationship to musculoskeletal injury: an exploratory ergonomic study. Appl. Ergon. 36 (5), 609e618. Village, J., Trask, C., Luong, N., Chow, Y., Johnson, P., Koehoorn, M., et al., 2009. Development and evaluation of an observational Back-Exposure Sampling Tool (Back-EST) for work-related back injury risk factors. Appl. Ergon. 40, 538e544. Waters, T.R., Putz-Anderson, V., Garg, A., Fine, L.J., 1993 Jul. Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics 36 (7), 749e776. Wells, R., Norman, R., Neumann, P., Andrews, D., Frank, J., Shannon, H., et al., 1997 Jan. Assessment of physical work load in epidemiologic studies: common measurement metrics for exposure assessment. Ergonomics 40 (1), 51e61. Winkel, J., Mathiassen, S.E., 1994 Jun. Assessment of physical work load in epidemiologic studies: concepts, issues and operational considerations. Ergonomics 37 (6), 979e988. WorkSafeBC, 2006. From awareness to action statistics 2005. Accessed from. http:// www.worksafebc.com/publications/reports/statistics_reports/assets/pdf/stats 2005.pdf (accessed 22.11.2006.).
Contents lists available at ScienceDirect
International Journal of Industrial Ergonomics journal homepage: www.elsevier.com/locate/ergon
EMG estimated mean, peak, and cumulative spinal compression of workers in five heavy industries Catherine Trask a, e, *, Kay Teschke b, e, Jim Morrison c, Pete Johnson d, Judy Village e, Mieke Koehoorn b, e a
Centre for Musculoskeletal Research, University of Gävle, SE e 801 76 Gävle, Sweden School of Population and Public Health, University of British Columbia, 5804 Fairview Avenue Vancouver, BC V6T 1Z3 Canada c School of Kinesiology, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6 Canada d Department of Environmental and Occupational Health Sciences, University of Washington Box 357234- 4225 Roosevelt Way NE Suite 100 Seattle, 98195e7234 WA, USA e School of Environmental Health, University of British Columbia 375-2206 East Mall Vancouver, BC V6T 1Z3 Canada b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 June 2009 Received in revised form 12 January 2010 Accepted 26 February 2010 Available online 29 March 2010
The goal of this study was to explore the use of compression-normalized electromyography (CNEMG) to estimate mean, peak, and cumulative loading of the low back in workers of five heavy industries and to compare the estimates to the NIOSH guidelines. Full-shift (5.5e10.3 h) EMG measurements were collected from 105 workers and transformed into units of low back compressive force (Newtons). The mean, peak, and cumulative CNEMG as well as the percentage of work time spent above 3400 N and 6800 N thresholds were calculated. Mean CNEMG (sd) was 1564 N (796), peak was 2721 (1545), and cumulative was 38 MN s (22). Mean time spent above the NIOSH threshold of 3400 N was on average 6.3% of shift, while mean time spent above the 6400 N threshold was around 1%. CNEMG allowed the feasible investigation of tasks and jobs that would be virtually impossible with more advanced biomechanical methods and represents a more objective measure of exposure than observation or self-report. CNEMG is a relatively new method with methodological limitations, however, CNEMG's strength may be as an objective, relative measure of exposure in studies where low back EMG is collected in a relatively systematic and structured manner. Ó 2010 Elsevier B.V. All rights reserved.
Keywords: Electromyography Back injury Exposure assessment Construction Compression-normalized EMG
1. Introduction In the Canadian province of British Columbia, back injuries account for approximately 25% of all workers' compensation claims, 23% of workdays lost, and 20% of total claims costs. Onequarter of back injury claims are made by workers in five heavy industries: construction, forestry, transportation, warehousing and wood products (WorkSafeBC, 2005). Work tasks and activities in heavy industry often involve considerable time in awkward postures and manual materials handling (MMH) (Teschke et al., 2009), and these types of exposures are thought to be related to musculoskeletal injuries (Keyserling, 2000). Multiple lab investigations have demonstrated that nonneutral postures and MMH increase muscle activity, which is in turn related to spinal moments and spinal compression (Marras
* Corresponding author. Centre for Musculoskeletal Research, University of Gävle, SE e 801 76 Gävle, Sweden. Tel.: þ46 604 221 0553. E-mail addresses: [email protected] (C. Trask), [email protected] (K. Teschke), [email protected] (J. Morrison), [email protected] (P. Johnson), [email protected] (J. Village), [email protected] (M. Koehoorn). 0169-8141/$ e see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ergon.2010.02.006
and Sommerich, 1991; Potvin, 2008;). These studies have yielded many excellent biomechanical modeling methods that account for asymmetrical, complex motion and dynamic segment/load inertias (Potvin, 2008). Unfortunately, heavy industry presents extensive measurement challenges to researchers (Trask et al., 2007); the more complex lab-based methods are too cumbersome to be applicable for many work areas or tasks, and too expensive to be used for a large number of workers or a long sampling duration. Although with some loss of precision when compared to labbased methods, portable electromyography (EMG) may provide a practical alternative to complex biomechanical methods because it can be used to estimate the low back compressive forces throughout the diverse activities of a heavy industry work shift. EMG of trunk muscles during work tasks has been proposed as a method for estimating spinal loads using a method called compression-normalized EMG (CNEMG) (Mientjes et al., 1999). CNEMG is a direct measurement method which allows for estimation of spinal compression in workplaces where traditional biomechanical methods, such as video analysis or motion-capture systems, are not feasible. Bao et al. (2009) call measurement of
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
