- Email: [email protected]
Upper trapezius muscle activity patterns during repetitive manual material handling and work with a computer mouse.
Journal of Electromyography and Kinesiology 9 (1999) 317–325 www.elsevier.com/locate/jelekin
Upper trapezius muscle activity patterns during repetitive manual material handling and work with a computer mouse. C. Jensen *, L. Finsen, K. Hansen, H. Christensen Department of Research on Monotonous Repetitive Work, National Institute of Occupational Health, DK-2100 Copenhagen 0, Denmark Received for publication 2 March 1999
Abstract Firstly, upper trapezius EMG activity patterns were recorded on the dominant side of 6 industrial production workers and on the side operating a computer mouse of 14 computer-aided design (CAD) operators to study differences in acute muscular response related to the repetitiveness of the exposure. The work tasks were performed with median arm movement frequencies ranging from 5 min−1 to 13 min−1 and were characterized by work cycle times ranging from less than 30 sec to several days. However, the static and median EMG levels and EMG gap frequencies were similar for all work tasks indicating that shoulder muscle loads may be unaffected by large variations in arm movement frequencies and work cycle times. An exposure variation analyses (EVA) showed that the EMG activity patterns recorded during production work were more repetitive than during CAD work, whereas CAD work was associated with more static muscle activity patterns, both may be associated with a risk of developing musculoskeletal symptoms. Secondly, upper trapezius EMG activity patterns recorded on the mouse side of the CAD operators were compared with those recorded on the non-mouse side to study differences in muscular responses potentially related to the risk of developing shoulder symptoms which were more prevalent on the mouse side. The number of EMG gaps on the mouse side were significantly lower than the values for the upper trapezius on the non-mouse side indicating that more continous activity was present in the upper trapezius muscle on the mouse side and EVA analyses showed a more repetitive muscle activity pattern on the mouse side. These findings may be of importance to explain differences in the prevalence of shoulder symptoms. 1999 Elsevier Science Ltd. All rights reserved. Keywords: Arm movement; Work cycle time; APDF; EMG-gaps; EVA; Shoulder symptoms
1. Introduction Repetitive monotonous work which represents a risk for developing musculoskeletal disorders, is characterised by highly repetitive movements and often short work cycle times. Silverstein et al. [17] reported that a highly repetitive exposure, defined as work tasks with cycle times below 30 seconds or performing the same movements for more than 50% of the cycle time, was associated with an increased risk of developing hand/wrist disorders. Kilbom [12,13] focused on movements and suggested limits for the number of arm, hand and finger movements which indicated high risk for the development of disorders based on a review of existing literature. As current models describe musculoskeletal health effects of work exposures to be mediated through
* Corresponding author. E-mail: [email protected]
acute physiological responses [1], electromyographic (EMG) recordings of muscle activity have often been used to study exposure/effect associations, especially between exposure levels and shoulder symptoms [2,4,7,9,18,19,22]. Muscular rest periods during work and their influence on the development of symptoms have also been addressed [5,6]. However, detailed knowledge about relationships between work cycle times, arm movements, shoulder muscle activity and musculoskeletal symptoms in occupational settings is still lacking. In a study of computer aided design work (CAD) we reported that musculoskeletal symptoms among CAD-operators were more frequent in the shoulder, arm and hand which almost continously operated a computer mouse than on the contralateral side where the hand was used for intermittent keyboard or paper work [8]. Thus, the 12-month prevalence of symptoms in the shoulder region on the side operating the mouse was 52%, whereas the prevalence was 19% on the
1050-6411/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 9 ) 0 0 0 0 7 - 3
318
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
‘non-mouse’ side (n ⫽ 132). EMG recordings showed statistically higher mean levels of EMG activity on the upper trapezius muscle on the side operating the mouse than on the other side, however, the differences in EMG levels seemed too small to be important for a differential development of disorders in the two shoulders. Thus, EMG analyses that focus on other aspects than the level of muscle activity should be used to study the exposure in repetitive work, preferably focusing on rest breaks and variability of the EMG pattern [14,23]. The aim of the present study was in a first step to study the association between upper trapezius muscle activity patterns and work cycle time, arm movement frequency and the duration of pauses in arm movements by comparing recordings from 3 different work tasks involving repetitive arm movements. In a second step the EMG recordings from both upper trapezius muscles of the CAD-operators were compared with respect to rest breaks and variability. This was done to assess whether other differences existed that were more likely to be associated with the large differences in symptom prevalences between the two sides than the previously reported levels of muscle activity [8]. Thus, all EMG signals were analysed both with respect to the level and the repetitiveness of the EMG patterns including a quantification of EMG micropauses.
was filled up from floor level to the top just above shoulder height. At one of the machines this was performed with a device hanging from the ceiling and operated by both hands (task 1). This enabled the worker to collect and carry a larger number of metal ends at a time than at the other machine, where the ends were collected in stacks weighing 1–4 kg and lifted manually to the container by pressing the metal ends firmly together with both hands (task 2). The CAD-operators were employed in a company of consulting engineers, where they produced technical drawings for a variety of constructions. This was performed on computers usually with a digitizer in front of the computer, a puck in one hand and a keyboard beside the digitizer. In a few cases an ordinary PC-mouse was used without a digitizer. In the present paper both the puck and the PC-mouse is referred to as a ‘mouse’. 2.3. Observations
Six industrial production workers (5 female and 1 male) and 14 CAD-operators (all females) from two different companies participated in the study after having provided written informed consent. The mean age of the production workers was 38 years (SD ⫽ 11 years) and the mean age of the CAD-operators was 39 years (SD ⫽ 10 years).
Observations of cycle times and upper arm movements were performed using video recordings of the workers normal work activities. Regarding upper arm movements the number of repetitive movements across the glenohumeral joint and the duration of pauses was registered for the arm on the dominant side of the production workers and on the side operating the mouse of the CAD-operators. A ‘pause’ was registered when the upper arm was hanging vertically along the side of the body and/or resting with support of the forearm or hand. On the other side of the CAD-operators (non-mouse side) the number of repetitive arm movements was registered. Data were continuously collected using a minicomputer by pressing predefined keys each time an event was observed. Video recordings of the production workers were performed for 10 min simultaneously with EMG recordings. Video recordings of the CAD-operators were performed for 5 min at the beginning of a twenty-five min EMG recording.
