Comparison of occupational exposure methods relevant to musculoskeletal disorders: Worker–workstation interaction in an office environment

Journal of Electromyography and Kinesiology 22 (2012) 176–185

Contents lists available at SciVerse ScienceDirect

Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin

Comparison of occupational exposure methods relevant to musculoskeletal disorders: Worker–workstation interaction in an office environment Dwayne Van Eerd a,b,⇑, Sheilah Hogg-Johnson a,c, Donald C. Cole a,c, Richard Wells a,d, Anjali Mazumder a,e,f a

Institute for Work & Health, Toronto, Canada School of Public Health and Health Systems, University of Waterloo, Waterloo, Canada c Dalla Lana School of Public Health, University of Toronto, Toronto, Canada d Department of Kinesiology, University of Waterloo, Waterloo, Canada e Department of Statistics, University of Oxford, Oxford, UK f Forensic Science Service, London, UK b

a r t i c l e

i n f o

Article history: Received 11 April 2011 Received in revised form 1 December 2011 Accepted 2 December 2011

Keywords: Exposure assessment Measurement methods Electromyography (EMG) Computer workstations Ergonomics

a b s t r a c t Work related musculoskeletal disorders have been associated with office work yet exposure quantification is challenging and not measured consistently. Our objective was to examine associations within and across exposure measurements guided by a conceptual model of three measurement locations: external to the body, at the interface, and internal to the body. Forty-one office workers (71% female), mean age 41 years (SD = 9.6), mean height 168 cm (SD = 10.3), and mean weight 74 kg (SD = 19), were recruited from a large urban newspaper. Four methods of quantifying mechanical exposure were used linked to locations: equipment dimensions (external), relative fit and postures (interface), and EMG (internal). We explored: (1) a within-location analysis of relationships among methods; and (2) a cross-location analysis of relationships among methods. Exposure method comparisons showed mostly weak correlations among equipment variables, moderate correlations among posture variables, and strong or moderate correlations among EMG variables. For the majority of pair-wise comparisons between exposure measures across locations, the correlations were weak or moderate. Comparisons of relative fit revealed some differences in dimensions, postures, and EMG measures. Few strong associations between various exposure measures were found, although worker-reported relative fit holds promise. Future work might link exposure methods (specific measures) with locations for particular purposes. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction The prevention of work-related musculoskeletal disorders (MSDs) among office workers relies on accurate identification of exposure to occupational hazards. Mechanical exposures are multifactoral and related to workplace design and individual working technique, including posture and muscular loads, have been linked to MSDs (Lindegard et al., 2003; Gerr et al., 2000; Karlqvist et al., 1996). Job types and tasks involving repetitive or monotonous exertions as well as task duration have also been linked to MSDs (Marcus et al., 2002; Homan and Armstrong, 2003; Polanyi et al., 1997; Nordander et al., 2000). MSD symptoms have also been related to long hours of mouse work (Andersen et al., 2008; Gerr et al., 2004) and mouse location (Jensen, 2003; Karlqvist et al., ⇑ Corresponding author at: Institute for Work & Health, 481 University Avenue, Suite 800, Toronto, Ontario, Canada M5G 2E9. Fax: +1 416 927 4167. E-mail address: [email protected] (D. Van Eerd). 1050-6411/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2011.12.001

1998; IJmker et al., 2007) as well as static postures and muscle activation (Jensen et al., 1998). The lack of well-defined, reliable, and common metrics for mechanical exposure assessment remains a methodological challenge (Punnett and Wegman, 2004; Spielholz et al., 2001; Gerr et al., 1996; Hagberg, 1992; Takala et al., 2010) and limits inference of causal relationships. Mechanical exposures may be measured using self-report methods, observational techniques, and direct measurements. Each method quantifies aspects of exposure (van der Beek and Frings-Dresen, 1998; Westgaard and Winkel, 1997) with varying degrees of measurement error (Wells et al., 1997; van der Beek and Frings-Dresen, 1998; Burdorf, 1995). In a recent literature review (Takala et al., 2010), 30 different methods for systematic observation of a person at work were identified yet no gold standard emerged. Laboratory studies have shown changes in workstation configuration and dimensions affect the magnitude of the electromyogram (Straker et al., 2008; Delisle et al., 2006). The relationships are

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

usually expressed as changes over a small number of experimental conditions. When independent variables such as keyboard angle had more than two positions, relationships for postures often demonstrated linear relationships however EMG showed a mixture of relationships, linear, flat, concave-up or down non-linear relationships (Kotani et al., 2007; Simoneau et al., 2003). This could lead to low correlations between workplace features, postures, and EMG. Comparisons of exposure methods have typically not shown strong relationships in workplace settings (Homan and Armstrong, 2003; Spielholz et al., 2001). Few office workstation dimensions have been associated with postures in workplace settings (Gerr et al., 2000, 2004). Muscular load (based on surface electromyography (EMG)) and either workstation attributes or postural measures (observed or self-report) also tend to show weak or inconsistent relationships in the workplace (Lindegard et al., 2003; Jensen et al., 1998; Hansson et al., 2000; Karlqvist et al., 1999, 1998). In Takala and colleagues’ review (2010), direct observation showed reasonable agreement with video based observation but agreement was lower between observational methods and technical measures. A better understanding of the relationships between various exposure measures is necessary to establish better measures and ultimately a better understanding of workplace exposures. This paper reports on part of a collaborative research project with a large urban newspaper (Cole et al., 2003). One aspect of this project was an exposure assessment incorporating various methods (self-report and direct measurement) that was guided by a conceptual exposure model (Wells et al., 2004). The model conceptualizes three locations of exposure measurement: external to the body, an interface, and internal to the body (Wells et al., 2004). Externally we can measure aspects such as equipment used and work organization. At the interface we can measure the interaction of the person with the environment over time, which may be characterized by the relative fit between the body and equipment used, postures adopted, and number and frequency of tasks performed (see Van Eerd et al., 2009, for more on the latter). Internally we can observe electromyographical signals as indicators of mechanical exposure (see Fig. 1). This paper’s primary objective was to examine associations among exposure measurement methods within and across potential exposure locations during the performance of common office