working forces “both important and challenging”, emphasizing that capturing the important forces is a critical component of measurement (Bao et al., 2009). One alternative to measuring only a few forces is to employ a method that allows measurement of a full-shift. EMG has advantages in that it allows workers to move freely throughout their working environments with minimal interruption, and has been used to assess muscle activity in apple pickers working in the field (Earle-Richardson et al., 2008) and estimate all day low back compressive forces in aides working in long term care (Village et al., 2005). A recent review shows mean, peak, and cumulative spinal compression to be related to low back injury (da Costa, 2009). Peak and cumulative spinal loads have been shown to be independently related to back injury, (Village et al., 2005; Norman et al., 1998; Kumar, 1990; Seidler et al., 2001; Kerr et al., 2001; Nelson and Hughes, 2009). Village et al. (2005) used full-shift EMG to estimate peak and cumulative spinal loading in 32 care aides and found a correlation between these measures and musculoskeletal injury rates as well as total tasks observed (Village et al., 2005). For example, a worker could have moderate peak loads that are sustained throughout the day, such that the total load on the back could surpass that of a worker with high but infrequent peak loads. Knowledge about cumulative load as an exposure is still developing, and it is not clear how it relates to classical peak loading measures. Understanding the relative contribution of peak and cumulative loads to overall occupational exposure in heavy industries can help target future exposure assessment and intervention strategies. Spinal loading studies have typically assessed exposure in cyclical manufacturing and assembly tasks (Norman et al., 1998; Kerr et al., 2001) or via self-report (Kumar, 1990; Seidler et al., 2001). However, comparatively little has been done on non-cyclical tasks, such as those found in resource industries and construction, perhaps because the most precise and accurate exposure assessment methods are nearly impossible to apply in these industries. Self-report and observation methods can be used to assess exposure, but these methods are more subjective and less precise than direct measurement (Winkel and Mathiassen, 1994; Burdorf, 1992; Burdorf et al., 1997; Wells et al., 1997). EMG represents a compromise between lab-based biomechanical methods and subjective methods; it is an objective, direct measurement exposure assessment method that can be used to attain full-shift estimates of non-cyclical activities in heavy industries. EMG also provides a unique opportunity to estimate mean, peak, and cumulative exposures concurrently, in spinal compression units that can be used as a ‘common metric’ to compare to other research results and to exposure guidelines (Wells et al., 1997). The goal of this study was to explore using full-shift EMG data to calculate CNEMG estimates of mean, peak, and cumulative spinal compression in workers in five heavy industries, and to interpret these estimated exposures in light of existing studies and guidelines. 2. Methods 2.1. Worker recruitment and sampling strategy This analysis was undertaken as part of a larger study of low back injury risk factor measurement in the heavy industrial sectors of construction, forestry, transportation, wood products, and warehousing in the Canadian province of British Columbia. A random sample of workers with a lost-time workers' compensation claim for a back injury in 2001 were contacted by phone and invited to participate; 75% of workers contacted agreed to participate. Employers were then contacted for permission to visit the worksite and make full-shift exposure measurements and to allow
449
recruitment of additional co-workers; 73% of employers gave permission. Prior to measurement, researchers visited each participating worksite to recruit up to 4 co-workers of the original WCB-selected worker, bringing the total sample to 126 individuals. All workers had production jobs without modified duties. The resulting sample represented 50 different worksites and a wide variety of job titles, tasks and task patterns. The work ranged from static, continuous tasks such as delivery truck driving, to dynamic but monotonous tasks such as repeatedly stacking wood off a conveyor belt, to dynamic and varied work such as maintenance work with no repeated tasks within a day. 2.2. EMG data collection and processing Full-shift (5.5e10.3 h, excluding breaks) EMG measurements were made using a portable data collection system with on-board memory (ME3000P4/ME3000P8, Mega Electronics, Finland) and disposable AgeAgCl electrodes (Blue Sensor N-00-S, Ambu, Denmark). Electrodes were placed bilaterally over the erector spinae at approximately the level of L4, with a 20 mm inter-electrode spacing and a ground electrode and preamplifier placed on the posterior aspect of the iliac crest. Signals were collected at 1000 Hz and filtered internally using an 8e500 Hz band pass filter. Rootmean-square values were data-logged at 10 Hz. During work breaks, data from the portable system were downloaded to a laptop computer. After data collection, EMG signals were imported to a customdeveloped LABVIEW program. All shift-long EMG measures were visually inspected for signal drift and noise spikes. Observation records were also consulted to check for notes about worker comments on shifting of electrodes. Data were removed where the observation records or workers comments indicated there were shifting electrodes, observable signal drift or suspicious noise spikes in the data. 2.3. Observation and posture measurements CNEMG methodology has been shown to be less accurate when assessing non-symmetrical tasks i.e. lateral bending or twisting (Mientjes et al., 1999). To investigate, Concurrent observation and trunk inclinometer data were collected to investigate the extent of non-symmetrical postures during measured workshifts. Oneminute sampling observations of trunk postures were made over the whole work shift by trained researchers as described in Village et al. (2009). A portable, data-logging torso-mounted inclinometer (VC-323, Microstrain Inc, Williston, VT, USA) was also used to collect shift-long data as described in Teschke et al. (2009). 2.4. Transformation to spinal compression: compressionnormalized EMG EMG voltage was transformed into units of spinal compression force (Newtons) using a linear calibration equation based on submaximal reference calibration postures performed during work time at the beginning and end of the shift (approximately 1 h for set-up and take-down). The first set of reference contractions involved standing upright with the hands by the sides. The second set of reference contractions involved static 45 forward trunk flexion while holding an 11.5 kg weight directly below the shoulders. The reference contractions were performed twice for 5 s each time at the beginning of each shift. As reported by Jackson et al. (2009) methodological variance due to EMG calibration in the trapezius is less than 5%, and collecting more than two repeated reference contractions over two days yields only marginal improvements in precision. Although the Jackson et al. study was
