Human Factors: The Journal of the Human Factors and Ergonomics Society http://hfs.sagepub.com/
Should We Be More on the Ball?: The Efficacy of Accommodation Training on Lumbar Spine Posture, Muscle Activity, and Perceived Discomfort During Stability Ball Sitting Jennie A. Jackson, Priyanka Banerjee-Guénette, Diane E. Gregory and Jack P. Callaghan Human Factors: The Journal of the Human Factors and Ergonomics Society 2013 55: 1064 originally published online 28 March 2013 DOI: 10.1177/0018720813482326 The online version of this article can be found at: http://hfs.sagepub.com/content/55/6/1064
Published by: http://www.sagepublications.com
On behalf of:
Human Factors and Ergonomics Society
Additional services and information for Human Factors: The Journal of the Human Factors and Ergonomics Society can be found at: Email Alerts: http://hfs.sagepub.com/cgi/alerts Subscriptions: http://hfs.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Dec 19, 2013 Downloaded from hfs.sagepub.com Bibliotheque de l'Universite Laval on28, June 7,2013 2014 OnlineFirst Versionat of Record - Mar
What is This?
482326 2013
HFS55610.1177/0018720813482326Month XXXX - Human FactorsEffect of Accommodation on Stability Ball Sitting
Should We Be More on the Ball? The Efficacy of Accommodation Training on Lumbar Spine Posture, Muscle Activity, and Perceived Discomfort During Stability Ball Sitting Jennie A. Jackson, University of Gävle, Gävle, Sweden, and Uppsala University, Uppsala, Sweden, Priyanka Banerjee-Guénette, University of Waterloo, Waterloo, Canada, Diane E. Gregory, Wilfrid Laurier University, Waterloo, Canada, and Jack P. Callaghan University of Waterloo, Waterloo, Canada Objective: The aim of this study was to evaluate the efficacy of a 9-day accommodation protocol on reducing perceived discomfort while sitting on a stability ball (SB); trunk muscle activity levels and lumbar spinal postures were also considered. Background: Previous studies have compared SB sitting with office chair sitting with few observed differences in muscle activity or posture; however, greater discomfort during SB sitting has been reported. These findings may indicate an accommodation period is necessary to acclimate to SB sitting. Method: For this study, 6 males and 6 females completed two separate, 2-hr sitting sessions on an SB. Half the participants completed a 9-day accommodation period between the visits, whereas the other half did not use an SB during the time. On both occasions, self-reported perceived discomfort ratings were collected along with erector spinae and abdominal muscle activity and lumbar spinal postures. Results: Discomfort ratings were reduced in female participants following the accommodation; no effects on muscle activation or lumbar spine postures were observed. Conclusion: Accommodation training may reduce perceived low-back discomfort in females. Trunk muscle activity and lumbar spine postures during seated office work on an SB did not differ between groups; however, greater sample power was required to conclusively address these variables. Application: Regarding whether to use an SB in place of a standard office chair, this study indicates that females electing to use an SB can decrease discomfort by following an accommodation protocol; no evidence was found to indicate that SB chair use will improve trunk strength or posture, even following an accommodation period. Keywords: sitting, low-back pain, spine biomechanics, office work, ergonomics, chair design, discomfort Address correspondence to Jack P. Callaghan, Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, Waterloo, ON N2L 3G1, Canada; e-mail: [email protected]. HUMAN FACTORS Vol. 55, No. 6, December 2013, pp. 1064–1076 DOI: 10.1177/0018720813482326 Copyright © 2013, Human Factors and Ergonomics Society.