177

tasks. Given the complexity of relationships and the uncertainties associated with specific within- and cross-location relationships in our conceptual model, it is not a model validation nor an hypothesis testing study but rather an exploratory study elucidating the multifactoral nature of exposures in office work. 2. Methods 2.1. Subjects Forty-one office workers (71% female), mean age of 41 years (SD = 9.6), were recruited from a large urban newspaper and provided informed written consent as per approval of the Research Ethics Boards of McMaster University and University of Waterloo. They performed clerical, administrative, sales, customer accounts, and call center jobs in the advertising, circulation and finance departments. Their mean height and weight were 168 cm (SD = 10.3) and 74 kg (SD = 19), respectively. All workers reported using the mouse with their right hand. 2.2. Exposure assessment methods We compared measures from four mechanical exposure methods (see Table 1). Equipment and posture measurements were taken by trained observers while workers performed tasks and EMG was collected concurrently with these observations over a 4 month period in a single calendar year. 2.2.1. External Trained observers took equipment measurements at the worker’s usual workstation with the worker present (see Appendix Table A1). Dimensions were measured using a tape measure (Ortiz et al., 1997). The orientation of workstation equipment was determined relative to the J key (center of the keyboard) or the floor, using a calibrated ‘‘bubble’’ level as a reference (true horizontal) when needed (Gerr et al., 2002; Ortiz et al., 1997). 2.2.2. Interface Relative fit. Workers completed a questionnaire which included demographic characteristics and workstation setup. For the latter,

Fig. 1. A model of the relationships between exposures in an office environment. Darker text shows the parts of the model examined in this paper.

178

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

Table 1 Measures taken for each location-exposure method related to the equipment used for each task.

a b

Interface

Internala – EMG (10th, 50th, 90th APDF and gap)

Equipment

External – dimensions (cm)

Questionnaire

Postures (°)

Keyboard

Keyboard height Keyboard–table edge distance Keyboard–seat distance Keyboard–monitor distance

Worker- reported keyboard position (relative fit)

Inner elbow (left, right) Shoulder flexion (left, right) Shoulder abduction (left, right) Wrist extension (left, right) Wrist ulnar deviation (left, right)

ECRB (left, right) Trapezius (left, right)

Mouse

Mouse height Mouse–table edge distance Mouse–seat distance Mouse–keyboard distance

Worker-reported mouse position (relative fit)

Not available

ECRB (right) Trapezius (right)

Phone

Not available

Hands-freeb vs. hand-held phone

Not available

ECRB (left, right) Trapezius (left, right)

APDF, amplitude probability distribution function; ECRB, extensor carpi radialis brevis. A hands-free phone was considered an optimal phone position while a hand-held phone was considered a non-optimal phone position.

workers used diagrams, developed in conjunction with the newspaper’s graphics staff (see Appendix Fig. A1), to indicate equipment location relative to a box placed directly in front of a seated person at elbow height. Participants were asked to indicate whether their keyboard and mouse were inside or outside the boxes on the diagram (Polanyi et al., 1997). If they responded ‘‘outside’’, they further indicated higher/lower than and/or to the left/right of the boxes. For phone, workers indicated whether they used a handsfree phone, a regular hand-held phone, a cell phone or a phone with shoulder rest (Cole et al., 2006). Postures. Trained observers assessed bilateral upper-extremity postures (shoulder, elbow, and wrist angles) while the workers performed their usual keying tasks (see Appendix Table A2). Hand-held goniometers were used, based on standard anatomical landmarks and body segments (Gerr et al., 2002; Ortiz et al., 1997). A calibrated ‘‘bubble’’ level provided a reference (true horizontal) to calculate angles as necessary. 2.2.3. Internal A portable EMG signal analysis system (ME300P8, Mega Electronics Ltd., Finland, CMRR 110 dB, 15–500 Hz, input impedance 10 GX) was used to record the workers’ muscle activity over 2 h on two separate days while workers performed their usual office tasks at their workstation. The skin surface was prepared by shaving the area – where necessary – and scrubbing with a water–alcohol solution. Standardized silver/silver-chloride disposable electrodes (Mediocost Blue Sensor N-00-S) were placed 2 cm apart. Surface EMG was collected at 1000 Hz and root mean square (RMS) EMG signals were calculated for each 100 ms window (10 Hz) from the extensor carpi radialis brevi (ECRB) and the middle trapezius sites bilaterally. For the ECRB site, the surface electrodes were placed one-third of the distance from the lateral epicondyle to the radial styloid. For the trapezius site, the surface electrodes were placed midway between C7 and the acromion. Signal quality and quiet level tests were performed in a dedicated office away from the workstation. The EMG measures were calibrated to maximum voluntary contraction (MVC). The MVC for the ECRB was determined by workers performing a maximal grasp simultaneously while extending their wrist. The MVC for the trapezius was achieved by workers pulling maximally against straps that were fixed to the floor and looped over their elbows while they stood with their arms straight and both shoulders abducted 90°. A VHS video camera (no audio), with a wide angle lens, captured the workers performing their usual tasks at their workstation, computer, office equipment, and much of their desk space. EMG collection started in view of the video camera recording to

permit synchronization during a 30 min period. Synchronization between video and EMG was accomplished by recording distinct periodic electronic signals or ‘‘marks’’ at the start, change, and end of tasks, using a switch on the EMG recording system, capturing the deployment of the switch on video (Moore et al., 2003), and producing a time stamped text file. Video was not available for all tasks on all workers, and use of a headset made interpretation of phone use difficult on video without audio. Hence, the number of workers varied for the three office tasks (keying n = 30, mousing n = 24, and phone n = 17).