450
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
performed on the trapezius in a lab environment, it supports the collection of two repeated reference contractions. For each of the reference positions, spinal compression was estimated using a computer-based link-segment model (Ergowatch 4D WATBAK, Waterloo University, Canada). The estimated spinal compression (in Newtons) and the muscle activity (in uV) for each position were used to generate a linear calibration equation. This process yielded a continuous estimate of “compression-normalized EMG” (CNEMG) for the whole workday. This is similar to the compression normalization methods used by Mientjes et al. (1999) and Village et al. (2005). 2.5. Investigation of potential fatigue effect of testing over a full work shift Given that fatigue is known to impact EMG, the data were analyzed to determine whether the EMG amplitude changed over the worker's shift. Four reference contractions were collected at the beginning and end of each worker's shift/measurement session. The reference contractions included: standing upright, standing with the trunk flexed at 45 , and standing at 45 with an 11.5 kg load and standing at 60 with an 11.5 kg load. There were two repetitions for each reference contraction, totaling 16 reference contractions per subject (2 calibration sessions 4 reference conditions 2 repeats per subject). Pairs of measurements were matched for subject and type of reference contraction, and paired t-tests used to calculate mean differences between pre- and post-shift measurements. Mixed effect models (using subject and measurement session as random effect terms) were used to estimate the components of variance in the EMG measures due to pre- versus post-shift timing, reference task, subject, and measurement session. A ‘measurement session’ refers to the unique electrode/skin interface for each shift measurement per worker. 2.6. Statistical analysis and comparison to guidelines Three EMG exposure metrics were calculated for each individual's work shift data: the arithmetic mean, the 90th percentile, and cumulative exposure. The arithmetic mean is a measure of central tendency in the intensity of exposure. The 90th percentile was included as an estimate of ‘peak’ intensity of spinal load as in previous studies (Mientjes et al., 1999; Jonsson, 1988; Mathiassen et al., 2002; Nordander et al., 2004; Moller et al., 2004). Cumulative exposure represents a combination of the intensity, duration, and frequency of exposure combined into a single ‘daily dose’. It was calculated as the sum of instantaneous values throughout the shift, expressed as MN sec (MegaNetwton-seconds) of cumulative EMG activity. Since shift lengths varied in this sample, cumulative spinal loads were calculated without normalizing to a standard day length to reflect the variation in daily dose. Standardizing cumulative exposure to a constant shift length would generate a rank order of exposures identical to that of the mean exposures. The mean and peak CNEMG levels were compared to levels approximating the NIOSH compression action limit guideline (3400 N), and an upper limit of twice the guideline (6800 N) (Waters et al., 1993). In addition, the percentage of work time spent above each threshold was calculated for each work shift. There are currently no institution-endorsed guidelines for cumulative compression limits analogous to the NIOSH compression guidelines. However, a case-control study (Norman et al., 1998) showed cumulative spinal compression is a significant and independent predictor of back pain. Mean cumulative exposure was 21 MN s in cases with back pain compared to 19.5 MN s in controls, and approximately 50% of workers with cumulative compressions of
23 MN s reported back pain. Although not a guideline per se, a threshold of 21 MN s was used as a benchmark for cumulative compression. 3. Results 3.1. Study participants Full-shift EMG measurements were collected from 103 of the 126 individuals. Repeated measurements were completed for 34% of the 103 included individuals, for a total of 138 worker-days. The interval between workers' first and second measurement days ranged from 1 to 439 days (mean ¼ 93 SD ¼ 64). Demographic data on the sample are summarized in Table 1. 3.2. Observation and posture measurement of on-symmetrical postures Inclinometer results showed lateral bending greater than 12 made up less than 20% of the workday on average. Direct observation results were similar showing a mean of 19.2% (sd ¼ 15%) of the work shift spent laterally bending. Twisting measures are not captured by the inclinometer, so only observational measures are presented. Twisting postures were observed on average 18.0% (sd ¼ 17%) of the work shift. These results are described in more detail elsewhere (Teschke et al., 2009). 3.3. Comparison of pre- and post-shift EMG calibration Paired t-tests showed significant differences between 916 pairs of pre- and post-shift measurements with pre-shift values on average 3.25 mV higher than post-shift values (mean reference contraction voltage was 32 mV). However, pre- and post-shift differences explained only 3.3% of the variability in EMG, This variability was dwarfed by the variability due to the reference task (23%), and due to the subject and measurement session (44.6% combined). 3.4. Compression-normalized EMG in heavy industry Table 2 summarizes the CNEMG measures, stratified by industry and job. Because the numbers of jobs represented in this study are small compared to the numbers in each of these industries, these results represent an initial window on exposure by job. The rank order of mean, peak, and cumulative CNEMG exposures was similar across industries. Mean CNEMG ranged from 1269 N (sd ¼ 654) in transportation to 1771 N (sd ¼ 593) in construction. There was less variation in peak CNEMG, which ranged from 2338 N (1679) in transportation to 3066 N (1065) in construction. Cumulative CNEMG showed a somewhat different ranking, lowest in transportation at 32 MN s (sd ¼ 0.24), and highest in forestry at 44 MN s (sd ¼ 0.20). Transportation work shifts consistently had the lowest measured exposures and construction the highest in all metrics except cumulative exposure. The highest mean and peak compressions were in tree fallers in the forestry industry; the lowest were among heavy equipment operators in transportation.