2.2. Work tasks
2.4. Electromyography
EMG recordings and observations of cycle times and upper arm movements were obtained while the subjects performed their normal work tasks. The production workers produced ends for metal cans and the data presented here were recorded from employees working at two similar machines. At both workstations the workers collected stacks of metal ends from the machine and packed the ends in containers placed next to the machine. At both machines most of the work time was spent collecting (including waiting for enough ends to collect a stack) and visually inspecting the metal ends on the conveyor belt in front of them. A shorter part of the work time was spent lifting the ends from the conveyor belt to the container, which
Bipolar surface electrodes were used for 10 min EMG recordings of the upper trapezius muscle (m. trapezius, pars desc.) on the dominant side of the production workers (n ⫽ 6). For the CAD-operators 25 min EMG recordings of the upper trapezius muscle were performed both on the mouse side (n ⫽ 14) and on the other side (n ⫽ 7). The center of each electrode pair was placed 2 cm lateral to the midpoint between the seventh cervical vertebrae and the lateral end of acromion with an interelectrode distance of 20 mm [10] (Ag-AgCl electrodes, Medicotest A/S, type 7 00 02-E, Denmark). The EMG signal was amplified, low-pass filtered (8.th order Butterworth filter, cut-off 400 Hz) and sampled on a computer (sampling frequency 1024 Hz). The signals were
2. Methods 2.1. Subjects
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
visually checked, high pass filtered (cut-off 10 Hz), fullwave rectified and root-mean-square (RMS) converted within moving windows of 100 ms duration. The resting EMG signal was recorded during a 5 sec period of instructed rest with visual feedback from an oscilloscope to eliminate visible EMG activity and the resting RMS amplitude was quadratically subtracted from all other EMG signals. For normalization EMG activity was recorded during three voluntary isometric contractions with bilateral 90° arm abduction with resistance just proximal to the elbow [16]. The subjects were instructed and encouraged to produce as much force as possible and hold the maximal force for at least 2–3 sec. The maximal EMG amplitude (EMGmax) during the reference contractions was calculated as the highest mean RMS amplitude obtained with a 1 sec window moving in steps of 100 msec. The EMG activity recorded during work was analyzed according to three different procedures: (1) the amplitude probability distribution function (APDF) analysis which quantifies EMG activity levels [11], (2) the EMG gap analysis which quantifies the numbers and duration of ‘silent’ periods in the EMG pattern, defined as periods with an EMG level below 0.5% EMGmax for at least 0.2 sec (gaps) or at least 0.6 sec (long gaps) [21] and (3) the exposure variation analysis (EVA) which quantifies the relative time that the EMG activity is distributed in categories according to the duration of activity sequences within specified amplitude intervals or classes [14]. The level classes used here were 0–0.3% EMGmax, 0.3–1% EMGmax, 1–3% EMGmax, 3–7% EMGmax, 7– 15% EMGmax, > 15% EMGmax and the time classes by which the duration of uninterrupted EMG sequences were categorised were 0–0.3 s, 0.3–1 s, 1–3 s, 3–7 s, 7– 15 s, > 15 s. The distribution of time spent within level classes and within period length classes are presented separately. 2.5. Design and statistics Firstly, associations between muscle activity patterns and work cycle time, arm movement frequency and the duration of pauses in arm movements were studied by: 쐌 comparing muscle activity patterns recorded on the production workers during the performance of task 1 and task 2, as the repetitiveness of the exposure differed for the two tasks, and by 쐌 comparing the muscle activity patterns recorded during work task 2 with those of the CAD-operators, as the cycle times differed considerably, but both tasks were performed with repetitive arm movements. When the paired recordings from task 1 and task 2 were analyzed by the APDF and gap analyses the Wilcoxon Signed Rank test was used to test for differences. Differences in the distributions of time spent within level
319
classes according to the EVA analyses between task 1 and task 2 were tested with a multivariate paired T2test. A similar but separate test was performed on the distributions within period length classes. The total time spent in different classes add up to 100%, but as the test requires statistically independent variables, data from one level class and from one period length class were excluded before the tests were carried out [15]. For the unpaired recordings from task 2 and from CAD-work the Mann-Whitney test was used to test for differences between the results obtained by APDF and gap analyses. Differences in the distributions of time spent within level classes and within period length classes between task 2 and CAD-work were tested with a MANOVA. Also here data from one level class and from one period length class were excluded before the tests were carried out. Secondly, a possible association between muscle activity patterns and musculoskeletal symptoms was studied among the CAD-operators by comparing the muscle activity patterns recorded on the mouse side with those recorded on the non-mouse side. The Mann-Whitney test was used for the results obtained by the gap analyses and a MANOVA was used for the EVA results as described above. A significance level of 0.05 was used.
3. Results The two work tasks of the production workers could be characterized by well-defined work cycles of relatively short duration and involved repetitive movements of the arms (Table 1). However, differences were found between the two work tasks as the work cycles in task 2 were shorter than 30 seconds, and work task 2 was performed with a higher frequency of upper arm movements than work task 1. Thus, work task 2 was characterized both by more repetitive work cycles and more repetitive movements than task 1. For the CAD-operators the duration of the work cycles was highly variable, lasting from minutes to days, and cycle times were not Table 1 Work cycle times and arm movements during two tasks involving manual material handling (metal can production) and during computeraided design (CAD) work based on observations of six workers for each task. Median values (min–max) are shown. Metal can production Task 1 Work cycle time (sec) 53 (41–64) Frequency of upper arm 5 (3–7) movements (min−1) Upper arm pause (% time) 37 (24–55) a
Based on observations of 14 subjects.
CAD-work
Task 2 23 (18–30) 13 (8–18)
12 (9–16)
30 (14–44)
2 (0–43)a
320
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
registered. The median frequency of upper arm movements during CAD-work on the mouse side was similar to that of task 2 for the production workers. Thus, the work tasks of the CAD-operators were characterized by repetitive movements, but not by repetitive work cycles. The median duration of pauses in upper arm movements was similar for the two tasks in metal-can production, but shorter for CAD-work (Table 1). Examples of RMS converted EMG recordings during task 1 and task 2 of one production worker and on the mouse side of one CAD-operator are shown in Fig. 1. For the whole work group the EMG recordings showed similar results for the two work tasks in metal can production both with respect to the median values of the level of EMG activity, the EMG gap pattern and the pattern of repetitiveness (Table 2). The static and median EMG levels were about 1% and 5% EMGmax, respectively, which was also similar to the EMG activity recorded during CAD-work. The tendency to lower peak EMG levels recorded during CAD-work as compared to task 2 in production work was not statistically signifi-
cant. Likewise, no significant differences were found in the frequency or duration of gaps between EMG recordings from CAD-work and work task 2. At the individual level there was no association between the recorded EMG gap time and the observed time that the arm rested on the table or hang passively along the side of the body (Fig. 2). The EVA analyses showed that the EMG level during CAD-work was between 3% and 7% EMGmax for almost half of the work time, whereas this was only the case for about 30% of the work time during production work. The difference in the distribution of time within different level classes between CAD-work and task 2 was almost significant (p ⫽ 0.056). Furthermore, the distribution of time spent in different period length classes was skewed in the direction of more time spent within shorter period lengths during work task 2 compared to CAD-work (p ⫽ 0.012). During CAD-work the median frequency of upper arm movements on the non-mouse side was 2 min−1 (range: 1–4 min−1), i.e. considerably lower than the number of repetitive arm movements on the mouse side (12 min−1,
Fig. 1. Examples of 5 minutes of EMG recordings from one production worker during the performance of task 1 (A) and task 2 (B) shown as the RMS-amplitude. Similarly, an example of an EMG recording from the mouse side of one CAD-operator is shown (C).
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
Table 2 EMG activity patterns of the upper trapezius muscle recorded during metal can production on the dominant side (two work tasks, n ⫽ 6) and during computer-aided design (CAD) work on the mouse side (n ⫽ 14). Median values (min–max) are shown. Metal can production Task 1 APDF-analyses: Static level 0.8 (0.5–2.9) (% EMGmax) Median level 4.2 (1.8–11.7) (% EMGmax) Peak level 15.8 (3.9–20.9) (% EMGmax) Gap analyses: Gaps, > 0.2 s 1.1 (0.6–10.0) (min−1) Long gaps, > 0.1 (0.0–2.6) 0.6 s (min−1) Gap time 1.0 (0.2–8.9) (% time) EVA analyses: Level in % % time EMGmax: 0–0.3 0.3 (0.0–6.5) 0.3–1 10.7 (1.2–25.1) 1–3 18.2 (5.7–54.5) 3–7 29.8 (12.9–40.9) 7–15 21.9 (0.3–50.6) > 15 10.3 (0.0–29.3) Period length in sec: 0–0.3 20.3 (9.9–30.2) 0.3–1 24.0 (15.6–33.4) 1–3 32.2 (27.1–37.4) 3–7 15.0 (6.8–24.6) 7–15 4.2 (1.5–15.0) > 15 1.2 (0.0–8.0) a
CAD-work
Task 2
1.2 (0.6–4.8)
1.4 (0.3–7.3)a
5.3 (2.5–16.1)
4.5 (1.0–11.8)a
14.5 (5.0–33.7)
8.4 (3.1–17.9)a
1.3 (0.0–8.2)
1.6 (0.0–19.0)
0.1 (0.0–1.4)
0.7 (0.0–5.8)
0.6 (0.0–5.8)
2.9 (0.0–18.1)
321
Table 1). The patterns of EMG gaps recorded on the two upper trapezius muscles during CAD-work were also significantly different (Table 3). The frequency of gaps was about ten-fold lower on the side operating the mouse than on the other side and the total gap time was sixfold shorter on the mouse side as compared to the other side. The distribution of time spent within specific level classes according to the EVA analyses showed more time spent at higher level classes on the mouse side (p ⫽ 0.036, Fig. 3). When analyzing the time spent within period length classes it was found that for the recordings on the mouse side a considerable part of the work time was spent within period length classes of 1– 3 s and 3–7 s while for the recordings on the other side the time spent within different classes was more evenly distributed between short, medium and long period lengths (p ⬍ 0.001). 4. Discussion
% time
% time
0.0 8.2 20.2 28.7 22.9 7.6
(0.0–1.0) 0.8 (0.0–6.9) (0.1–15.5) 3.8 (0.0–41.1) (3.4–46.2) 18.7 (0.3–66.2) (5.7–47.5) 46.7 (4.2–79.3) (3.2–46.7) 16.8 (0.2–76.2) (0.0–59.2) 0.9 (0.0–19.3)
21.8 31.0 35.0 11.5 1.6 0.0
(14.0–25.4) (19.9–34.8) (27.3–40.4) (6.9–20.4) (0.0–9.1) (0.0–6.7)
11.3 16.9 30.8 23.8 12.1 2.8
(9.2–19.9) (11.0–28.4) (24.1–38.9) (11.6–35.0) (1.1–19.5) (0.0–13.4)
Previously published in Jensen et al. [8].