Introduction The use of a stability ball (SB) as a piece of exercise equipment has become increasingly common in recent years. This device (also known as a Swiss, gym, exercise, balance, or physio ball) is often used in both healthy persons and those with low-back pain (LBP) to enhance core stability and improve balance and control. In the last decade, the device has also been promoted as an alternative to an office chair, primarily for those who suffer sittingrelated LBP. Manufacturers of SB seating solutions have claimed that using the device as an office chair will promote “active sitting” (Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a, n.d.b) involving constant small adjustments (Posture Perfect Solutions, 2010c) and body movements (Ball Chair Headquarters, 2012b), release muscle tension (Posture Perfect Solutions, 2010c; Sit-A-Round, n.d.a), strengthen core muscles (Ball Chair Headquarters, 2012a; Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a), and build core muscle endurance (Posture Perfect Solutions, 2010c). SB sitting is also said to improve seated spinal posture (Ball Chair Headquarters, 2012c; Posture Perfect Solutions, 2010b; Sit-A-Round, n.d.a) and relieve spinal compression (Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a). Finally, SB sitting is also said to reduce back pain (Ball Chair Headquarters, 2012c; Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a) and even prevent back injuries (Sit-A-Round, n.d.a). Established links have been shown both between increased trunk strength and reduced LBP (Bentsen, Lindgarde, & Manthorpe, 1997; Handa, Yamamoto, Tani, Kawakami, & Take-
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
masa, 2000), and between improved posture and reduced LBP (Harrison, Harrison, Croft, Harrison, & Troyanovich, 1999; Williams, Hawley, McKenzie, & van Wijmen, 1991). Given SB manufacturers’ claims, we were motivated to investigate the scientific merit of SB use in lieu of an office chair during computer work. In addition to the potential benefits of increased trunk strength and improved posture, several previously proposed mechanisms for LBP development may additionally explain what motivates SB manufacturer claims. During sitting, the lumbar spine assumes a more flexed (kyphotic) posture than in upright standing (Keegan, 1953; Lord, Small, Dinsay, & Watkins, 1997). A more kyphotic (or less lordotic) lumbar posture has several implications for the occurrence of LBP. First, flexed postures displace the nucleus pulposus posteriorly (Alexander, Hancock, Agouris, Smith, & MacSween, 2007), which can place substantial tensile stress on the posterior aspect of the intervertebral disc (Adams & Hutton, 1985; Adams, Mcnally, Chinn, & Dolan, 1994), potentially contributing to disc herniation. Next, static loads during sitting have also been proposed as a cause of LBP (Callaghan & McGill, 2001). When there is a lack of motion in the lumbar spine, the provision of sufficient nutrition to the intervertebral discs is compromised (Holm & Nachemson, 1983). Statically maintained contractions have also been linked to muscle fatigue attributable to a lack of rest periods (Jonsson, 1978; Veiersted, Westgaard, & Andersen, 1990). In short, if using an SB in lieu of an office chair leads to more dynamic sitting, both in terms of muscle activation levels and spinal motions, then SB usage may reduce, or even eliminate, previously proposed risk factors for LBP. Low-back disorders can be viewed as a progression of events beginning with the application of a physical load, followed by the development of discomfort, then physical symptoms, and finally, the development of a disorder (injury or illness; Ferguson & Marras, 1997). Monitoring trunk postures and trunk muscle activity levels is feasible for (indirectly) evaluating applied loads to the lumbar spine, and discomfort surveys are commonly used as an outcome measure to evaluate relative discomfort levels in different
1065
conditions. In this paradigm, decreased discomfort would be associated with decreased likelihood of progression along the disorder pathway and vice versa. To date, two case studies, three quantitative laboratory studies assessing biomechanical variables, and one quantitative laboratory study assessing energy expenditure have been published comparing SB with office chair sitting. From the two case studies, a total of three patients reported a reduction in LBP as a result of sitting on an SB, both in the absence of additional treatments (Merritt, 2001) and in concert with chiropractic care and exercise therapies (Merritt & Merritt, 2007), thus providing anecdotal evidence in support of SB manufacturer claims. From an empirical perspective, however, there has been little evidence supporting the proposed physical changes or the effectiveness of SB use in the treatment or prevention of LBP. Gregory, Dunk, and Callaghan (2006) explored differences in participants sitting on an SB and sitting on a standard office chair while performing computer workstation tasks for 1 hr. Participants reported higher whole-body discomfort while using the SB; however, only minimal differences in biomechanical responses were found. Increased activation levels of the thoracic erector spinae muscles were shown while participants were sitting on the