Table 1 Demographic characteristics of study participants (N ¼ 103). Variable
Value
% Male Height in cm (sd) Weight in kg (sd) Age, in years, on sampling day (sd)
95.3% 178.1 (7.9) 85.2 (16.1) 42.2 (12)
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
451
Table 2 Compression-normalized EMG exposure metrics measured in workers of five heavy industries (N ¼ 138 workshifts).
All industries Construction Construction carpenter Construction labourer Construction supervisor Construction trades Otherd Forestry Boomman Faller Heavy Equip.mechanic Log-Machine Operators Otherd Transportation Airport ramp attendant Automotive mechanic Heavy equip op Truck driver Otherd Warehousing Forklift operator Warehouse person Wood products Cabinet maker Forklift operator Lumber puller grader Papermaking Otherd a b c d
na
Meanb (sd)
Cumulativec (sd)
10th %ileb (sd)
90 th %ileb (sd)
% time above 3400 N (sd)
% time above 6800 N (sd)
138 25 7 6 4 3 5 30 10 4 4 4 8 33 6 6 2 6 13 23 21 2 27 6 9 5 5 2
1564 1771 1591 2076 1901 1650 1664 1727 1482 2738 2159 1351 1496 1269 1193 1062 863 1084 1452 1444 1465 1223 1653 1988 1623 1473 1433 1787
38 37 33 40 39 35 38 44 37 69 52 40 39 32 23 20 27 31 39 36 36 26 39 38 45 32 36 46
641 731 710 823 666 702 808 667 659 770 840 574 592 533 451 430 500 506 581 602 610 521 691 797 667 693 621 664
2721 3066 2739 3626 3306 2882 2816 2911 2555 4789 3484 2143 2350 2911 2218 1907 1515 1939 2746 2531 2573 2092 2821 3284 2825 2391 2441 3439
6.34 5.06 4.00 2.04 2.3 7.09 8.00 7.51 5.00 23.7 13.8 4.05 0.87 3.66 2.92 0.53 0.31 2.13 6.92 3.66 6.29 1.59 10.3 14.2 12.3 4.03 5.00 32.5
0.9 0.16 0.06 0.16 0.03 0.49 0.21 0.65 0.55 0.58 0.76 1.12 0.08 0.84 0.22 0.0 0.0 0.13 1.31 0.84 0.78 0.02 2.00 3.33 3.54 0.06 0.1 5.06
(796) (593) (218) (564) (1012) (695) (693) (760) (601) (1154) (709) (310) (525) (654) (272) (344) (176) (327) (533) (755) (787) (293) (1076) (1590) (1232) (441) (556) (1483)
(22) (17) (7) (4) (26) (2) (1.4) (20) (13) (38) (10) (15) (16) (23) (13) (8) (20) (10) (18) (27) (29) (5) (21) (18) (28) (8) (14) (47)
(244) (217) (147) (216) (86) (389) (393) (225) (269) (325) (228) (94) (129) (216) (90) (93) (5) (126) (193) (193) (200) (77) (311) (457) (368) (193) (118) (298)
(1545) (1065) (308) (1103) (2056) (994) (1423) (1247) (1025) (1738) (744) (566) (694) (1247) (489) (731) (573) (658) (1197) (1574) (1640) (511) (1945) (2606) (2358) (736) (1046) (3124)
(10.9) (0.7) (2.23) (0.7) (2.13) (8.44) (12) (0.77) (3.83) (0.77) (9.05) (3.31) (0.62) (6.83) (2.18) (0.68) (0.37) (1.31) (6.47) (0.7) (8.71) (1.69) (2.0) (29.8) (24.2) (2.24) (5.74) (2.0)
(2.99) (0.04) (0.071) (0.04) (0.043) (0.78) (0.37) (0.07) (0.5) (0.07) (0.28) (1.52) (0.11) (3.4) (0.24) (0.01) (0) (0.1) (2.22) (0.30) (1.56) (0.03) (0.50) (7.41) (7.29) (0.03) (0.14) (0.5)
Sample Size. Newtons of Spinal compression (N). MegaNewton-seconds (MN s). Jobs with few measurements combined to prevent identification of individuals.