2.3. EMG extraction and concatenation procedure Thirty minute segments of each synchronized video and EMG recording were analyzed by trained video coders using Observer Pro 4.0 (Noldus Information Technology, The Netherlands). The 30 min period was chosen by the video coder to maximize the time that a subject was present at their workstation. Task start and stop times were recorded to within 0.1 s for three office equipment-related tasks: (1) keying – hands touching or directly over keyboard; (2) mousing – hand touching or directly over mouse; and (3) phone – hand grasping handset. Task specific EMG data were extracted from the synchronized video and EMG recordings for each worker (Moore et al., 2003). The task specific EMG data were then concatenated across all portions of the task specific EMG e.g. mousing (see Fig. 2). The remainder of the EMG signal was then concatenated as non-task specific EMG sample e.g. not using a mouse. The amplitude probability distribution function (APDF) at the 10th (static), 50th (dynamic), and 90th (peak) percentile (Jonsson et al., 1988) and gaptime (<0.5% MVC value for at least 200 ms) in s/min (Veiersted et al., 1993) were calculated for each of the task specific EMG signals using custom software.

2.4. Statistical analyses Given our exposure model (Fig. 1), we explored relationships between various exposure methods within each location as well as relationships across locations (see Table 1). Given the number of comparisons and the exploratory nature of this work, we did not focus on statistically significant associations alone. We describe the nature of relationships between these exposure variables. All statistical analyses were performed using SAS V9 (SAS, 1990). S-plus was used to illustrate graphical results (Venables and Ripley, 1999).

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

179

Fig. 2. Schematic representation of the EMG extraction and concatenation procedure during mousing for the extensor carpi radialis muscle (ECRB) site. Timing of the hand touching or directly over mouse (on) and hand away from the mouse (off) were obtained from video recordings synchronized to the EMG.

2.4.1. Within-location analysis Summary statistics on all measures were computed for each exposure method within a location. For categorical questionnaire data, frequency distributions were calculated. We calculated ‘‘relative external measures’’ by dividing the various external measures by the subject’s height. However these ‘‘relative external measures’’ did not result in different patterns of associations and are not as readily interpretable so we do not report them in this paper. For all direct workstation and posture measurements, the data were normally distributed, permitting use of means and standard deviations (SD). Task-specific Pearson correlation coefficients were calculated to determine the linear strength between pair-wise measures within each exposure method. We considered 0.5 and greater as strong, 0.3–0.5 as moderate, and 0.1–0.3 as weak (both positive and negative). If more than one method was used within a location, then pair-wise comparisons were conducted in the same manner as cross-location comparisons (see Section 2.4.2). 2.4.2. Cross-location comparisons Scatterplots were created to illustrate the relationship between pairs of direct continuous measures (e.g. dimension measures vs. posture measures), followed by calculation of Pearson correlation coefficients of pairs of exposure measures (by task). T-tests were used to test for differences in the continuous direct exposure measures by worker reported relative fit (significance set at p 6 0.05), with illustration of any relationships via box plots. 3. Results Thirty-three workers had data for all exposure measures except for mouse specific (n = 24) and phone specific (n = 17) tasks.

single strong and statistically significant correlation between mouse height and mouse–seat distance (r = 0.53) was found, with weak correlations among other pairs (range: ±0.05 to 0.15). 3.1.2. Interface Relative fit. Eleven of the 33 workers reported that their keyboard was inside the box, 14 that their mouse was inside the box (see Appendix Fig. A1), and eight that both their keyboard and mouse were outside the box. Those reporting a keyboard location outside the box tended to indicate that it was too high while the mouse was either too high or too far to the side. Sixteen workers reported using hand-held phones. Workers reporting equipment outside of the box tended to spend less time (not statistically significant) performing the three equipment-related tasks during a normal workday. Postures. Among bilateral keyboard-related posture measures, strong and statistically significant correlations were observed between (see Table 2): left and right inner elbow angles (r = 0.64), and left and right shoulder angles (abduction (r = 0.58) and flexion (r = 0.52)). Among unilateral measures, there was a moderately strong and statistically significant negative correlation found between left shoulder abduction and wrist extension (r = 0.47) and a moderately strong and significant positive correlation for left shoulder and wrist ulnar deviation (r = 0.43) measures. Correlations of this strength were not seen on the right side. 3.1.3. External – interface Of the eight dimensional measures, mouse–table edge distance was significantly different for inside and outside the box, with a similar, but non-significant pattern, for keyboard–monitor distance (see Table 3). Of the 10 posture measures, the left shoulder flexion angle showed significant differences: 32.6° outside box vs. 18.4° for inside box reported keyboard locations.