4.1. Arm movements and upper trapezius muscle activity CAD-work using a mouse and both of the described work tasks in production work may be characterized as repetitive because a high number of similar arm movements were performed in each task for a major part of the work time (even though the movements during CADwork were not similar to the movements during production work). 4.1.1. Task 1 versus task 2 However, task 1 in production work seemed only half as repetitive as task 2 when comparing arm movement frequencies and work cycle times. The differences in repetitiveness between task 1 and task 2 were not reflected in the EMG patterns of the upper trapezius muscle indicating that the muscular load was the same when performing these tasks, i.e. a larger difference in the number of movements is required to alter the muscle activity pattern. Thus, job rotation between the two work tasks, as was normal procedure in the company, may not Table 3 EMG gaps in the recordings from the upper trapezius muscle on the mouse side (n ⫽ 14) and the non-mouse side (n ⫽ 7) recorded during computer-aided design (CAD) work. Upper trapezius non- Upper trapezius mouse side mouse side Gaps, > 0.2 s (min−1) Long gaps, > 0.6 s (min−1) Gap time (% time)
Fig. 2. Total EMG gap time vs the time that the arm is resting on the table or hanging passively along the side of the body (arm pause) for each worker.
18.0 (2.4–43.1)* 9.3 (1.1–17.0)* 17.3 (6.2–32.8)**
1.6 (0.0–6.1) 0.7 (0.0–2.3) 2.9 (0.0–5.3)
Median values (min–max) are shown. Significant pairwise differences are indicated with * p ⬍ 0.05 or **p ⬍ 0.01.
322
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
Fig. 3. Time spent within level—and period length classes according to EVA analyses of EMG recordings during CAD-work. Median percentages of the time recorded on the upper trapezius muscle on the mouse side (n ⫽ 14) and on the non-mouse side (n ⫽ 7) are shown.
create variation in shoulder muscle loads and consequently may not affect the risk of developing shoulder muscle disorders. The possibility exists that the EMG analysis procedures were not suitable to detect actual differences in the temporal RMS amplitude pattern. However, it seems unlikely that a more detailed analysis than the EVA analysis would be necessary to detect aspects of the EMG pattern that are important for the development of muscular disorders. The main problem seems to reduce the amount of possible information obtained by the EVA analysis and extract only the relevant information. 4.1.2. CAD-work versus task 2 During CAD-work with a similar arm movement frequency to that of task 2 the median level of EMG activity and the number of EMG gaps were also similar. The total gap time was not associated with the observed duration of upper arm pauses neither between tasks nor between individuals performing similar tasks. This is a similar finding to that of a study of light material handling and office work, where no correlation was found between time with an arm posture below 10° elevation and time with an upper trapezius EMG activity below
0.5% EMGmax [20]. The peak levels of EMG activity showed a tendency to be higher during production work. This was probably due to greater arm movements and handling of stacks of metal ends, which required higher arm and shoulder muscle forces during a minor part of the work cycle. However, caution should be exerted in interpreting the actual peak levels of activity during production work as the EMG activity was normalized in relation to a static contraction, but during most of the work time the upper arms were only slightly elevated, making estimates of static and median levels of EMG activity more reliable than the peak levels. This was shown in more detail by the EVA analyses, which also showed that a larger proportion of the work time was spent in shorter period length classes during production work. This indicated that production work was associated with a more repetitive muscle activity pattern than CAD-work, which seemed contradictory to the observed similar degree of repetitiveness of these two types of work with respect to arm movement frequencies. However, the EMG differences may also have been due to different arm postures. CAD-work was performed with a static posture, smaller arm movements and higher precision demands than production work which required a
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
323
static or intermittent-static activation pattern of the upper trapezius muscle with longer stays within the same level classes of the EVA classification system. A comparison of the risk of developing disorders between production and CAD-work based on recordings of exposure and muscle activity patterns would be difficult as both shortcycled repetitive muscle activity patterns as recorded during production work and more static activation patterns as recorded during CAD-work may represent a risk. The common factor, which may lead to an assessment of the risk as being high in both jobs, was the few EMG gaps recorded during all 3 work tasks. This indicated that part of the muscle was continuously active during work. Continuous activity in low-threshold motor units has been suggested to cause musculoskeletal disorders [5], but more scientific evidence is needed to use gap analyses in exposure assessment.
differences in several EMG variables between recordings from the two sides makes it difficult to identify EMG gaps as the single key factor causally related to the development of disorders. The level of EMG activity was also significantly higher on the mouse side (median level: 4.4% EMGmax) than on the other side (median level: 1.9% EMGmax), but it should be realized that the levels were small on both sides. Consequently, the small difference between the two sides appears physiologically insignificant compared to the considerably larger difference in EMG gap time and gap frequency. The EVA analyses showed in more detail that the fraction of the recording with low muscle activity levels, e.g. below 1% EMGmax, was much higher on the non-mouse side than on the other side, clearly emphasizing that muscular rest was seldom present for the upper trapezius muscle on the mouse side during mouse work.
4.2. CAD-work: Side differences in muscle activity patterns and symptom prevalence
4.2.2. Repetitiveness of the muscle activity The distribution of stays within different period length classes was also different between the two sides as the EMG recordings from the non-mouse side showed a more equal distribution between different period length classes, i.e. a more variable EMG pattern than recorded on the mouse side. A priori this seems favorable, however, whether the actual differences influence the risk of developing muscular disorders is not known.
4.2.1. Gaps as risk indicator During CAD-work the number of EMG gaps recorded on the upper trapezius muscle on the non-mouse side was considerably higher and the total gap time was much longer than on the mouse side. In view of the previously reported differences in shoulder symptom prevalence this indicates that lack of EMG gaps may be associated with the development of disorders. Similar to the present study some studies have shown that lack of gaps may be a risk factor for shoulder disorders among workers performing light manual work or office work [6,22], although others have found no difference in EMG gaps between workers with and without shoulder complaints [9,19]. In all of the above-mentioned studies symptomatic and non-symptomatic workers with identical jobs were compared. However, the ability to identify EMG gaps as a risk factor may have been weakened if other factors such as the individual vulnerability to acquire such disorders show large interindividual variation. In the present study the different prevalence of symptoms on the two shoulders were based on reports from a larger group of workers with highly different exposures on the two sides of the same individual. Thus, the mean differences between muscle activity parameters measured on the two shoulders were largely due to exposure contrasts and the large differences in some of the EMG parameters may be important factors influencing the development of pain. When comparing with the study by Veiersted et al. [22] the number of gaps recorded on symptomatic workers (5–10 gaps min−1) in their study was higher than the median number of gaps on the mouse side of the CAD-operators and the number of gaps recorded on nonsymptomatic workers (10–15 gaps min−1) in their study was lower than the median number of gaps recorded on the non-mouse side of the CAD-operators. However, the
5. Conclusions 5.1. Arm movement frequencies Firstly, in repetitive work the exposure should be changed considerably to alter shoulder muscle loads. A decrease in upper arm movement frequency from 13 to 5 movements min−1 during production work had no significant effect on upper trapezius muscle activity patterns such as the number of EMG gaps, whereas a comparison between the mouse- and non-mouse side during CADwork corresponding to arm movement frequencies of 12 and 2 min−1 revealed significant differences in muscle activity patterns. The difference in the number of gaps was especially pronounced indicating that given a sufficiently large change in exposure this parameter may be sensitive to such changes. In contrast no difference in the number of gaps was found between recordings from CAD-work and production work indicating that certain aspects of the exposure of these different work groups were similar. Thus, even though CAD-work and manual material handling differ in many respects and other factors than the movement frequency influence muscle activity, both may be characterized as repetitive work and the data may be taken as support of the guidelines presented by Kilbom [12]. She suggested that an upper arm movement frequency exceeding 2.5 min−1 during
324
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
repetitive work is associated with a high risk of developing musculoskeletal disorders. 5.2. Work cycle times Secondly, a general use of work cycle times to characterize work appears inadequate both to define work as repetitive and to predict muscle loads. A lack of association between work cycle time and forearm muscle load during repetitive meat-cutting work has also been reported by Christensen et al. [3]. Thus, movement frequencies rather than cycle times should be preferred to characterize repetitive work, especially if the aim is to predict the risk of developing disorders. 5.3. EMG analysis methods Thirdly, muscle activity patterns recorded during repetitive work should be analyzed according to other methods than the APDF analysis to obtain relevant information on the muscular load. In principle the EVA method appears suitable to provide all the information which is obtained by the APDF and the gap methods along with unique information regarding the repetitiveness of the muscle activity pattern. However, more efforts should be invested in studying how to interpret the relatively large amounts of information obtained by the EVA procedure. It seems insufficient to focus on the degree of repetitiveness, i.e. the time with stays in short period length classes (0–0.3 sec and 0.3–1 sec) as static muscle activity will show up as time spent in the longest period length classes (7–15 sec and > 15 sec). Both repetitive and static muscle activity should be avoided, whereas a variable muscle activity pattern may have positive health effects. A variable muscle activity pattern may be characterized by an equal partitioning of the time between all period length classes and some partitioning of the time between the level classes, including sufficient time at the lowest level classes to ensure adequate restitution between activity periods. More effort should also be invested in studying whether all the information of the EVA analysis is necessary to predict the risk of developing disorders. If EMG gap patterns are critical for the development of muscular problems it would be wiser to focus directly on these analyses. Future research should resolve these questions.