3.5. Comparisons to guidelines The mean CNEMG-estimated compression value for all industries combined (mean ¼ 1564 N) was only 42% of the NIOSH action limit guideline; each of the industry means was well below the NIOSH limit as well. Mean peak CNEMG measurements were much closer to the NIOSH values, but even the highest means (in construction) represented only 89% of the NIOSH limit. However, in every industry some measurements were above the 3400 N and 6800 N thresholds. Mean time spent above the 3400 N threshold was 6.3% for all workers (Table 2); wood products had the greatest time above this threshold at 10%, while transportation had the least at 3.7%. Wood products also had the most time spent above the 6800 N threshold at 2%; construction had the least at 0.16%. Although construction had the highest mean and peak CNEMG values, it had the lowest variability (in terms of standard deviations), and did not include the worst case jobs for time spent above NIOSH recommended limits. Compared to the cumulative compression levels from previous studies suggesting a likelihood of back pain, the cumulative compression measures were high. The mean CNEMG-estimated cumulative compression in every industry and almost every job exceeded the 21 MN s value associated with back pain reporting (Norman et al., 1998). 4. Discussion 4.1. Exposure intensity The average mean and peak CNEMG-estimated compression values are below the NIOSH-based 3400 N and 6800 N thresholds for all industries. It should be noted that the ‘peak’ values reported in this paper were 90th percentile values, so 10% of the exposure
values in a day were higher than these levels. All industries had workshifts that exceeded these levels at some point during the day, albeit for as little as 0.16% of the work shift. Clearly exposure intensity is problematic in this industrial population, as evidenced by the percentage of time spent above NIOSH thresholds. Although the percentage of time spent above a NIOSH guideline has not been explicitly compared to risk for back injury, any exceedance of the NIOSH guidelines is considered to introduce elevated risk for some workers (Waters et al., 1993) and presents an opportunity for improvement via appropriate controls. 4.2. Cumulative exposure Cumulative exposure has been shown to be a significant and independent predictor of back pain (Village et al., 2005; Norman et al., 1998/12; Kumar, 1990; Seidler et al., 2001). The mean cumulative compression estimates were substantially above a level shown to be related to back pain reporting, in both autoworkers (Norman et al., 1998) and compensated injury rates in care aides in a long term care facility(Village et al., 2005). This would seem to suggest that mean or peak exposure intensity, as well as the duration and frequency of these exposures, may be problematic in these heavy industries. An example can be seen in the higher cumulative compression values for forestry compared to construction; the difference might result from longer shift lengths in forestry where the average measured shift length (excluding breaks and the approximately 1-h study set-up period) was 7.4 h versus 5.8 h in construction. 4.3. Implications for back injury risk Workers in some jobs had high exposures in this sample. For example, the amount of time spent above the NIOSH guideline
452
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
among fallers (23.7%, sd ¼ 0.77%), heavy equipment mechanics (13.8% sd ¼ 9.05), cabinet makers (14.2% sd ¼ 29.8) and forklift operators (12.3% sd ¼ 24.2) suggests that even if these workers were young and very fit, they would be at increased risk of back injury. Since these jobs were observed by researchers during data collection to spend a large proportion of time manual materials handling and bending over in a sustained way (fallers and heavy equipment mechanics) or bending repetitively while lifting (forklift operators and cabinet makers), intervention efforts might begin with strategies to control these characteristics of the work tasks. Finding compression-normalized EMG values that exceed NIOSH guidelines in a heavy industrial working population is not surprising, especially since these workers spend substantial time performing MMH activities and in flexion postures (Teschke et al., 2009) and such exposures are known to be related to both spinal compression (Potvin, 2008) and back injury (Keyserling, 2000; Nelson and Hughes, 2009). The NIOSH action limit was exceeded about 5% of working time. The guidelines that form the NIOSH lifting index were developed based on combined biomechanical, physiological, and psychophysical criteria and were developed to be protective for the majority of the working population (Waters et al., 1993). The data presented in this paper represent primarily biomechanical exposures, and the physiological demands of the work are also likely to contribute to risk in a way undetected by these comparisons to NIOSH guidelines. Even more telling than the high prevalence of risk factors are the high rates of back injuries in heavy industry. Back injury claims in these industries are elevated above other industries (WorkSafeBC, 2005). The majority (77%) of the workers in this study, including those who were not selected based on a prior workers compensation claim, reported low back pain in the last 6 months (Trask et al., 2008). Together, these facts make a compelling argument for controlling exposures in these industries. 4.4. Implications for sampling strategies The differences between mean, peak, and cumulative measures of CNEMG speak to the importance of exposure metric selection and the sampling strategies employed in ergonomics. If only mean and peak exposure were consulted, the work tasks might be deemed ‘acceptable’ because they did not exceed the NIOSH guidelines. Similarly, if the full variability in tasks is not captured, exposure misclassification results. Measuring for multiple full shifts can be costly, but task-based sampling requires advance understanding about the type of work tasks and the scheduling of these tasks, so that each task may be targeted for measurement. Worstcase sampling requires an additional judgment about which tasks are likely to have the highest exposures. EMG is a valuable exposure assessment tool in that it is possible to efficiently measure many different exposure metrics over long periods with one data collection instrument. This makes it possible to investigate the effects of multiple risk factors in epidemiological studies. This may ultimately help pinpoint the aspects of exposure that are important to back injury and prioritize the aspects of work that should be modified using control strategies. Given that the literature on cumulative exposure is relatively young and the exposures in heavy industry seem to be high, this is an area that merits further study. 4.5. Limitations 4.5.1. Single equivalent muscle and asymmetrical movement The method described in this paper used only one muscle location for estimating spinal compression. Although measurements from the erector spinae muscles were the most practical for