3.1. Within-location associations 3.1.1. External Among the keyboard-related dimension measures, one pair showed a strong and statistically significant correlation, keyboard height and keyboard–seat distance (r = 0.82), another pair a moderate correlation, keyboard height and keyboard–monitor distance (r = 0.31), and four other pairs, weak correlations (ranging between ±0.05 and 0.27). Among mouse-related dimension measures, a

3.1.4. Internal Bilateral EMG measures (e.g. between left and right ERCB) showed moderate to strong and mostly significant correlations for keying, ranging between: (0.37, 0.82) for ECRB, and (0.18, 0.72) for trapezius, gaptime being the highest (in both muscle groups). The only other significant bilateral relationship was observed in the static measure of the trapezius muscles during phone use (r = 0.52). EMG measures for unilateral muscles (e.g. left ECRB

180

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

Table 2 Correlation matrix for postures measures (interface) while performing keying tasks (n = 33). Left

Left Inner elbow Shoulder abduction Shoulder flexion Wrist extension Wrist ulnar deviation Right Inner elbow Shoulder abduction Shoulder flexion Wrist extension Wrist ulnar deviation

Right

Inner elbow

Shoulder abduction

Shoulder flexion

Wrist extension

Wrist uln dev

1.00 0.20

1.00

0.54* 0.10 0.16

Inner elbow

Shoulder abduction

0.25 0.47* 0.43*

1.00 0.18 0.21

1.00 0.25

1.00

0.64* 0.07

0.11 0.58*

0.60* 0.11

0.35* 0.43*

0.02 0.30

1.00 0.23

1.00

0.41* 0.10 0.16

0.16 0.10 0.29

0.52* 0.16 0.02

0.19 0.26 0.01

0.16 0.22 0.26

0.61* 0.31 0.05

0.07 0.19 0.02

Shoulder flexion

Wrist extension

Wrist uln dev

1.00 0.14 0.05

1.00 0.03

1.00

Uln dev, ulnar deviation. p 6 0.05.

*

Table 3 Dimension measures (external) and posture measures (interface) by relative fit of equipment (interface) (n = 33). Relative fit of equipmenta Inside box (IN)

IN–OUT difference

p-Value

Outside box (OUT)

Mean (SD)

Range

Mean (SD)

Range

Dimension measure (cm) Keyboard height Keyboard–table edge distance Keyboard–seat distance Keyboard–monitor distance Mouse height Mouse–table edge distance Mouse–seat distance Mouse–keyboard distance

72.93 (4.12) 5.18 (5.39) 22.33 (5.00) 4.02 (5.19) 69.50 (3.20) 13.00 (2.35) 18.68 (3.84) 36.76 (16.40)

66.50–87.00 11.00 to 14.00 14.50–40.00 7.50 to 12.80 60.50–73.00 10.00–19.00 8.00–25.00 25.00 to 47.00

74.55 (5.55) 4.82 (7.11) 23.36 (5.73) 6.53 (4.63) 70.41 (2.36) 15.71 (4.85) 19.48 (2.89) 40.96 (5.56)

68.50–89.50 14.00 to 14.00 16.00–37.50 1.00 to 13.50 67.00–76.00 11.00–29.00 16.00–25.00 34.00–56.00

Posture measure (°) Left elbow angle Right elbow angle Left shoulder abduction Right shoulder abduction Left shoulder flexion Right shoulder flexion Left wrist extension Right wrist extension Left wrist ulnar deviation Right wrist ulnar deviation

108.59 (13.25) 109.86 (15.62) 15.62 (9.71) 21.41 (12.76) 18.41 (10.25) 18.96 (10.35) 14.05 (9.71) 16.73 (9.48) 10.41 (7.90) 7.59 (4.18)

90–135 85–154 5–38 5–50 0–35 0–35 0–36 0–40 0–35 0–20

116.45 (13.61) 114.36 (15.36) 14.82 (5.76) 15.73 (8.50) 32.55 (12.92) 21.46 (21.38) 19.64 (13.52) 20.18 (22.56) 8.18 (2.36) 9.82 (8.73)

92–137 101–158 5–23 0–30 15–60 27 to 56 4–40 0–72 3–10 0–30

1.62 0.36 1.03 2.51 0.91 2.71 0.80 4.2

0.35 0.87 0.60 0.19 0.39 0.05* 0.52 0.37

7.86 4.5 0.86 5.68 14.14 2.5 5.59 3.46 2.23 2.23

0.12 0.44 0.79 0.19 0.00* 0.65 0.18 0.54 0.37 0.33

a *

See Appendix Fig. A1. p 6 0.05.

and left trapezius) across all three tasks exhibited weak correlations except for: gaptime for the left muscles (r = 0.40) and right muscles (r = 0.36) during keying; and dynamic (r = 0.52) and peak (r = 0.68) for left muscles during phone use (see Table 4). Unilateral EMG measures showed weaker correlations for keying with moderate and significant correlations for gaptime on right and left sides (0.36 and 0.40). The correlations for mousing were consistently low. There were however significant stronger correlations for phone use on the left side for dynamic and peak (0.52 and 0.60, respectively) (see Table 4).

pair-wise keyboard-related dimension and EMG measures, the correlations ranged from 0 to 0.39, and from 0 to 0.47 for the 32 mouse pairings. Of the 160 pair-wise keyboard-related posture and EMG measures, the correlations ranged from 0 to 0.53 (see Table 5).

3.2. Cross-location comparisons

3.2.2. External and internal Seven keyboard-related dimensions and EMG measures had correlations between 0.35 and 0.39, the largest between keyboard height and the left trapezius gaptime. Among the fewer mousingspecific comparisons (right hand only), the right ECRB gaptime and mouse–keyboard distance measures had a correlation of 0.47.

For the large number of pair-wise comparisons, scatterplots did not suggest any non-linear relationships but there were few strong correlations. For the 70 keyboard-related dimension and posture pairings, correlation (r) values ranged from 0 to 0.51. For the 64

3.2.1. External and interface The largest correlation was between keyboard–table edge distance and right wrist ulnar deviation (r = 0.51), with less for right shoulder abduction (r = 0.35).