ball, although this difference was relatively small (less than 1% maximum voluntary contraction [MVC]). An increase in anterior pelvic tilt was also shown in participants during SB sitting; however, no differences were found in lumbar spine postures (Gregory et al., 2006). McGill, Kavcic, and Harvey (2006) also reported increased discomfort among participants sitting on an SB compared with sitting on a stool while watching a movie for 30 min; however, they did not find any biomechanical differences between stool and SB sitting in terms of muscle activations, lumbar spine postures, lumbar spine stability, or lumbar spine compression. Most recently, Kingma and van Dieen (2009) investigated differences in participants sitting on an SB and sitting on a standard office chair while performing a 1-hr typing task. Unlike in previous studies, increased trunk motion and increased
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1066 December 2013 - Human Factors Table 1: Mean (± SD) Values for Age, Height, Mass, and Body Mass Index (BMI) for Male and Female Participants Sex
Age (years)
Height (m)
Mass (kg)
BMI (kg·m–2)
Male (n = 6) Female (n = 6)
24.5 ± 3.7 23.2 ± 3.9
1.82 ± 0.04 1.66 ± 0.09
86.4 ± 7.4 62.3 ± 7.6
25.8 ± 1.9 22.8 ± 2.3
variation in trunk muscle activity were observed during ball sitting. An increase in lumbar electromyography (EMG) data was also observed, along with increased spinal shrinkage for ball trials compared with office chair trials, that is, an increase in factors that could be associated with increased risk of LBP. Increased energy expenditure during clerical work performed while sitting on an SB compared with an office chair has also been reported (Beers, Roemmich, Epstein, & Horvath, 2008). This finding would be expected with increased lumbar (L3) muscle activity, as was reported by Kingma and van Dieen (2009), but could also result from increased activity in other deep muscles for which involvement was not measured in any of the three biomechanical papers. Absent from all prior empirical studies is the inclusion of a recommended accommodation training period to accompany an individual’s introduction to an SB. For example, Posture Perfect Solutions (2007) recommends a 4-week “accommodation” period to become accustomed to sitting on the SB for an entire 8-hr working day and to gradually build up strength. This protocol progressively increases the daily time spent sitting on the ball until a “full working day level of fitness is achieved” (Posture Perfect Solutions, 2007). Consider that SBs were originally introduced to take advantage of the increased stress on the neuromuscular system offered by the unstable surface training (Behm, Drinkwater, Willardson, & Cowley, 2010). Furthermore, intermuscular coordination for a specific task must be developed through practice. The provision of an accommodation period to allow users to adjust to sitting on an SB may therefore be an important factor in the efficacy of an SB as an office chair. The lack of user accommodation in prior experimental evaluations may explain why changes in posture and muscle activity have not, to date, been documented.
To facilitate a lab-based study comparable in duration to previous sitting studies and representative of a typical working-day continuous exposure, a 2-hr sitting task was selected. According to the Posture Perfect accommodation protocol, 9 days of training are required to develop a base fitness for a 2-hour period of ball sitting (Posture Perfect Solutions, 2007). Thus, the purpose of this intervention study was to evaluate the effect of a 9-day, progressive accommodation protocol on user discomfort scores during a 2-hr office work task performed on an SB. Our primary hypothesis was that a progressive accommodation protocol would reduce the magnitude of self-reported discomfort development across the working time. Our secondary aim was to investigate the effect of the accommodation training on trunk muscle activity and lumbar spine postural changes. Method Participants
The study was completed by 6 males and 6 females (Table 1). Potential participants were excluded if they had experienced LBP in the preceding 12 months, had undergone spinal surgery, or had a known spinal disorder. All participants were regular computer users and right-handed mousers, and none had previously used an SB. All participants reviewed and signed information and consent forms that had been approved by the University of Waterloo’s Office of Research Ethics. A priori power calculations indicated that substantially larger group sizes (n = 34) would be required to ensure a power level of .8 at an alpha level of .05. Such group sizes were outside the scope of the current study; however, this initial pilot work was conducted to explore discomfort variables while providing preliminary descriptions for EMG and kinematic variables.
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
1067
Table 2: Recommended Accommodation Training Protocol for Use With the Evolution Chair™ Week
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
1 2 3 4
0:15 1:45 3:30 5:15
0:20 2:00 3:45 5:30
0:30 2:15 4:00 6:00
0:40 2:30 4:15 3:00
1:00 2:45 4:30 6:30
1:15 3:00 4:45 7:00
1:30 3:15 5:00 8:00
Note. Bold font indicates the portion of the accommodation protocol used in the present study.