Acknowledgements We thank Henrik B. Olsen and Paolo Capodaglio for their assistance with data collection and analysis.
References [1] Armstrong TJ, Buckle P, Fine LJ. et al., A conceptual model for work-related neck and upper-limb musculoskeletal disorders. Scand J Work Environ Health 1993;19(2):73–84. [2] Christensen H. Muscle activity and fatigue in the shoulder muscles of assembly-plant employees. Scand J Work Environ Health 1986;12(6):582–7. [3] Christensen H, Søgaard K, Pilegaard M, et al. The importance of the work/rest pattern as a risk factor in repetitive monotonous work (submitted). [4] Hagberg M, Sundelin G. Discomfort and load on the upper trapezius muscle when operating a wordprocessor. Ergonomics 1986;29(12):1637–45. [5] Ha¨gg GM. Static work loads and occupational myalgia - a new explanation model. In: Anderson PA, Hobart DJ, Danoff JV, editors. Electromyographical Kinesiology. Elsevier Science Publishers B.V., 1991:141–144. [6] Ha¨gg GM, Åstro¨m A. Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders. Int Arch Occup Environ Health 1997;69:423–32. [7] Jensen BR, Schibye B, Søgaard K. et al., Shoulder muscle load and muscle fatigue among industrial sewing-machine operators. Eur J Appl Physiol 1993;67:467–75. [8] Jensen C, Borg V, Finsen L. Job demands, muscle activity and musculoskeletal symptoms in relation to computer mouse work. Scand J Work Environ Health 1998;24(5):418–24. [9] Jensen C, Nilsen K, Hansen K. et al., Trapezius muscle load as a risk indicator for occupational shoulder-neck complaints. Int Arch Occup Environ Health 1993;64:415–23. [10] Jensen C, Vasseljen O, Westgaard RH. The influence of electrode position on bipolar surface electromyogram recordings of the upper trapezius muscle. Eur J Appl Physiol 1993;67:266–73. [11] Jonsson B. Measurement and evaluation of local muscular strain in the shoulder during constrained work. J Hum Ergol 1982;11:73–88. [12] Kilbom Å. Repetitive work of the upper extremity: Part I- Guidelines for the practitioner. Int J Ind Erg 1994;14:51–7. [13] Kilbom Å. Repetitive work of the upper extremity: Part II - The scientific basis (knowledge base) for the guide. Int J Ind Erg 1994;14:59–86. [14] Mathiassen SE, Winkel J. Quantifying variation in physical load using exposure-vs-time data. Ergonomics 1991;34(12):1455–68. [15] Mathiassen SE, Winkel J. Physiological comparison of three interventions in light assembly work: reduced work pace, increased break allowance and shortened working days. Int Arch Occup Environ Health 1996;68:94–108. [16] Mathiassen SE, Winkel J, Ha¨gg GM. Normalization of surface EMG amplitude from the upper trapezius muscle in ergonomic studies - A review. J Electromyogr Kinesiol 1995;5(4):197–226. [17] Silverstein BA, Fine LJ, Armstrong TJ. Hand wrist cumulative trauma disorders in industry. Br J Ind Med 1986;43(11):779–84. [18] Takala E, Viikari-Juntura E. Muscular activity in simulated light work among subjects with frequent neck-shoulder pain. Int J Ind Erg 1991;8:157–64. [19] Vasseljen O, Westgaard RH. A case-control study of trapezius muscle activity in office and manual workers with shoulder and neck pain and symptom-free controls. Int Arch Occup Environ Health 1995;67:11–18. [20] Vasseljen O, Westgaard RH. Arm and trunk posture during work in relation to shoulder and neck pain and trapezius activity. Clin Biomech 1997;12:22–31. [21] Veiersted KB, Westgaard RH, Andersen P. Pattern of muscle activity during sterotyped work and its relation to muscle pain. Int Arch Occup Environ Health 1990;62:31–41. [22] Veiersted KB, Westgaard RH, Andersen P. Electromyographic
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
evaluation of muscular work pattern as a predictor of trapezius myalgia. Scand J Work Environ Health 1993;19(4):284–90. [23] Winkel J, Westgaard R. Occupational and individual risk factors for shoulder-neck complaints: Part II—The scientific basis (literature review) for the guide. Int J Ind Erg 1992;10:85–104. Chris Jensen received the M.Sc degree in zoophysiology at the University of Trondheim in Norway. The Ph.D. thesis was received in 1995 after studies at the Norwegian Institute of Technology where he became trained in work physiology. Chris Jensen is a senior researcher since 1998 at the National Institute of Occupational Health, Denmark where he has been engaged in research on monotonous, repetitive work since 1995. Chris Jensen’s research at the University of Trondheim has focused on understanding the physiology of shoulder muscles especially in relation to repetitive work. At the National Institute of Occupational Health in Denmark his research is focused on work exposures and risk factors for the development of musculoskeletal disorders in computer operators. Lotte Finsen received the M.Sc degree in biology and physical training in 1991 and the Ph.D. degree in physiology in 1995, University of Copenhagen. Since 1989 she has been working with basic and applied physiology at the department of Physiology at the National Institute of Occupational Health in Copenhagen. Her research interests focus on the mechanisms behind development of occupational related musculoskeletal disorders which she studies using electromyography, kinesiology and biomechanical models. From 1994–1998 she has been a researcher at the Department for research in repetitive monotonous
325
work, and from 1998 at the Department of Visual Display Unit Work. She serves as a referee for scientific journals. Klaus Hansen is a physiotherapist since 1985. He has worked at Bispebjerg Hospital, Copenhagen from 1985 to 1995, from 1991 to 1994 as department head-physiotherapist. Klaus Hansen has worked at the National Institute of Occupational Health, Denmark since 1995, where he has been engaged in research on monotonous, repetitive work and computer work. His primary focus is on the effect of work factors on the physical capacity of the human body.
Hanne Christensen received the M.Sc degree in physiology and the Ph.D. degree in medicine from Copenhagen University. She has been working in the area of work physiology in the department of Physiology, National Institute of Occupational Health, Copenhagen, since 1986, and as a senior researcher since 1991. From 1994–1998 she has been head of the Department for research in repetitive monotonous work, and 1998–2001 head of the Department of Visual Display Unit Work, both at the Dainsh NIOH. Her research interest is mechanism behind developing of musculoskeletal symptoms and disorders, especially quantitated by electromyography. She serves as a referee for several scientific journals and is a member of the editorial board in Scandinavian Journal of Work, Environment & Health. She is awarded from The Danish Society of casualties from Polio, Road Accidents and Accidents (1986), from The Foundation of Work Protections (1987) and The Danish Society of Manual Medicine (1987).