181

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185 Table 4 Bilateral and unilateral correlations of EMG measures (internal) for each office task. Equipment

Muscles

Keying (n = 30)

Bilateral Left and right ECRB Left and right trapezius Unilateral Left ECRB and trapezius Right ECRB and trapezius

EMG measure Static

*

Mousing (n = 24)

Unilateral Right ECRB and trapezius

Phone use (n = 17)

Bilateral Left and right ECRB Left and right trapezius Unilateral Left ECRB and trapezius Right ECRB and trapezius

Dynamic

Peak

‘‘Gaptime’’

0.37* 0.40*

0.43* 0.33*

0.52* 0.18

0.82* 0.72*

0.08 0.07

0.10 0.17

0.26 0.08

0.40* 0.36*

0.21

0.11

0.11

0.23

0.08 0.52*

0.28 0.09

0.16 0.33

0.12 0.39

0.20 0.01

0.52* 0.12

0.68* 0.15

0.27 0.33

p 6 0.05.

Table 5 Summary of moderate or greater correlations between pairs of exposures measures. Equipment

Exposure method

r

Dimensions (external)

Postures (interface)

Keyboard

Keyboard–table edge distance Keyboard–table edge distance

R Sh Abduction R Wrist uln dev

Keyboard

Keyboard height Keyboard height Keyboard height Keyboard–seat distance Keyboard–monitor distance Keyboard–monitor distance Keyboard–monitor distance

L Trapezius (50th) R ECRB (90th) L Trapezius (gap) L ECRB (10th) R ECRB (90th) R Trapezius (50th) L Trapezius (50th)

0.36 0.37 0.39 0.35 0.36 0.25 0.36

Mouse

Mouse–keyboard distance Mouse–keyboard distance

R ECRB (gap) R Trapezius (gap)

0.47 0.40

R Trapezius (10th) R Trapezius (50th) R ECRB (50th) R ECRB (10th) R ECRB (50th) L ECRB (10th) L ECRB (50th) L ECRB (gap) R Trapezius (50th) R Trapezius (90th) L Trapezius (10th) L Trapezius (50th)

0.44 0.40 0.33 0.49 0.37 0.45 0.39 0.53 0.42 0.40 0.42 0.37

Keyboard

R Sh Abduction R Sh Abduction R Sh Abduction L Sh flexion L Sh flexion L Wrist ext L Wrist ext L Wrist ext R Wrist ext R Wrist ext R Wrist uln dev R Wrist uln dev

EMG by muscle (internal) (APDF%, gap) 0.35 0.51

Note. R, right; L, left; Sh, shoulder; ext, extension; uln dev, ulnar deviation; APDF, amplitude probability distribution function; ECRB, extensor carpi radialis brevis.

3.2.3. Interface and internal Postures. Among the 10 pair-wise analyses conducted between posture and EMG measures (r of 0.33–0.53), most involved static and dynamic EMG measures on the same side. For instance, the strongest correlation was found between the keying-specific EMG gaptime measure for left ECRB and left wrist extension angle (r = 0.53).

4. Discussion

3.2.4. Interface and internal Relative fit. For the mouse, those indicating it as outside the box had significantly higher muscle activity/loadings for the right ECRB (APDF measures) than those indicating it as inside (Fig. 3, left). For the keyboard, workers reporting keyboards inside the box demonstrated longer gaptimes across all muscles (Fig. 3, right) than those reporting keyboards outside the box. Those using a hands-free phone vs. those using a hand-held phone showed differences in dynamic (1.62 vs. 5.62% MVC) and peak (7.70 vs. 14.67% MVC) measures for the right trapezius muscle (data not shown).

4.1. Within-location comparisons

In this paper we examined the possible association between a variety of exposure measures in an office-based workplace. Our exploration of relationships between commonly used exposure measurement methods was categorized by location relative to the human body. We uncovered a number of patterns of co-variation and a few important differences.

Equipment dimension measures generally showed weak associations among each other. The stronger association found between equipment (e.g. mouse) location and its relative distance from the floor compared with the worker’s chair-seat relative to the floor may be because it accounts for workers’ stature, better reflecting the worker–workstation interface. Among the posture measures, bilateral shoulder measures showed stronger correlations than unilateral measures. Posture

182

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

Fig. 3. Left hand panel: distribution of 10th, 50th, and 90th percentile EMG levels from the amplitude probability distribution function (APDF) from the right extensor carpi radilis brevis (ECRB) site for mouse reported inside box (IN) vs. mouse reported outside box (OUT) while mousing. Right hand panel: distribution of gaptime across all muscles for keyboard reported inside box (IN) vs. keyboard reported outside box (OUT) while keying. (Note. L, left; R, right.) Note boxplots show the median value (white line), the top and bottom of the shaded box shows the inter-quartile range and the horizontal lines above and below the box (and connected by the vertical lines) show the minimum and maximum values. Horizontal lines above (or below) the max (or min) indicate outliers.