workers performing frequent bending, co-contraction of the paraspinal muscles, such as the rectus abdominus, obliques, or latissimus dorsi, is an important source of spinal compression and is not accounted for in the exposures presented. Co-contraction may be one reason why CNEMG was shown to be less accurate during non-symmetrical lateral bending or twisting or combined movement tasks (Mientjes et al., 1999). In addition, the linear model used to transform EMG volts to Newtons of spinal compression was based on a static biomechanical model and the reference postures used for calibration were static postures in the sagittal plane. Lateral bending and twisting comprised 19.2% and 18% of the workday, respectively. Although these movements should have a small effect on the daily spinal compressive dose estimates, it still represents a non-trivial amount of time of over the whole workday (roughly 86 min). 4.5.2. The length-tension and velocity-tension relationships There are several overlapping issues relating to joint position or posture and EMG that can affect the CNEMG estimates, including the muscle length and velocity of change in muscle length. Tension in a muscle is developed from passive structures, such as tendons and fascia, as well as the active muscle fibres. Therefore the tension in a muscle is due to not only the activation of muscle fibres as measured by EMG but also the passive tensions in the soft tissue structures which increases with increasing muscle length (NIOSH, 1992). Muscles develop more tension while lengthening than while contracting; the faster an eccentric contraction is, the more force can be produced. In contrast, faster concentric contractions produce less force as the overlap of sliding filaments increases (NIOSH, 1992). These muscle length, tension, and velocity relationships are important in heavy industrial work; our examination of posture in heavy industry work shifts (Teschke et al., 2009) showed the tasks to involve both a wide range of postures (and therefore muscle lengths) as well as varying movement speeds (and therefore velocities). However, the CNEMG calibration procedure in this study (and in other CNEMG studies) uses only two static calibration reference postures. It is likely that muscle length and tension interact with muscle activation (as measured by EMG) in a very complex manner to produce the compression forces in the spine. Unfortunately, there is no way to partition and account for the changes in CNEMG amplitude due to muscle length or velocity. As described by Kothiyal and Kayis (2001), these limitations necessitate cautious interpretation of EMG (or in this case, CNEMG) results. 4.5.3. Fatigue Under heavy to moderate forces, when muscles fatigue and metabolites accumulate, EMG amplitude usually increases and the conduction velocities in the muscle decrease, resulting in a shift in the mean power frequency of the EMG signal to lower frequencies (NIOSH, 1992). Although power spectral data could not be derived from the RMS EMG data, the study did show a 3.5 mV increase in EMG amplitude between the pre-shift and post-shift reference measures. Subject, measurement session, and task condition accounted for the vast majority of explained EMG variability in the reference measures, while pre- and post-shift measurement accounted for only 3.3%. 4.5.4. Shear forces Although not as widely-studied or well-understood as compressive forces, shear forces have been shown to be an important risk factor for back injury (Potvin, 2008; Kerr et al., 2001). Shear forces act perpendicular to the compressive forces and may contribute to destabilization and structural damage to the spine (Fathallah et al., 1998). One limitation of the CNEMG
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
methodology is that it does not provide estimates of shear forces, and therefore a potentially important risk factor is not accounted for. 4.5.5. Pain Pain has been shown to have an impact on muscle recruitment and EMG, where those with pain activate their erector spinae muscles for longer than those without. This, in addition to other differences in spinal mechanics, lead to higher spinal loads in the pain group (Marras et al., 2004). Back pain is very common, 70e85% lifetime prevalence (Andersson, 1999) and known to be higher among workers in heavy industry, who have elevated rates of back strain claims (WorkSafeBC, 2005). In the heavy industrial population studied in this paper, 77% reported having back pain in the last 6 months; working with pain is a fact of life for these workers. It would be very difficult to find workers without back pain, and such workers would certainly not be a representative sample of their industries. However, given that the CNEMG methodology was developed using healthy participants, this discrepancy must be considered when interpreting results of the current study. 4.6. Interpretation and application of the CNEMG estimates of spinal compression Despite the limitations of EMG, it has been used successfully as a measure of working exposures in the upper limbs and shoulder (Earle-Richardson et al., 2008; Jones and Kumar, 2007a,b; Kothiyal and Kayis, 2001; Jin et al., 2009; Akesson et al., 1997; Østensvik et al., 2008; Sporrong et al., 1999; Straker and Mekhora, 2000; Capodaglio et al., 1996), the trunk (Earle-Richardson et al., 2008; Village et al., 2005; Keir and MacDonell, 2004; Lavender et al., 2007a,b, Lavender et al., 2007c), as well as the hip and lower limbs (Earle-Richardson et al., 2008; Jin et al., 2009). However, CNEMG is still a young method in terms of the number of published studies that have employed it; notably a lab-based study simulated lifting tasks (Mientjes et al., 1999), and an occupational field study of long-term care nurses (Village et al., 2005). Nonetheless, the Village et al. study demonstrated a relationship between CNEMG and injury reporting, indicating that CNEMG may have some predictive validity for back injury. However, the many limitations of CNEMG methodology mean that the CNEMG-estimated spinal compression values