Upper trapezius muscle activity patterns during repetitive manual material handling and work with a computer mouse. C. Jensen *, L. Finsen, K. Hansen, H. Christensen Department of Research on Monotonous Repetitive Work, National Institute of Occupational Health, DK-2100 Copenhagen 0, Denmark Received for publication 2 March 1999
Abstract Firstly, upper trapezius EMG activity patterns were recorded on the dominant side of 6 industrial production workers and on the side operating a computer mouse of 14 computer-aided design (CAD) operators to study differences in acute muscular response related to the repetitiveness of the exposure. The work tasks were performed with median arm movement frequencies ranging from 5 min−1 to 13 min−1 and were characterized by work cycle times ranging from less than 30 sec to several days. However, the static and median EMG levels and EMG gap frequencies were similar for all work tasks indicating that shoulder muscle loads may be unaffected by large variations in arm movement frequencies and work cycle times. An exposure variation analyses (EVA) showed that the EMG activity patterns recorded during production work were more repetitive than during CAD work, whereas CAD work was associated with more static muscle activity patterns, both may be associated with a risk of developing musculoskeletal symptoms. Secondly, upper trapezius EMG activity patterns recorded on the mouse side of the CAD operators were compared with those recorded on the non-mouse side to study differences in muscular responses potentially related to the risk of developing shoulder symptoms which were more prevalent on the mouse side. The number of EMG gaps on the mouse side were significantly lower than the values for the upper trapezius on the non-mouse side indicating that more continous activity was present in the upper trapezius muscle on the mouse side and EVA analyses showed a more repetitive muscle activity pattern on the mouse side. These findings may be of importance to explain differences in the prevalence of shoulder symptoms. 1999 Elsevier Science Ltd. All rights reserved. Keywords: Arm movement; Work cycle time; APDF; EMG-gaps; EVA; Shoulder symptoms
1. Introduction Repetitive monotonous work which represents a risk for developing musculoskeletal disorders, is characterised by highly repetitive movements and often short work cycle times. Silverstein et al. [17] reported that a highly repetitive exposure, defined as work tasks with cycle times below 30 seconds or performing the same movements for more than 50% of the cycle time, was associated with an increased risk of developing hand/wrist disorders. Kilbom [12,13] focused on movements and suggested limits for the number of arm, hand and finger movements which indicated high risk for the development of disorders based on a review of existing literature. As current models describe musculoskeletal health effects of work exposures to be mediated through
* Corresponding author. E-mail: [email protected]
acute physiological responses [1], electromyographic (EMG) recordings of muscle activity have often been used to study exposure/effect associations, especially between exposure levels and shoulder symptoms [2,4,7,9,18,19,22]. Muscular rest periods during work and their influence on the development of symptoms have also been addressed [5,6]. However, detailed knowledge about relationships between work cycle times, arm movements, shoulder muscle activity and musculoskeletal symptoms in occupational settings is still lacking. In a study of computer aided design work (CAD) we reported that musculoskeletal symptoms among CAD-operators were more frequent in the shoulder, arm and hand which almost continously operated a computer mouse than on the contralateral side where the hand was used for intermittent keyboard or paper work [8]. Thus, the 12-month prevalence of symptoms in the shoulder region on the side operating the mouse was 52%, whereas the prevalence was 19% on the
1050-6411/99/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 9 ) 0 0 0 0 7 - 3
318
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
‘non-mouse’ side (n ⫽ 132). EMG recordings showed statistically higher mean levels of EMG activity on the upper trapezius muscle on the side operating the mouse than on the other side, however, the differences in EMG levels seemed too small to be important for a differential development of disorders in the two shoulders. Thus, EMG analyses that focus on other aspects than the level of muscle activity should be used to study the exposure in repetitive work, preferably focusing on rest breaks and variability of the EMG pattern [14,23]. The aim of the present study was in a first step to study the association between upper trapezius muscle activity patterns and work cycle time, arm movement frequency and the duration of pauses in arm movements by comparing recordings from 3 different work tasks involving repetitive arm movements. In a second step the EMG recordings from both upper trapezius muscles of the CAD-operators were compared with respect to rest breaks and variability. This was done to assess whether other differences existed that were more likely to be associated with the large differences in symptom prevalences between the two sides than the previously reported levels of muscle activity [8]. Thus, all EMG signals were analysed both with respect to the level and the repetitiveness of the EMG patterns including a quantification of EMG micropauses.
was filled up from floor level to the top just above shoulder height. At one of the machines this was performed with a device hanging from the ceiling and operated by both hands (task 1). This enabled the worker to collect and carry a larger number of metal ends at a time than at the other machine, where the ends were collected in stacks weighing 1–4 kg and lifted manually to the container by pressing the metal ends firmly together with both hands (task 2). The CAD-operators were employed in a company of consulting engineers, where they produced technical drawings for a variety of constructions. This was performed on computers usually with a digitizer in front of the computer, a puck in one hand and a keyboard beside the digitizer. In a few cases an ordinary PC-mouse was used without a digitizer. In the present paper both the puck and the PC-mouse is referred to as a ‘mouse’. 2.3. Observations
Six industrial production workers (5 female and 1 male) and 14 CAD-operators (all females) from two different companies participated in the study after having provided written informed consent. The mean age of the production workers was 38 years (SD ⫽ 11 years) and the mean age of the CAD-operators was 39 years (SD ⫽ 10 years).
Observations of cycle times and upper arm movements were performed using video recordings of the workers normal work activities. Regarding upper arm movements the number of repetitive movements across the glenohumeral joint and the duration of pauses was registered for the arm on the dominant side of the production workers and on the side operating the mouse of the CAD-operators. A ‘pause’ was registered when the upper arm was hanging vertically along the side of the body and/or resting with support of the forearm or hand. On the other side of the CAD-operators (non-mouse side) the number of repetitive arm movements was registered. Data were continuously collected using a minicomputer by pressing predefined keys each time an event was observed. Video recordings of the production workers were performed for 10 min simultaneously with EMG recordings. Video recordings of the CAD-operators were performed for 5 min at the beginning of a twenty-five min EMG recording.