Accommodation Protocol
Participants were recruited in matched pairs based on height, weight, and sex. One individual from each pair was randomly assigned to the control group and the other to the accommodation group. All participants completed two data collection sessions interspaced by 9 days. During the interim period, accommodation group participants followed the first 9 days of the 28-day protocol recommended for users of the Evolution Chair™ (Posture Perfect Solutions, 2007) to reach a total sitting duration of 2 hr, the length of the data collection protocol (Table 2, shaded portion). Accommodation group participants were each supplied with an SB, were taught how to customize ball and base height to their stature, and were asked to complete the accommodation protocol at home. Following each day’s at-home accommodation session, participants were asked to rate their discomfort using a 100-mm visual analog scale (VAS) and state whether they would have preferred to sit on a traditional office chair or on the SB and were given the chance to provide additional comments. Participants in the control group had no exposure to the SB between the two laboratory visits. Instrumentation
Muscle activation levels and three-dimensional lumbar spine kinematics were collected in 14-min blocks throughout the 2-hr sitting protocol (eight blocks in total). VAS scores for upper-back, lower-back, buttocks, and overall discomfort were collected prior to commencing the seated office task (baseline) and after each 14-min block. One minute was allotted between blocks for participants to complete this survey; during this time, participants remained seated.
Surface EMG was collected bilaterally from two abdominal and two erector spinae muscles: external oblique, 15 cm lateral to the umbilicus; internal oblique, below external oblique and superior to the inguinal ligament; thoracic erector spinae, 5 cm lateral to T9 vertebrae; and lumbar erector spinae, 3 cm lateral to L3 vertebrae. A reference electrode was placed over the right iliac crest. EMG was recorded using pairs of AgAgCl electrodes (Blue Sensor, Medicotest Incorporated, Ølystykke, Denmark) with an interelectrode distance of approximately 3 cm. Raw EMG signals were band-pass filtered (10 to 1000 Hz) and differentially amplified using a Bortec Amplifier (Model AMT-8, Bortec Biomedical Limited, Calgary, Alberta, Canada; common-mode rejection ratio 115 dB at 60 Hz, 10 GΩ input impedance). Lumbar spine kinematic data were recorded with the use of active markers and Optotrak Certus position sensors (Northern Digital Inc., Waterloo, Ontario, Canada). Rigid plates with three noncollinear markers were applied to the skin over the spinous process of vertebra T12 and over the sacrum to facilitate the calculation of relative, three-dimensional lumbar spine angles. Time-synched EMG and kinematic data were collected with the use of Toolbench software (Northern Digital, Inc., Waterloo, Ontario, Canada). EMG data were collected at 2,048 samples per second and kinematic data at 32 Hz. Normalization
Isometric MVCs were performed for each of the instrumented muscle groups according to the procedures described by McGill (1991). A minimum of two MVC trials were performed for each muscle; an additional trial was performed if the participant or experimenter felt
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1068 December 2013 - Human Factors
Figure 1. (A) Stability ball (SB) with accompanying four-point plastic base on casters. (B) Experimental setup for 2-hr work task performed on an SB. Ball inflation and caster risers were set to ensure the angle between the thigh and the trunk was between 110° to 120°, as per the manufacturer’s suggestion.
a maximum effort had not been achieved. Adequate rest was given between MVC trials to prevent fatigue. An upright standing trial was collected as a neutral lumbar spine angle calibration. Maximum lumbar range of motion (ROMmax) trials were collected about all three axes from a standing posture with knees in extension, arms at the sides, and head and neck in a neutral posture with respect to the torso. Lumbar spine angle data during all seated tasks were expressed as percentages of the ROMmax collected during these trials. Data Collection Protocol
During both data collection sessions, participants sat on an Evolution Chair, which consisted of an SB on a four-point base equipped with casters (Figure 1A). The SB was fit to each participant according to the manufacturers’ specifications, ensuring a starting hip posture of 110° to 120° flexion with both feet planted on the ground (Figure 1B). The experiment was conducted without shoes to standardize across participants and to match the anticipated accommodation training conditions: It was expected most participants would remove their shoes at