measures reflect the workers’ relationship to their office equipment (Marcus et al., 2002; Gerr et al., 2000) and may account for consistent individual characteristics hence exhibit stronger associations amongst themselves. Hansson et al. (2009, 2010), in a very wide range of jobs, also showed strong relationships between postures on the right and left for the shoulder and to a somewhat lesser extent, the wrist. Similar to Hansson et al. (2009, 2010) we found strong correlations between the magnitudes of the APDF values within a muscle and negative relationships between APDF and gaptime measures (data not shown). Self-report measures of relative fit between worker and equipment were reflected in small differences in average posture measures. For example, the mean difference in left elbow angle between keyboard inside and outside the box was nearly 8°, indicating that workers reporting the keyboard as outside the box had larger inner elbow angles than those reporting the keyboard inside the box. However, standard deviations were larger than any differences, perhaps due to variation in individual perceptions of relative fit. 4.2. Cross-location comparisons Correlations between equipment dimensions, postures, and EMG were relatively weak (highest being 0.53). Consistent with previous studies, only two of 70 pairings of dimensions and postures (keyboard table-edge with wrist ulnar deviation and shoulder abduction) were significant, suggesting that at a group level, equipment dimensions are not strong determinants of worker postures (Gerr et al., 2000). Only seven out of 96 dimension and EMG comparisons and 10 out of 160 posture and EMG comparisons were found to be at least moderate, suggesting that individual dimension/posture measures are not strong determinants of muscle activity. Nevertheless, keyboard height and keyboard–monitor distance were more predictive of static and dynamic trapezius measures than other dimension measures. Among postures, shoulder abduction and wrist extension and ulnar deviation angles were associated with static and dynamic trapezius measures while shoulder flexion and wrist extension were associated with static and dynamic ECRB measures. Dimension and posture measures were found to be associated with static and dynamic levels of muscle activity as well as gaptime. The relative fit of the keyboard and mouse – being in or out of the box – were however associated with measures of muscle

activation. Longer gaptimes (more rest) were found in workers reporting their keyboard as inside the box. Similarly, right ECRB activity was higher in workers reporting their mouse as outside the box, indicating increased muscle loadings. Revelation of such differences in field conditions strengthens similar laboratory findings in VDT work (Karlqvist et al., 1999, 1998; Moffet et al., 2002; Straker et al., 2008). Overall the findings of this study suggest that measuring mechanical exposure in office workers requires a variety of methods addressing different body locations. With relatively few strong correlations among exposure measures, researchers and practitioners should be aware that different measures may be assessing different aspects of mechanical exposure. To best capture exposure among office workers likely requires a mix of methods and variety of measures. 4.3. Limitations The major limitations of this study are the differences in collection time and exposure which may have contributed to attenuation of some relationships between measures within and across exposure methods. A more targeted sampling strategy with more concurrent measures for all methods (Loomis and Kromhout, 2004; Mathiassen et al., 2002) may have increased co-variation or revealed more distinct relationships. We were also limited in the number of office tasks we could compare given that we could not interfere with participants’ work. We note that the relationships noted here may be related to the fairly constrained postures typical in office work and may not generalize to other, non-office, work environments. 5. Conclusion Our comparison of common methods representing external, interface, and internal exposures found few strong associations between exposure measures either within or across exposure locations (Fig. 1). A combination of multiple exposure methods is therefore likely required to permit a more accurate estimate of risk associated with the multi-dimensional domain of mechanical exposure. Determining the mix of methods (and specific measures) which best capture mechanical exposure, particularly in an office environment, would be helpful in future studies of exposure.

183

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

Acknowledgements

References

Study funding: US NIOSH/NIH and the Center for VDT & Health. Sponsorship: the Institute for Work & Health, an independent notfor-profit research organization which receives support from the Ontario Workplace Safety & Insurance Board. Study participants, other members of the Worksite Upper Extremity Group and the RSI (Repetitive Strain Injury) Watch Steering Committee. Anne Moore and Cynthia Chen made contributions to analysis and previous manuscript drafts.

Andersen JH, Harhoff M, Grimstrup S, Vilstrup I, Lassen CF, Brandt LPA, et al. Computer mouse use predicts acute pain but not prolonged or chronic pain in the neck and shoulder. Occup Environ Med 2008;65(2):126–31. Burdorf A. Reducing random measurement error in assessing postural load on the back in epidemiologic surveys. Scand J Work Environ Health 1995;21(1):15–23. Cole DC, Ferrier S, Hogg-Johnson S, Kennedy C, Lee H, Van Eerd D, et al. WMSD: evaluating interventions among office workers. Toronto: Institute for Work & Health; Report No. National Institute for Occupational Safety and Health (NIOSH) Research Grant No. 5 R01 OH003708-03; 2003. Cole DC, Hogg-Johnson S, Manno M, Ibrahim S, Wells RP, Ferrier SE, et al. Reducing musculoskeletal burden through ergonomic program implementation in a large newspaper. Int Arch Occup Environ Health 2006;80(2):98–108. Delisle A, Larivière C, Plamondon A, Imbeau D. Comparison of three computer office workstations offering forearm support: impact on upper limb posture and muscle activation. Ergonomics 2006;49(2):139–60. Gerr F, Marcus M, Ensor C, Kleinbaum D, Cohen S, Edwards A, et al. A prospective study of computer users. I: Study design and incidence of musculoskeletal symptoms and disorders. Am J Ind Med 2002;41:221–35. Gerr F, Marcus M, Monteilh C. Epidemiology of musculoskeletal disorders among computer users: lessons learned from the role of posture and keyboard use. J Electromyogr Kinesiol 2004;14:25–31. Gerr F, Marcus M, Ortiz DJ. Methodological limitations in the study of video display terminal use and upper extremity musculoskeletal disorders. Am J Ind Med 1996;29(6):649–56. Gerr F, Marcus M, Ortiz D, White B, Jones W, Cohen S, et al. Computer users’ postures and associations with workstation characteristics. AIHA J 2000;61(2):223–30. Hagberg M. Exposure variables in ergonomic epidemiology. Am J Ind Med 1992;21(1):91–100. Hansson G-Å, Balogh I, Ohlsson K, Granqvist L, Nordander C, Arvidsson I, et al. Physical workload in various types of work. Part I: Wrist and forearm. Int J Ind Ergon 2009;39:221–33. Hansson G-Å, Balogh I, Ohlsson K, Granqvist L, Nordander C, Arvidsson I, et al. Physical workload in various types of work. Part II: Neck, shoulder and upper arm. Int J Ind Ergon 2010;40:267–81.