2.2. Work tasks
2.4. Electromyography
EMG recordings and observations of cycle times and upper arm movements were obtained while the subjects performed their normal work tasks. The production workers produced ends for metal cans and the data presented here were recorded from employees working at two similar machines. At both workstations the workers collected stacks of metal ends from the machine and packed the ends in containers placed next to the machine. At both machines most of the work time was spent collecting (including waiting for enough ends to collect a stack) and visually inspecting the metal ends on the conveyor belt in front of them. A shorter part of the work time was spent lifting the ends from the conveyor belt to the container, which
Bipolar surface electrodes were used for 10 min EMG recordings of the upper trapezius muscle (m. trapezius, pars desc.) on the dominant side of the production workers (n ⫽ 6). For the CAD-operators 25 min EMG recordings of the upper trapezius muscle were performed both on the mouse side (n ⫽ 14) and on the other side (n ⫽ 7). The center of each electrode pair was placed 2 cm lateral to the midpoint between the seventh cervical vertebrae and the lateral end of acromion with an interelectrode distance of 20 mm [10] (Ag-AgCl electrodes, Medicotest A/S, type 7 00 02-E, Denmark). The EMG signal was amplified, low-pass filtered (8.th order Butterworth filter, cut-off 400 Hz) and sampled on a computer (sampling frequency 1024 Hz). The signals were
2. Methods 2.1. Subjects
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
visually checked, high pass filtered (cut-off 10 Hz), fullwave rectified and root-mean-square (RMS) converted within moving windows of 100 ms duration. The resting EMG signal was recorded during a 5 sec period of instructed rest with visual feedback from an oscilloscope to eliminate visible EMG activity and the resting RMS amplitude was quadratically subtracted from all other EMG signals. For normalization EMG activity was recorded during three voluntary isometric contractions with bilateral 90° arm abduction with resistance just proximal to the elbow [16]. The subjects were instructed and encouraged to produce as much force as possible and hold the maximal force for at least 2–3 sec. The maximal EMG amplitude (EMGmax) during the reference contractions was calculated as the highest mean RMS amplitude obtained with a 1 sec window moving in steps of 100 msec. The EMG activity recorded during work was analyzed according to three different procedures: (1) the amplitude probability distribution function (APDF) analysis which quantifies EMG activity levels [11], (2) the EMG gap analysis which quantifies the numbers and duration of ‘silent’ periods in the EMG pattern, defined as periods with an EMG level below 0.5% EMGmax for at least 0.2 sec (gaps) or at least 0.6 sec (long gaps) [21] and (3) the exposure variation analysis (EVA) which quantifies the relative time that the EMG activity is distributed in categories according to the duration of activity sequences within specified amplitude intervals or classes [14]. The level classes used here were 0–0.3% EMGmax, 0.3–1% EMGmax, 1–3% EMGmax, 3–7% EMGmax, 7– 15% EMGmax, > 15% EMGmax and the time classes by which the duration of uninterrupted EMG sequences were categorised were 0–0.3 s, 0.3–1 s, 1–3 s, 3–7 s, 7– 15 s, > 15 s. The distribution of time spent within level classes and within period length classes are presented separately. 2.5. Design and statistics Firstly, associations between muscle activity patterns and work cycle time, arm movement frequency and the duration of pauses in arm movements were studied by: 쐌 comparing muscle activity patterns recorded on the production workers during the performance of task 1 and task 2, as the repetitiveness of the exposure differed for the two tasks, and by 쐌 comparing the muscle activity patterns recorded during work task 2 with those of the CAD-operators, as the cycle times differed considerably, but both tasks were performed with repetitive arm movements. When the paired recordings from task 1 and task 2 were analyzed by the APDF and gap analyses the Wilcoxon Signed Rank test was used to test for differences. Differences in the distributions of time spent within level
319
classes according to the EVA analyses between task 1 and task 2 were tested with a multivariate paired T2test. A similar but separate test was performed on the distributions within period length classes. The total time spent in different classes add up to 100%, but as the test requires statistically independent variables, data from one level class and from one period length class were excluded before the tests were carried out [15]. For the unpaired recordings from task 2 and from CAD-work the Mann-Whitney test was used to test for differences between the results obtained by APDF and gap analyses. Differences in the distributions of time spent within level classes and within period length classes between task 2 and CAD-work were tested with a MANOVA. Also here data from one level class and from one period length class were excluded before the tests were carried out. Secondly, a possible association between muscle activity patterns and musculoskeletal symptoms was studied among the CAD-operators by comparing the muscle activity patterns recorded on the mouse side with those recorded on the non-mouse side. The Mann-Whitney test was used for the results obtained by the gap analyses and a MANOVA was used for the EVA results as described above. A significance level of 0.05 was used.
3. Results The two work tasks of the production workers could be characterized by well-defined work cycles of relatively short duration and involved repetitive movements of the arms (Table 1). However, differences were found between the two work tasks as the work cycles in task 2 were shorter than 30 seconds, and work task 2 was performed with a higher frequency of upper arm movements than work task 1. Thus, work task 2 was characterized both by more repetitive work cycles and more repetitive movements than task 1. For the CAD-operators the duration of the work cycles was highly variable, lasting from minutes to days, and cycle times were not Table 1 Work cycle times and arm movements during two tasks involving manual material handling (metal can production) and during computeraided design (CAD) work based on observations of six workers for each task. Median values (min–max) are shown. Metal can production Task 1 Work cycle time (sec) 53 (41–64) Frequency of upper arm 5 (3–7) movements (min−1) Upper arm pause (% time) 37 (24–55) a
Based on observations of 14 subjects.
CAD-work
Task 2 23 (18–30) 13 (8–18)
12 (9–16)
30 (14–44)
2 (0–43)a
320
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
registered. The median frequency of upper arm movements during CAD-work on the mouse side was similar to that of task 2 for the production workers. Thus, the work tasks of the CAD-operators were characterized by repetitive movements, but not by repetitive work cycles. The median duration of pauses in upper arm movements was similar for the two tasks in metal-can production, but shorter for CAD-work (Table 1). Examples of RMS converted EMG recordings during task 1 and task 2 of one production worker and on the mouse side of one CAD-operator are shown in Fig. 1. For the whole work group the EMG recordings showed similar results for the two work tasks in metal can production both with respect to the median values of the level of EMG activity, the EMG gap pattern and the pattern of repetitiveness (Table 2). The static and median EMG levels were about 1% and 5% EMGmax, respectively, which was also similar to the EMG activity recorded during CAD-work. The tendency to lower peak EMG levels recorded during CAD-work as compared to task 2 in production work was not statistically signifi-
cant. Likewise, no significant differences were found in the frequency or duration of gaps between EMG recordings from CAD-work and work task 2. At the individual level there was no association between the recorded EMG gap time and the observed time that the arm rested on the table or hang passively along the side of the body (Fig. 2). The EVA analyses showed that the EMG level during CAD-work was between 3% and 7% EMGmax for almost half of the work time, whereas this was only the case for about 30% of the work time during production work. The difference in the distribution of time within different level classes between CAD-work and task 2 was almost significant (p ⫽ 0.056). Furthermore, the distribution of time spent in different period length classes was skewed in the direction of more time spent within shorter period lengths during work task 2 compared to CAD-work (p ⫽ 0.012). During CAD-work the median frequency of upper arm movements on the non-mouse side was 2 min−1 (range: 1–4 min−1), i.e. considerably lower than the number of repetitive arm movements on the mouse side (12 min−1,
Fig. 1. Examples of 5 minutes of EMG recordings from one production worker during the performance of task 1 (A) and task 2 (B) shown as the RMS-amplitude. Similarly, an example of an EMG recording from the mouse side of one CAD-operator is shown (C).
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
Table 2 EMG activity patterns of the upper trapezius muscle recorded during metal can production on the dominant side (two work tasks, n ⫽ 6) and during computer-aided design (CAD) work on the mouse side (n ⫽ 14). Median values (min–max) are shown. Metal can production Task 1 APDF-analyses: Static level 0.8 (0.5–2.9) (% EMGmax) Median level 4.2 (1.8–11.7) (% EMGmax) Peak level 15.8 (3.9–20.9) (% EMGmax) Gap analyses: Gaps, > 0.2 s 1.1 (0.6–10.0) (min−1) Long gaps, > 0.1 (0.0–2.6) 0.6 s (min−1) Gap time 1.0 (0.2–8.9) (% time) EVA analyses: Level in % % time EMGmax: 0–0.3 0.3 (0.0–6.5) 0.3–1 10.7 (1.2–25.1) 1–3 18.2 (5.7–54.5) 3–7 29.8 (12.9–40.9) 7–15 21.9 (0.3–50.6) > 15 10.3 (0.0–29.3) Period length in sec: 0–0.3 20.3 (9.9–30.2) 0.3–1 24.0 (15.6–33.4) 1–3 32.2 (27.1–37.4) 3–7 15.0 (6.8–24.6) 7–15 4.2 (1.5–15.0) > 15 1.2 (0.0–8.0) a
CAD-work
Task 2
1.2 (0.6–4.8)
1.4 (0.3–7.3)a
5.3 (2.5–16.1)
4.5 (1.0–11.8)a
14.5 (5.0–33.7)
8.4 (3.1–17.9)a
1.3 (0.0–8.2)
1.6 (0.0–19.0)
0.1 (0.0–1.4)
0.7 (0.0–5.8)
0.6 (0.0–5.8)
2.9 (0.0–18.1)
321
Table 1). The patterns of EMG gaps recorded on the two upper trapezius muscles during CAD-work were also significantly different (Table 3). The frequency of gaps was about ten-fold lower on the side operating the mouse than on the other side and the total gap time was sixfold shorter on the mouse side as compared to the other side. The distribution of time spent within specific level classes according to the EVA analyses showed more time spent at higher level classes on the mouse side (p ⫽ 0.036, Fig. 3). When analyzing the time spent within period length classes it was found that for the recordings on the mouse side a considerable part of the work time was spent within period length classes of 1– 3 s and 3–7 s while for the recordings on the other side the time spent within different classes was more evenly distributed between short, medium and long period lengths (p ⬍ 0.001). 4. Discussion
% time
% time
0.0 8.2 20.2 28.7 22.9 7.6
(0.0–1.0) 0.8 (0.0–6.9) (0.1–15.5) 3.8 (0.0–41.1) (3.4–46.2) 18.7 (0.3–66.2) (5.7–47.5) 46.7 (4.2–79.3) (3.2–46.7) 16.8 (0.2–76.2) (0.0–59.2) 0.9 (0.0–19.3)
21.8 31.0 35.0 11.5 1.6 0.0
(14.0–25.4) (19.9–34.8) (27.3–40.4) (6.9–20.4) (0.0–9.1) (0.0–6.7)
11.3 16.9 30.8 23.8 12.1 2.8
(9.2–19.9) (11.0–28.4) (24.1–38.9) (11.6–35.0) (1.1–19.5) (0.0–13.4)
Previously published in Jensen et al. [8].