home. We adjusted the height of the ball using height inserts between the base and the casters and/or by altering the amount of air in the ball within the specified ball diameter range of 50 to 55 cm. The ball diameter and use of any height inserts were recorded to ensure consistency between days and for use by accommodation group participants in setting up their SB. The height of the workstation was adjusted so the starting elbow angle was 90° when the participant sat in an upright posture with relaxed shoulders. The experimental task was a 2-hr, seated, standardized typing and mousing task. Participants copied text from a series of articles provided in hard copy by typing the text into a blank computer document. After each paragraph, participants were required to highlight each row of typed text in a different color using only the mouse; no keyboard shortcuts were permitted. A document holder was provided and was positioned on the right side of the monitor. Participants were required to keep both feet flat on the floor and were not permitted to stand up at any time during the 2-hr protocol. Between 14-min blocks, when subjective discomfort ratings were recorded, participants were permitted to have a
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
drink of water and readjust their position on the ball, if necessary. Data Analysis
EMG data were full wave rectified, low-pass filtered with the use of a Butterworth filter with 2.5 Hz cutoff frequency (Brereton & McGill, 1998) to produce linear enveloped EMG and normalized to the maximum values obtained during MVC testing. For each muscle (eight channels) and each 14-min working block (eight blocks in total), average EMG and total number of EMG gaps were calculated. A gap was defined as a drop in EMG activity below 0.5% MVC for longer than 0.2 s and with a maximum duration of 2 s (Veiersted et al., 1990). Kinematic data were dual-pass filtered with the use of a Butterworth filter with a 6 Hz cutoff frequency (Winter, 2009). For each 14-min working block, mean lumbar flexion angle and number of postural shifts (Gregory & Callaghan, 2008) were determined for each axis of motion: flexion-extension, lateral bend, and axial twist. Perceived discomfort ratings recorded on the VAS were measured to the nearest millimeter for each body area (upper back, lower back, buttocks, and for overall discomfort). Comments from the take-home surveys completed by accommodation group participants during the training period were tabulated. Statistical Analysis
To determine whether differences occurred in posture, muscle activation, or discomfort between experimental groups, between sexes, or over time, we conducted a four-way, mixed general linear model with between-subject factors group (control or accommodation) and sex (male or female) and within-subject factors day (Visit 1 or 2) and block (eight work blocks). To compensate for multiple comparisons, we set the alpha level to p ≤ .01. Results Perceived Discomfort
In general, participants tended to report increased discomfort during the course of both 2-hr seated work periods (Visit 1 and Visit 2) in all regions examined (range: p = .01 to p =
1069
.04); however, the relative magnitude between visits of the increase in discomfort varied by group (control vs. accommodation) and by sex. Specifically, female accommodation participants developed less discomfort on the second visit as compared with the first, whereas only minimal differences in the level of developed discomfort were observed between visits in the female control group. For example, in the low back, the female accommodation group showed, on average, a 25.3-mm smaller increase in discomfort scores on the second visit compared with the first, whereas the female control group demonstrated, on average, a 2.3-mm smaller increase in this score. Interestingly, the opposite was observed in males: The male accommodation group showed, on average, a 3.0-mm smaller increase in low-back discomfort on the second visit as compared with the first, whereas the male control group showed, on average, a 15.0mm smaller increase in this score (Figure 2). A significant three-way interaction (Sex × Group × Visit) was found for overall discomfort (p = .01), with a similar trend shown for the lower back (p = .03) and buttocks (p = .04). All discomfort rating variables showed similar interaction patterns: Both control and experimental groups showed less discomfort on Visit 2, but the magnitude of the decrease differed by group, and opposite patterns for groups were observed between sexes (see Figure 2). Post hoc calculations showed that the effect size for between-group difference in selfreported discomfort measures was moderate. For example, the effect size for overall discomfort was 0.53. A post hoc power calculation indicated low power (0.2) for the finding of a significant group difference in overall discomfort reporting. Given the small sample size employed, this finding was not unexpected. Comments from the take-home surveys completed by accommodation group participants during the accommodation training often indicated increased discomfort and/or fatigue with increased time spent on the ball, for example: “long periods hurt”; “good for 30 min, but low back started to bother me”; “a bit sore in the back after the first half hour.” In addition, participants voted whether they would have preferred to sit on the ball or a regular office chair
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1070 December 2013 - Human Factors
Figure 2. Low-back discomfort scores for females and males by test group. Error bars denote ±1 standard deviation. Block denotes sequential bouts of time across the 2-hr task (14 min per block). Visit 1 data shown in solid line; Visit 2 data shown in dashed line.