presented must be considered estimates of working exposures. To give an idea of the potential discrepancies, Mientjes et al. (1999) found difference between CNEMG and measures from a lab-based biomechanical model of 12e25% for simple tasks, and 25e45% for dynamic tasks or axial twisting. Although it is illustrative to compare the spinal compression estimates reported in the current study to those of previous studies and to NIOSH guidelines, biases due to the limitations of CNEMG estimates are unknown and as such, CNEMG may be best used as a relative measure for comparison within populations measured in the same way or for comparisons before and after an intervention. Assessing exposures in the working environments of heavy industry presents necessary choices and tradeoffs. What is lost in the accuracy of each individual estimate is gained in the ability to perform full-shift measurements and low cost (relative to labbased biomechanical methods). These benefits allowed collection of 135 full shifts of data in our study. CNEMG is imperfect and caution is required when interpreting its estimates, but not more so than other exposure assessment methodologies. For example, observation is often used to assess physical working exposures without great cost savings yet more subjectivity in the results (Trask et al., 2007). Similarly, self-reports and even job titles have been used as exposure assessment techniques to examine
453
relationships with back injury (Nelson and Hughes, 2009). According to Schultz and Andersson (1981), “.even a rough estimate of the loads generated by an activity will usually suffice for the solution of practical problems.” Given the widespread use of far less precise methods of exposure assessment, a direct measurement method such as CNEMG has much to contribute. Direct measurement of exposure to low back risk factors in heavy industry is challenging, with inclement, dusty, and vibrating environments, very high mobility, extreme postures, confined spaces, and unpredictable task schedules and work methods (Trask et al., 2007). However, the current study was able to capture a tremendous variety of tasks, jobs, and work conditions beyond stationary, repetitive assembly work. Although it may represent a compromise in accuracy compared to more advanced biomechanical methods, CNEMG allows for investigation of tasks and jobs that would be impossible to measure using video-based or motioncapture methods, and yields better objectivity and precision than observation or self-report. In this respect, CNEMG represents a quantitative ‘middle ground’ in the spectrum of exposure assessment tools available in ergonomics today. 5. Conclusion CNEMG provided a feasible, quantitative alternative for measuring risk factors for back injury in the construction, forestry, transportation, warehousing, and wood products industries. Mean and 90th percentile CNEMG-estimated compression was highest in construction and cumulative compression was highest in forestry. Transportation jobs had the lowest CNEMG measures across all metrics, likely because of the more sedentary driving tasks seen in this industry. When comparing CNEMG levels to NIOSH guidelines, there were non-trivial excursions above exposure thresholds. Overall, estimated low back compressive forces in heavy industry were in a range that would be considered ‘high risk’ when compared to NIOSH guidelines, although there was variability across industries and jobs. Even in this exploratory study, the percentage of time exceeding guideline thresholds (particularly in wood products and forestry) suggests that controlling exposure through engineering or administrative interventions is warranted. Acknowledgements The authors would like to thank WorkSafeBC, the Michael Smith Foundation for Health Research, and the Canadian Institutes for “Health Research Strategic Training Program: “Bridging Public Health, Engineering, and Policy Research” for financial and in kind support;” Yat Chow, Nancy Luong, Kevin Hong, James Cooper and Melissa Knott for data collection; and especially the workers and employers who participated in this study. References Akesson, I., Hansson, G.A., Balogh, I., Moritz, U., Skerfving, S., 1997. Quantifying work load in neck, shoulders and wrists in female dentists. Int. Arch. Occup. Environ. Health 69 (6), 461e474. Andersson, G.B., 1999 Aug 14. Epidemiological features of chronic low-back pain. Lancet 354 (9178), 581e585. Bao, S., Spielholz, P., Howard, N., Silverstein, B., 2009. Force measurement in field ergonomics research and application. Int. J. Ind. Ergon. 39, 333e340. Burdorf, A., 1992 Feb. Exposure assessment of risk factors for disorders of the back in occupational epidemiology. Scand. J. Work Environ. Health 18 (1), 1e9. Burdorf, A., Rossignol, M., Fathallah, F.A., Snook, S.H., Herrick, R.F., 1997 Aug. Challenges in assessing risk factors in epidemiologic studies on back disorders. Am. J. Ind. Med. 32 (2), 142e152. Capodaglio, P., Jensen, C., Christensen, H., 1996 JuleAug. Quantification of muscular activity in the shoulder region during monotonous repetitive work. Med. Lav 87 (4), 305e313.
454
C. Trask et al. / International Journal of Industrial Ergonomics 40 (2010) 448e454
da Costa, B.R., Vieira, E.R., 2009 Sep 14. Risk factors for work-related musculoskeletal disorders: a systematic review of recent longitudinal studies. Am. J. Ind. Med. 14, 14. Earle-Richardson, G., Jenkins, P., Fulmer, S., Mason, C., Burdick, P., May, J., 2008 May. An ergonomic intervention to reduce back strain among apple harvest workers in New York State. Appl. Ergon. 36 (3), 327e334. Fathallah, F.A., Marras, W.S., Parnianpour, M., 1998 Mar 15. An assessment of complex spinal loads during dynamic lifting tasks. Spine (Phila Pa 1976) 23 (6), 706e716. Jackson, J.A., Mathiassen, S.E., Dempsey, P.G., 2009 Jun. Methodological variance associated with normalization of occupational upper trapezius EMG using submaximal reference contractions. J. Electromyogr. Kinesiol. 19 (3), 416e427. Jin, S., McCulloch, R., Mirka, G.A., 2009. Biomechanical evaluation of postures assumed when harvesting bush crops. Int. Arch. Occup. Environ. Health 39, 347e352. Jones, T., Kumar, S., 2007a. Assessment of physical demands and comparison of multiple exposure definitions in a repetitive sawmill job: board edger operator. Ergonomics 50 (5), 676e693. Jones, T., Kumar, T., 2007b. Assessment of physical exposures and comparison of exposure definitions in a repetitive sawmill occupation: trim-saw operator. Work 28 (2), 183e196. Jonsson, B., 1988. The static load component in muscle work. Eur. J. Appl. Physiol. Occup. Physiol. 