4.1. Arm movements and upper trapezius muscle activity CAD-work using a mouse and both of the described work tasks in production work may be characterized as repetitive because a high number of similar arm movements were performed in each task for a major part of the work time (even though the movements during CADwork were not similar to the movements during production work). 4.1.1. Task 1 versus task 2 However, task 1 in production work seemed only half as repetitive as task 2 when comparing arm movement frequencies and work cycle times. The differences in repetitiveness between task 1 and task 2 were not reflected in the EMG patterns of the upper trapezius muscle indicating that the muscular load was the same when performing these tasks, i.e. a larger difference in the number of movements is required to alter the muscle activity pattern. Thus, job rotation between the two work tasks, as was normal procedure in the company, may not Table 3 EMG gaps in the recordings from the upper trapezius muscle on the mouse side (n ⫽ 14) and the non-mouse side (n ⫽ 7) recorded during computer-aided design (CAD) work. Upper trapezius non- Upper trapezius mouse side mouse side Gaps, > 0.2 s (min−1) Long gaps, > 0.6 s (min−1) Gap time (% time)
Fig. 2. Total EMG gap time vs the time that the arm is resting on the table or hanging passively along the side of the body (arm pause) for each worker.
18.0 (2.4–43.1)* 9.3 (1.1–17.0)* 17.3 (6.2–32.8)**
1.6 (0.0–6.1) 0.7 (0.0–2.3) 2.9 (0.0–5.3)
Median values (min–max) are shown. Significant pairwise differences are indicated with * p ⬍ 0.05 or **p ⬍ 0.01.
322
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
Fig. 3. Time spent within level—and period length classes according to EVA analyses of EMG recordings during CAD-work. Median percentages of the time recorded on the upper trapezius muscle on the mouse side (n ⫽ 14) and on the non-mouse side (n ⫽ 7) are shown.
create variation in shoulder muscle loads and consequently may not affect the risk of developing shoulder muscle disorders. The possibility exists that the EMG analysis procedures were not suitable to detect actual differences in the temporal RMS amplitude pattern. However, it seems unlikely that a more detailed analysis than the EVA analysis would be necessary to detect aspects of the EMG pattern that are important for the development of muscular disorders. The main problem seems to reduce the amount of possible information obtained by the EVA analysis and extract only the relevant information. 4.1.2. CAD-work versus task 2 During CAD-work with a similar arm movement frequency to that of task 2 the median level of EMG activity and the number of EMG gaps were also similar. The total gap time was not associated with the observed duration of upper arm pauses neither between tasks nor between individuals performing similar tasks. This is a similar finding to that of a study of light material handling and office work, where no correlation was found between time with an arm posture below 10° elevation and time with an upper trapezius EMG activity below
0.5% EMGmax [20]. The peak levels of EMG activity showed a tendency to be higher during production work. This was probably due to greater arm movements and handling of stacks of metal ends, which required higher arm and shoulder muscle forces during a minor part of the work cycle. However, caution should be exerted in interpreting the actual peak levels of activity during production work as the EMG activity was normalized in relation to a static contraction, but during most of the work time the upper arms were only slightly elevated, making estimates of static and median levels of EMG activity more reliable than the peak levels. This was shown in more detail by the EVA analyses, which also showed that a larger proportion of the work time was spent in shorter period length classes during production work. This indicated that production work was associated with a more repetitive muscle activity pattern than CAD-work, which seemed contradictory to the observed similar degree of repetitiveness of these two types of work with respect to arm movement frequencies. However, the EMG differences may also have been due to different arm postures. CAD-work was performed with a static posture, smaller arm movements and higher precision demands than production work which required a
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
323
static or intermittent-static activation pattern of the upper trapezius muscle with longer stays within the same level classes of the EVA classification system. A comparison of the risk of developing disorders between production and CAD-work based on recordings of exposure and muscle activity patterns would be difficult as both shortcycled repetitive muscle activity patterns as recorded during production work and more static activation patterns as recorded during CAD-work may represent a risk. The common factor, which may lead to an assessment of the risk as being high in both jobs, was the few EMG gaps recorded during all 3 work tasks. This indicated that part of the muscle was continuously active during work. Continuous activity in low-threshold motor units has been suggested to cause musculoskeletal disorders [5], but more scientific evidence is needed to use gap analyses in exposure assessment.
differences in several EMG variables between recordings from the two sides makes it difficult to identify EMG gaps as the single key factor causally related to the development of disorders. The level of EMG activity was also significantly higher on the mouse side (median level: 4.4% EMGmax) than on the other side (median level: 1.9% EMGmax), but it should be realized that the levels were small on both sides. Consequently, the small difference between the two sides appears physiologically insignificant compared to the considerably larger difference in EMG gap time and gap frequency. The EVA analyses showed in more detail that the fraction of the recording with low muscle activity levels, e.g. below 1% EMGmax, was much higher on the non-mouse side than on the other side, clearly emphasizing that muscular rest was seldom present for the upper trapezius muscle on the mouse side during mouse work.
4.2. CAD-work: Side differences in muscle activity patterns and symptom prevalence
4.2.2. Repetitiveness of the muscle activity The distribution of stays within different period length classes was also different between the two sides as the EMG recordings from the non-mouse side showed a more equal distribution between different period length classes, i.e. a more variable EMG pattern than recorded on the mouse side. A priori this seems favorable, however, whether the actual differences influence the risk of developing muscular disorders is not known.
4.2.1. Gaps as risk indicator During CAD-work the number of EMG gaps recorded on the upper trapezius muscle on the non-mouse side was considerably higher and the total gap time was much longer than on the mouse side. In view of the previously reported differences in shoulder symptom prevalence this indicates that lack of EMG gaps may be associated with the development of disorders. Similar to the present study some studies have shown that lack of gaps may be a risk factor for shoulder disorders among workers performing light manual work or office work [6,22], although others have found no difference in EMG gaps between workers with and without shoulder complaints [9,19]. In all of the above-mentioned studies symptomatic and non-symptomatic workers with identical jobs were compared. However, the ability to identify EMG gaps as a risk factor may have been weakened if other factors such as the individual vulnerability to acquire such disorders show large interindividual variation. In the present study the different prevalence of symptoms on the two shoulders were based on reports from a larger group of workers with highly different exposures on the two sides of the same individual. Thus, the mean differences between muscle activity parameters measured on the two shoulders were largely due to exposure contrasts and the large differences in some of the EMG parameters may be important factors influencing the development of pain. When comparing with the study by Veiersted et al. [22] the number of gaps recorded on symptomatic workers (5–10 gaps min−1) in their study was higher than the median number of gaps on the mouse side of the CAD-operators and the number of gaps recorded on nonsymptomatic workers (10–15 gaps min−1) in their study was lower than the median number of gaps recorded on the non-mouse side of the CAD-operators. However, the
5. Conclusions 5.1. Arm movement frequencies Firstly, in repetitive work the exposure should be changed considerably to alter shoulder muscle loads. A decrease in upper arm movement frequency from 13 to 5 movements min−1 during production work had no significant effect on upper trapezius muscle activity patterns such as the number of EMG gaps, whereas a comparison between the mouse- and non-mouse side during CADwork corresponding to arm movement frequencies of 12 and 2 min−1 revealed significant differences in muscle activity patterns. The difference in the number of gaps was especially pronounced indicating that given a sufficiently large change in exposure this parameter may be sensitive to such changes. In contrast no difference in the number of gaps was found between recordings from CAD-work and production work indicating that certain aspects of the exposure of these different work groups were similar. Thus, even though CAD-work and manual material handling differ in many respects and other factors than the movement frequency influence muscle activity, both may be characterized as repetitive work and the data may be taken as support of the guidelines presented by Kilbom [12]. She suggested that an upper arm movement frequency exceeding 2.5 min−1 during
324
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
repetitive work is associated with a high risk of developing musculoskeletal disorders. 5.2. Work cycle times Secondly, a general use of work cycle times to characterize work appears inadequate both to define work as repetitive and to predict muscle loads. A lack of association between work cycle time and forearm muscle load during repetitive meat-cutting work has also been reported by Christensen et al. [3]. Thus, movement frequencies rather than cycle times should be preferred to characterize repetitive work, especially if the aim is to predict the risk of developing disorders. 5.3. EMG analysis methods Thirdly, muscle activity patterns recorded during repetitive work should be analyzed according to other methods than the APDF analysis to obtain relevant information on the muscular load. In principle the EVA method appears suitable to provide all the information which is obtained by the APDF and the gap methods along with unique information regarding the repetitiveness of the muscle activity pattern. However, more efforts should be invested in studying how to interpret the relatively large amounts of information obtained by the EVA procedure. It seems insufficient to focus on the degree of repetitiveness, i.e. the time with stays in short period length classes (0–0.3 sec and 0.3–1 sec) as static muscle activity will show up as time spent in the longest period length classes (7–15 sec and > 15 sec). Both repetitive and static muscle activity should be avoided, whereas a variable muscle activity pattern may have positive health effects. A variable muscle activity pattern may be characterized by an equal partitioning of the time between all period length classes and some partitioning of the time between the level classes, including sufficient time at the lowest level classes to ensure adequate restitution between activity periods. More effort should also be invested in studying whether all the information of the EVA analysis is necessary to predict the risk of developing disorders. If EMG gap patterns are critical for the development of muscular problems it would be wiser to focus directly on these analyses. Future research should resolve these questions.