for the duration of that day’s training session. Of a possible 53 votes (6 participants × 9 training days, with one missing vote), 62% of votes were in favor of their own office chair, whereas only 38% of votes were in favor of the SB. In terms of sex, males preferred the SB 50% of the time, whereas females preferred the ball only 30% of the time. In general, sitting preference did not change during the 9-day accommodation; that is, participants who preferred the ball on Day 1 also preferred the ball on Day 9 and similarly for those preferring the chair. Muscle Activation
Muscle activity levels during the 2-hr seated working task were, on average, low across groups and sexes: mean erector spinae EMG, 2.4% to 5.0% MVC (individual participant range = 0.4% to 13.9% MVC); mean abdominal EMG, 2.0% to 3.2% MVC (individual range = 0.5% to 8.3% MVC) (Table 3). No changes
were found in mean muscle activation levels between visits for any of the eight muscle sites for either the accommodation or the control group (p > .01). There were also no sex differences in the average EMG amplitudes (p > .01) at any of the recording sites. The three-way interactions between visit, experimental group, and block for the mean EMG amplitudes of the left thoracic erector spinae and left lumbar erector spinae muscles tended to differ between accommodation and control participants, indicating varying group responses in muscle activation levels during the course of the 2-hr seated task between visits (p = .04 and p = .05, respectively). The interaction was, however, driven by differing responses between experimental and control groups on the first visit and was not a result of the accommodation training (see Figure 3). No differences were found in the number of EMG gaps for each muscle between groups
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
1071
Table 3: Pooled Estimates by Sex (n = 6) of Mean (SD) Values for Main Outcome Variables Pre- and Postintervention Period Control Measure
Sex
1. Erector spinae EMG (% MVC)
Visit 2
Visit 1
L TES
2. Abdominal EMG (% MVC) 3. Lumbar angle (degrees) 4. No. of postural shifts
Visit 1
Accommodation
Male Female R TES Male Female L LES Male Female R LES Male Female L EO Male Female R EO Male Female L IO Male Female R IO Male Female Flexion Male Female Flexion Male Female Lateral bend Male Female Rotation Male Female
Visit 2
4.3 (2.3) 5.4 (2.0)
3.7 (1.9) 3.8 (0.7)
6.1 (3.8) 6.9 (4.2)
4.1 (2.1) 6.1 (2.6)
3.8 (2.6) 5.7 (1.5)
3.6 (2.3) 3.9 (1.6)
4.3 (2.1) 4.8 (0.9)
5.6 (2.4) 3.6 (0.7)
1.6 (0.9) 3.3 (1.2)
1.9 (1.0) 2.6 (1.3)
2.8 (0.8) 2.2 (0.9)
2.9 (0.7) 2.2 (1.0)
1.3 (0.6) 2.5 (1.0)
2.4 (1.0) 1.9 (0.8)
3.5 (1.5) 2.1 (0.7)
2.8 (0.8) 2.4 (0.8)
2.0 (0.8) 1.8 (0.6)
1.5 (0.1) 2.1 (1.1)
2.1 (0.7) 3.2 (0.6)
1.6 (0.4) 3.7 (1.6)
1.5 (0.3) 2.0 (0.7)
1.5 (0.1) 2.3 (0.9)
1.9 (1.4) 2.2 (0.9)
1.8 (0.9) 2.3 (0.6)
3.4 (2.1) 2.6 (1.7)
3.5 (3.4) 2.6 (1.9)
2.7 (0.7) 4.8 (0.8)
3.6 (1.3) 2.6 (1.3)
3.5 (2.4) 2.1 (0.5)
2.7 (2.0) 3.4 (2.4)
4.1 (1.0) 5.1 (0.7)
62.0 (17.1) 57.2 (16.9)
62.7 (9.0) 54.8 (22.5)
54.0 (26.4) 56.4 (11.8)
70.8 (39.8) 62.9 (21.2)
69.9 (54.4) 186.7 (286.5)
62.5 (48.3) 74.4 (63.5)
3.4 (1.0) 3.6 (2.2) 54.0 (24.6) 59.7 (12.3) 80.9 (69.6) 51.2 (36.3)
297.7 (118.0) 306.1 (175.1) 330.9 (130.4) 605.4 (401.1)
361.3 (192.1) 758.9 (663.7) 304.9 (143.7) 494.4 (321.5)
282.9 (86.4) 199.1 (59.0) 282.9 (96.0) 462.9 (362.9)
314.9 (130.7) 339.3 (101.1) 291.3 (131.6) 309.3 (128.5) (continued)
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1072 December 2013 - Human Factors Table 3: (continued) Control Measure
Sex
Visit 1
Accommodation Visit 2
Visit 1
Lower-back
Male
0–30
90–120 Female 0–30 90–120 Overall Male 0–30 90–120 Female 0–30 90–120
Visit 2
2.3 (3.9)
1.2 (1.7)
2.6 (2.8)
1.3 (2.1)
16.0 (14.0)
5.3 (5.4)
5.3 (3.0)
3.4 (3.2)
2.0 (2.4) 5.3 (3.7)
0.7 (1.0) 2.3 (2.3)
8.3 (7.0) 29.0 (3.8)
1.7 (1.9) 4.6 (4.7)
4.4 (3.8) 39.6 (29.3)
2.9 (3.9) 12.7 (11.5)
1.8 (1.6) 7.6 (6.8)
1.7 (3.0) 6.0 (6.4)
2.0 (1.7) 11.8 (7.0)
0.8 (0.7) 6.8 (3.5)
12.2 (6.5) 38.2 (8.5)
3.8 (2.8) 6.6 (5.2)
Note. Variables in Sections 1 through 4 are pooled estimates across the 2-hr work task; variables in Section 5 are pooled across the first (0–30 min) and last (90–120 min) half hour of each data collection (mean across Blocks 0 to 2 and 6 to 8, respectively). Standard deviations shown in parentheses. EMG = electromyography; L = left; R = right; TES = thoracic erector spinae; LES = lumbar erector spinae; EO = external oblique; IO = internal oblique.