Appendix A

Fig. A1. Graphic on questionnaire for relative fit (inside or outside box) for keyboard and mouse locations.

Table A1 Description of workstation dimension measures (external) taken. Workstation dimension measure

Description

Seat height Keyboard height Keyboard–table edge distance Keyboard–seat distance Keyboard–monitor distance Mouse height Mouse–table edge distance Mouse–seat distance Mouse–keyboard distance

Floor to height of top surface of seat pan Floor to height of the ‘‘J’’ key Distance from ‘‘J’’ key to the table edge Distance between keyboard height and seat height Distance from ‘‘J’’ key to center screen Floor to height of the mouse pad Distance from center pad to the table edge Distance between mouse height and seat height Horizontal distance from ‘‘J’’ key to mouse pad center

Table A2 Description of posture measures (interface). Worker posture measure

Description

Inner elbow

Viewed from the side: angle between the upper arm and the forearm

Shoulder abduction

Viewed from behind: angle between the upper arm and the torso

Shoulder flexion

Viewed from the side: angle between the upper arm and the torso

Wrist extension

Viewed from the side: angle between the dorsal surface of the hand and a line extending from the ulna

Wrist ulnar deviation

Viewed from above: angle created from a line in the center of the dorsal hand and a line extending from the center of the dorsal forearm (between the ulna and the radius)

Visual depiction

184

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185

Hansson G-Å, Balogh I, Ohlsson K, Pålsson B, Rylander L, Skerfving S. Impact of physical exposure on neck and upper limb disorders in female workers. Appl Ergon 2000;31:301–10. Homan MM, Armstrong TJ. Evaluation of three methodologies for assessing work activity during computer use. AIHA J 2003;64:48–55. IJmker S, Huysmans MA, Blatter BM, van der Beek AJ, van Mechelen W, Bongers PM. Should office workers spend fewer hours at their computer? A systematic review of the literature. Occup Environ Med 2007;64:211–22. Jensen C. Development of neck and hand–wrist symptoms in relation to duration of computer use at work. Scand J Work Environ Health 2003;29(3): 197–205. Jensen C, Borg V, Finsen L, Hansen K, Juul-Kristensen B, Christensen H. Job demands, muscle activity and musculoskeletal symptoms in relation to work and computer mouse. Scand J Work Environ Health 1998;24(5):418–24. Jonsson BG, Persson J, Kilbom Å. Disorders of the cervicobrachial region among female workers in the electronics industry: a two-year follow up. Int J Ind Ergon 1988;3:1–12. Karlqvist L, Bernmark E, Ekenvall L, Hagberg M, Isaksson A, Rostö T. Computer mouse position as a determinant of posture, muscular load and perceived exertion. Scand J Work Environ Health 1998;24(1):62–73. Karlqvist L, Bernmark E, Ekenvall L, Hagberg M, Isaksson A, Rostö T. Computer mouse and track-ball operation: similarities and differences in posture, muscular load and perceived exertion. Int J Ind Ergon 1999;23:157–69. Karlqvist L, Hagberg M, Köster M, Wenemark M, Nell R. Musculoskeletal symptoms among computer-assisted design operators and evaluation of a self-assessment questionnaire. Int J Occup Environ Health 1996;2:185–94. Kotani K, Barrero LH, Lee DL, Dennerlein JT. Effect of horizontal position of the computer keyboard on upper extremity posture and muscular load during computer work. Ergonomics 2007;50(9):1419–32. Lindegard A, Wahlström J, Hagberg M, Hansson G-Å, Jonsson P, Wigaeus Törnqvist E. The impact of working technique on physical loads – an exposure profile among newspaper editors. Ergonomics 2003;46(6):598–615. Loomis D, Kromhout H. Exposure variability: concepts and applications in occupational epidemiology. Am J Ind Med 2004;45(1):113–22. Marcus M, Gerr F, Monteilh C, Ortiz DJ, Gentry E, Cohen S, et al. A prospective study of computer users. II: Postural risk factors for musculoskeletal symptoms and disorders. Am J Ind Med 2002;41:236–49. Mathiassen SE, Burdorf A, van der Beek AJ. Statistical power and measurement allocation in ergonomic intervention studies assessing upper trapezius EMG amplitude – a case study of assembly work. J Electromyogr Kinesiol 2002;12:45–7. Moffet H, Hagberg M, Hansson-Risberg E, Karlqvist L. Influence of laptop computer design and working position on physical exposure variables. Clin Biomech (Bristol, Avon) 2002;17(5):368–75. Moore A, Wells R, Van Eerd D, Krajcarski S, Banina M, Cole DC, et al. Separation and summation of EMG recordings by task using video records. In: Proceedings of the international society of bio-mechanics XIXth congress, Dunedin, New Zealand, July 6–12; 2003. Nordander C, Hansson G-Å, Rylander L, Asterlan P, Byström UJ, Ohlsson K, et al. Muscular rest and gap frequency as EMG measures of physical exposure: the impact of work tasks and individual related factors. Ergonomics 2000;43(11):1904–19. Ortiz DJ, Marcus M, Gerr F, Jones W, Cohen S. Measurement variability in upper extremity posture among VDT users. Appl Ergon. 1997;28:139–43. Polanyi MFD, Cole DC, Beaton DE, Chung J, Wells R, Abdolell M. Upper limb workrelated musculoskeletal disorders among newspaper employees: crosssectional survey results. Am J Ind Med 1997;32:620–8. Punnett L, Wegman DH. Work-related musculoskeletal disorders: the epidemiologic evidence and the debate. J Electromyogr Kinesiol 2004;14:13–23. SAS Institute Inc.. SAS procedures guide. Version 6. 3rd ed. Cary, NC: SAS Institute; 1990. Simoneau GG, Marklin RW, Berman JE. Effect of computer keyboard slope on wrist position and forearm electromyography of typists without musculoskeletal disorders. Phys Ther 2003;83:816–30. Spielholz P, Silverstein B, Morgan M, Checkoway H, Kaufman J. Comparison of self-report, video observation and direct measurement methods for upper extremity musculoskeletal disorder physical risk factors. Ergonomics 2001;44(6):588–613. Straker L, Pollock C, Burgess-Limerick R, Skoss R, Coleman J. The impact of computer height and desk design on muscle activity during information technology work by young adults. J Electromyogr Kinesiol 2008;18:606–17. Takala EP, Pehkonen I, Forsman M, Hansson GA, Mathiassen SE, Neumann WP, et al. Systematic evaluation of observational methods assessing biomechanical exposures at work. Scand J Work Environ Health 2010;36:3–24. van der Beek AJ, Frings-Dresen MH. Assessment of mechanical exposure in ergonomic epidemiology. Occup Environ Med 1998;55(5):291–9. Van Eerd D, Hogg-Johnson S, Mazumder A, Cole D, Wells R, Moore A. Task exposures in an office environment: a comparison of methods. Ergonomics 2009;52(10):1248–58. Veiersted KB, Westgaard RH, Andersen P. Electromyographic evaluation of muscular work pattern as a predictor of trapezius myalgia. Scand J Work Environ Health 1993;19(4):284–90. Venables WN, Ripley BD. Modern applied statistics with S-PLUS. 3rd ed. Springer; 1999.