57, 305e310. Keir, P.J., MacDonell, C.W., 2004 Feb 26. Muscle activity during patient transfers: a preliminary study on the influence of lift assists and experience. Ergonomics 47 (3), 296e306. Kerr, M.S., Frank, J.W., Shannon, H.S., Norman, R.W., Wells, R.P., Neumann, W.P., et al., 2001 Jul. Biomechanical and psychosocial risk factors for low back pain at work. Am. J. Public Health 91 (7), 1069e1075. Keyserling, W.M., 2000. Workplace risk factors and occupational musculoskeletal disorders, part 1: a review of biomechanical and psychophysical research on risk factors associated with low-back pain. AIHAJ 61 (1), 39e50. Kothiyal, K., Kayis, B., 2001. Workplace layout for seated manual handling tasks: an electromyography study. Int. J. Ind. Ergon. 27 (1), 19e32. Kumar, S., 1990 Dec. Cumulative load as a risk factor for back pain. Spine 15 (12), 1311e1316. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Gacki-Smith, J., Kohok, A.K., 2007a. Designing ergonomic interventions for EMS workers, Part I: transporting patients down the stairs. Appl. Ergon. 38 (1), 71e81. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Kohok, A.K., Gacki-Smith, J., 2007b. Designing ergonomic interventions for emergency medical services workersepart III: bed to stairchair transfers. Appl. Ergon. 38 (5), 581e589. Lavender, S.A., Conrad, K.M., Reichelt, P.A., Kohok, A.K., 2007c. Gacki-Smith J. Designing ergonomic interventions for EMS workers e part II: lateral transfers. Appl. Ergon. Mar 38 (2), 227e236. Marras, W.S., Ferguson, S.A., Burr, D., Davis, K.G., Gupta, P., 2004 JaneFeb. Spine loading in patients with low back pain during asymmetric lifting exertions. Spine J. 4 (1), 64e75. Marras, W.S., Sommerich, C.M., 1991 Apr. A three-dimensional motion model of loads on the lumbar spine: II. Model validation. Hum. Factors 33 (2), 139e149. Mathiassen, S.E., Burdorf, A., van der Beek, A.J., 2002 Feb. Statistical power and measurement allocation in ergonomic intervention studies assessing upper trapezius EMG amplitude. A case study of assembly work. J. Electromyogr. Kinesiol. 12 (1), 45e57. Mientjes, M.I., Norman, R.W., Wells, R.P., McGill, S.M., 1999 Jun. Assessment of an EMG-based method for continuous estimates of low back compression during asymmetrical occupational tasks. Ergonomics 42 (6), 868e879. Moller, T., Mathiassen, S.E., Franzon, H., Kihlberg, S., 2004 Jan 15. Job enlargement and mechanical exposure variability in cyclic assembly work. Ergonomics 47 (1), 19e40. Nelson, N., Hughes, R.E., 2009. Quantifying relationships between selected workrelated risk factors and back pain: a systematic review of objective
biomechanical measures and cost-related health outcomes. Int. J. Ind. Ergon. 39, 202e210. NIOSH, 1992 March. Selected Topics in Surface Electromyography for Use in the Occupational Setting: Expert Perspective. Report No.: Publication No. DHHS (NIOSH). 91e100. Nordander, C., Balogh, I., Mathiassen, S.E., Ohlsson, K., Unge, J., Skerfving, S., et al., 2004 Aug. Precision of measurements of physical workload during standardised manual handling. Part I: surface electromyography of m. trapezius, m. infraspinatus and the forearm extensors. J. Electromyogr. Kinesiol. 14 (4), 443e454. Norman, R., Wells, R., Neumann, P., Frank, J., Shannon, H., Kerr, M., 1998/12. A comparison of peak vs cumulative physical work exposure risk factors for the reporting of low back pain in the automotive industry. Clin. Biomech. 13 (8), 561e573. Østensvik, T., Veiersted, K.B., Cuchet, E., Nilsen, P., Hanse, J.J., Carlzon, C., et al., 2008 NoveDec. A search for risk factors of upper extremity disorders among forest machine operators: a comparison between France and Norway. Int. J. Ind. Ergon. 38 (11e12), 1017e1027. Potvin, J.R., 2008 Dec. Occupational spine biomechanics: a journey to the spinal frontier. J. Electromyogr. Kinesiol. 18 (6), 891e899. Schultz, A.B., Andersson, G.B., 1981 JaneFeb. Analysis of loads on the lumbar spine. Spine 6 (1), 76e82. Seidler, A., Bolm-Audorff, U., Heiskel, H., Henkel, N., Roth-Kuver, B., Kaiser, U., et al., 2001 Nov. The role of cumulative physical work load in lumbar spine disease: risk factors for lumbar osteochondrosis and spondylosis associated with chronic complaints. Occup. Environ. Med. 58 (11), 735e746. Sporrong, H., Sandsjo, L., Kadefors, R., Herberts, P., 1999 Dec. Assessment of workload and arm position during different work sequences: a study with portable devices on construction workers. Appl. Ergon. 30 (6), 495e503. Straker, L., Mekhora, K., 2000 Sep. An evaluation of visual display unit placement by electromyography, posture, discomfort and preference. Int. J. Ind. Ergon. 26 (3), 389e398. Teschke, K., Trask, C., Village, J., Johnson, P., Koehoorn, M., 2009. Measuring posture for epidemiology: comparing inclinometry, observations, and self-reports. Ergonomics 52 (9), 1067e1078. Trask, C., Teschke, K., Village, J., Chow, Y., Johnson, P., Luong, N., et al., 2007. Measuring low back injury risk factors in challenging work environments: an evaluation of cost and feasibility. Am. J. Ind. Med. 50 (9), 687e696. Trask C.M., Teschke K., Koehoorn M., 2008. Back Pain and Disability Reporting by Workers in Heavy Industry: Implications for Case-control Studies. Association of Canadian Ergonomists 39th Annual Conference October 5e8 Alymer, Quebec, Canada. Village, J., Frazer, M., Cohen, M., Leyland, A., Park, I., Yassi, A., 2005 Sep. Electromyography as a measure of peak and cumulative workload in intermediate care and its relationship to musculoskeletal injury: an exploratory ergonomic study. Appl. Ergon. 36 (5), 609e618. Village, J., Trask, C., Luong, N., Chow, Y., Johnson, P., Koehoorn, M., et al., 2009. Development and evaluation of an observational Back-Exposure Sampling Tool (Back-EST) for work-related back injury risk factors. Appl. Ergon. 40, 538e544. Waters, T.R., Putz-Anderson, V., Garg, A., Fine, L.J., 1993 Jul. Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics 36 (7), 749e776. Wells, R., Norman, R., Neumann, P., Andrews, D., Frank, J., Shannon, H., et al., 1997 Jan. Assessment of physical work load in epidemiologic studies: common measurement metrics for exposure assessment. Ergonomics 40 (1), 51e61. Winkel, J., Mathiassen, S.E., 1994 Jun. Assessment of physical work load in epidemiologic studies: concepts, issues and operational considerations. Ergonomics 37 (6), 979e988. WorkSafeBC, 2006. From awareness to action statistics 2005. Accessed from. http:// www.worksafebc.com/publications/reports/statistics_reports/assets/pdf/stats 2005.pdf (accessed 22.11.2006.).