Acknowledgements We thank Henrik B. Olsen and Paolo Capodaglio for their assistance with data collection and analysis.
References [1] Armstrong TJ, Buckle P, Fine LJ. et al., A conceptual model for work-related neck and upper-limb musculoskeletal disorders. Scand J Work Environ Health 1993;19(2):73–84. [2] Christensen H. Muscle activity and fatigue in the shoulder muscles of assembly-plant employees. Scand J Work Environ Health 1986;12(6):582–7. [3] Christensen H, Søgaard K, Pilegaard M, et al. The importance of the work/rest pattern as a risk factor in repetitive monotonous work (submitted). [4] Hagberg M, Sundelin G. Discomfort and load on the upper trapezius muscle when operating a wordprocessor. Ergonomics 1986;29(12):1637–45. [5] Ha¨gg GM. Static work loads and occupational myalgia - a new explanation model. In: Anderson PA, Hobart DJ, Danoff JV, editors. Electromyographical Kinesiology. Elsevier Science Publishers B.V., 1991:141–144. [6] Ha¨gg GM, Åstro¨m A. Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders. Int Arch Occup Environ Health 1997;69:423–32. [7] Jensen BR, Schibye B, Søgaard K. et al., Shoulder muscle load and muscle fatigue among industrial sewing-machine operators. Eur J Appl Physiol 1993;67:467–75. [8] Jensen C, Borg V, Finsen L. Job demands, muscle activity and musculoskeletal symptoms in relation to computer mouse work. Scand J Work Environ Health 1998;24(5):418–24. [9] Jensen C, Nilsen K, Hansen K. et al., Trapezius muscle load as a risk indicator for occupational shoulder-neck complaints. Int Arch Occup Environ Health 1993;64:415–23. [10] Jensen C, Vasseljen O, Westgaard RH. The influence of electrode position on bipolar surface electromyogram recordings of the upper trapezius muscle. Eur J Appl Physiol 1993;67:266–73. [11] Jonsson B. Measurement and evaluation of local muscular strain in the shoulder during constrained work. J Hum Ergol 1982;11:73–88. [12] Kilbom Å. Repetitive work of the upper extremity: Part I- Guidelines for the practitioner. Int J Ind Erg 1994;14:51–7. [13] Kilbom Å. Repetitive work of the upper extremity: Part II - The scientific basis (knowledge base) for the guide. Int J Ind Erg 1994;14:59–86. [14] Mathiassen SE, Winkel J. Quantifying variation in physical load using exposure-vs-time data. Ergonomics 1991;34(12):1455–68. [15] Mathiassen SE, Winkel J. Physiological comparison of three interventions in light assembly work: reduced work pace, increased break allowance and shortened working days. Int Arch Occup Environ Health 1996;68:94–108. [16] Mathiassen SE, Winkel J, Ha¨gg GM. Normalization of surface EMG amplitude from the upper trapezius muscle in ergonomic studies - A review. J Electromyogr Kinesiol 1995;5(4):197–226. [17] Silverstein BA, Fine LJ, Armstrong TJ. Hand wrist cumulative trauma disorders in industry. Br J Ind Med 1986;43(11):779–84. [18] Takala E, Viikari-Juntura E. Muscular activity in simulated light work among subjects with frequent neck-shoulder pain. Int J Ind Erg 1991;8:157–64. [19] Vasseljen O, Westgaard RH. A case-control study of trapezius muscle activity in office and manual workers with shoulder and neck pain and symptom-free controls. Int Arch Occup Environ Health 1995;67:11–18. [20] Vasseljen O, Westgaard RH. Arm and trunk posture during work in relation to shoulder and neck pain and trapezius activity. Clin Biomech 1997;12:22–31. [21] Veiersted KB, Westgaard RH, Andersen P. Pattern of muscle activity during sterotyped work and its relation to muscle pain. Int Arch Occup Environ Health 1990;62:31–41. [22] Veiersted KB, Westgaard RH, Andersen P. Electromyographic
C. Jensen et al. / Journal of Electromyography and Kinesiology 9 (1999) 317–325
evaluation of muscular work pattern as a predictor of trapezius myalgia. Scand J Work Environ Health 1993;19(4):284–90. [23] Winkel J, Westgaard R. Occupational and individual risk factors for shoulder-neck complaints: Part II—The scientific basis (literature review) for the guide. Int J Ind Erg 1992;10:85–104. Chris Jensen received the M.Sc degree in zoophysiology at the University of Trondheim in Norway. The Ph.D. thesis was received in 1995 after studies at the Norwegian Institute of Technology where he became trained in work physiology. Chris Jensen is a senior researcher since 1998 at the National Institute of Occupational Health, Denmark where he has been engaged in research on monotonous, repetitive work since 1995. Chris Jensen’s research at the University of Trondheim has focused on understanding the physiology of shoulder muscles especially in relation to repetitive work. At the National Institute of Occupational Health in Denmark his research is focused on work exposures and risk factors for the development of musculoskeletal disorders in computer operators. Lotte Finsen received the M.Sc degree in biology and physical training in 1991 and the Ph.D. degree in physiology in 1995, University of Copenhagen. Since 1989 she has been working with basic and applied physiology at the department of Physiology at the National Institute of Occupational Health in Copenhagen. Her research interests focus on the mechanisms behind development of occupational related musculoskeletal disorders which she studies using electromyography, kinesiology and biomechanical models. From 1994–1998 she has been a researcher at the Department for research in repetitive monotonous
325
work, and from 1998 at the Department of Visual Display Unit Work. She serves as a referee for scientific journals. Klaus Hansen is a physiotherapist since 1985. He has worked at Bispebjerg Hospital, Copenhagen from 1985 to 1995, from 1991 to 1994 as department head-physiotherapist. Klaus Hansen has worked at the National Institute of Occupational Health, Denmark since 1995, where he has been engaged in research on monotonous, repetitive work and computer work. His primary focus is on the effect of work factors on the physical capacity of the human body.
Hanne Christensen received the M.Sc degree in physiology and the Ph.D. degree in medicine from Copenhagen University. She has been working in the area of work physiology in the department of Physiology, National Institute of Occupational Health, Copenhagen, since 1986, and as a senior researcher since 1991. From 1994–1998 she has been head of the Department for research in repetitive monotonous work, and 1998–2001 head of the Department of Visual Display Unit Work, both at the Dainsh NIOH. Her research interest is mechanism behind developing of musculoskeletal symptoms and disorders, especially quantitated by electromyography. She serves as a referee for several scientific journals and is a member of the editorial board in Scandinavian Journal of Work, Environment & Health. She is awarded from The Danish Society of casualties from Polio, Road Accidents and Accidents (1986), from The Foundation of Work Protections (1987) and The Danish Society of Manual Medicine (1987).