Wells R, Norman R, Neumann P, Andrews D, Frank J, Shannon H, et al. Assessment of physical work load in epidemiologic studies: common measurement metrics for exposure assessment. Ergonomics 1997;40(1):51–61. Wells R, Van Eerd D, Hagg G. Mechanical exposure concepts using force as the agent. Scand J Work Environ Health 2004;30(3):179–90. Westgaard RH, Winkel J. Ergonomic intervention research for improved musculoskeletal health: a critical review. Int J Ind Ergon 1997;20:463–500.

Dwayne Van Eerd is an associate scientist at the Institute for Work & Health. He has both an M.Sc. and B.Sc. in kinesiology from the University of Waterloo and an M.Sc. in health research methodology from McMaster University. He is currently enrolled in the doctoral program in Applied Health Sciences at the University of Waterloo. Involved in clinical work since 1990, Van Eerd has designed and implemented rehabilitation and ergonomic programs for injured workers. He has been a researcher with the Institute for Work & Health since 1997. The focus of Van Eerd’s research has been on the classification and prevention of work-related musculoskeletal disorders. His research projects have included an exploration of the mechanical exposures involved in office work and the implementation of participatory interventions. Recently, Van Eerd has led or participated in a number of systematic reviews of the scientific literature addressing prevention of musculoskeletal disorders.

Dr. Sheilah Hogg-Johnson is the Interim Scientific Director at the Institute for Work & Health. She is also the chief privacy officer at the Institute. Hogg-Johnson is an assistant professor in the Dalla Lana School of Public Health, Faculty of Medicine at the University of Toronto. She holds a PhD in biostatistics from the University of Toronto and a MMath in statistics from the University of Waterloo. Hogg-Johnson is a specialist in statistical methodology and prognostic modeling. Her current research interests include identifying drivers of long disability duration and the role of inspection and enforcement in preventing work injuries.

Dr. Donald C Cole trained as a physician and practiced primary care, public health, occupational health and environmental health in a variety of settings in Canada and lower and middle income countries. After a residency at McMaster, he qualified as a Royal College fellow in Occupational Medicine (1990) and Community Medicine (1992). With the International Potato Center he has co-lead research on pesticides, urban agriculture and nutrition. A Tri-Council Eco-Research fellowship in environmental epidemiology led to research on environmental contaminants, ecosystems and human health. The role of Interim Director of Research at the Institute for Work & Health fostered research on evaluation of complex workplace interventions to reduce the burden of musculoskeletal disorders and other morbidity. Through the Dalla Lana School of Public Health, he currently teaches, mentors, and contributes research evidence to public health practice and policy, with an interest in evaluation of global health capacity development.

Dr. Richard Wells is a Professor in the Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, where he specializes in work-related musculoskeletal disorders of the back and upper limbs; their causes, patho-physiology and prevention. He addresses these issues using anatomical and functional anatomical studies in cadavers and volunteers, by biomechanical modeling of the structures affected. Another approach develops measurement, recording and processing approaches to document exposure at work and using these methods in epidemiological studies to assess the work-relatedness of various workplace exposures. Lastly he is involved in the development of workplace processes and interventions to prevent musculoskeletal disorders.

D. Van Eerd et al. / Journal of Electromyography and Kinesiology 22 (2012) 176–185 Dr Anjali Mazumder is a Statistical Consultant and Researcher with over 9 years of experience working in UK and Canadian institutions, conducting research and providing statistical support to a wide range of projects in forensic science, health, education, and international development. She holds an MSc in Statistics and an MA in Measurement and Evaluation both from the University of Toronto. She also holds a doctorate in Statistics from the University of Oxford. Her primary interests are in the measurement, evaluation, and interpretation of evidence for decision-making under uncertainty. She is currently developing statistical methods for research and development of new evaluation methods and technologies for forensic intelligence, supporting casework, training scientists in statistics and probabilistic reasoning, and informing procedures and guidelines across a range of evidence types. She is actively engaged in several initiatives that promote the communication of statistics and probabilistic reasoning in serving the justice system and the broader remit of decision-making in public policy in the UK.

185