Figure 3. Left thoracic erector spinae mean muscle activation level for each 14-min data collection block during the 2-hr working task by group and visit (data are collapsed across sex). Error bars denote ±1 standard deviation.
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
by visit for any of the eight muscles studied (p > .01). Kinematics
The mean lumbar flexion posture adopted across the 2-hr working period did not change significantly between visits for either the control (Visit 1, 59.6% ROMmax; Visit 2, 58.7% ROMmax) or the accommodation group (Visit 1, 55.2% ROMmax; Visit 2, 56.7% ROMmax). Furthermore, no postural differences were found between sexes, nor were interactions found at any level (p > .01; Table 3). The number of postural shifts (per block) was calculated to determine whether the pattern of lumbar spinal movements across the 2-hr work task differed by visit between groups. A trend toward larger numbers of lateral bending movements was found on Visit 2 compared with Visit 1 (p = .04); however, this difference was observed in both groups and in both men and women. No change was present in postural response patterns across the 2-hr task (i.e., across blocks) between visits. Discussion
In this study, we examined differences in self-reported discomfort levels during a 2-hr seated work task performed on an SB between a control group and a group undergoing an accommodation training protocol; differences between two sequential visits were compared between groups. Decreased overall discomfort scores resulted following the accommodation training, predominantly in female participants; there was also a trend toward decreased lowback and buttocks discomfort following the accommodation period. Kinematic and muscle activation levels were also investigated; however, these variables were examined in a pilot study sense as the group sizes used provided only minimal power for detecting differences in these variables. No empirical evidence was found to support the claim that accommodation training for sitting on an SB will affect trunk muscle activity and lumbar posture. Participants showed no change in the average amplitude of muscle activity between visits or in the muscle recruitment and rest patterns, as evaluated with a gaps analysis. No between-visit difference was
1073
found in mean lumbar flexion angle for either group (control or accommodation), indicating that an accommodation period did not alter the initial posture adopted on the SB. There was therefore no evidence to indicate that a short, progressive accommodation protocol altered muscular involvement or promoted dynamic sitting while working seated on an SB. Together, these findings indicate that initial perceived discomfort for initial users can be mitigated with an accommodation protocol, even in the absence of any postural or muscular changes. Since control group participants also showed reductions in discomfort between Visit 1 and Visit 2, it is unlikely that the accommodation training is the only factor contributing to the decline in discomfort; task familiarity may have also played a role. Authors of other studies who have previously examined SB sitting did not incorporate the effect of a strict accommodation training period on lumbar posture, muscle activity, or discomfort. Earlier studies by Gregory et al. (2006), McGill et al. (2006), and Kingma and van Dieen (2009) showed relatively low muscle activation levels during SB sitting, ranging from
Should We Be More on the Ball?: The Efficacy of Accommodation Training on Lumbar Spine Posture, Muscle Activity, and Perceived Discomfort During Stability Ball Sitting Jennie A. Jackson, Priyanka Banerjee-Guénette, Diane E. Gregory and Jack P. Callaghan Human Factors: The Journal of the Human Factors and Ergonomics Society 2013 55: 1064 originally published online 28 March 2013 DOI: 10.1177/0018720813482326 The online version of this article can be found at: http://hfs.sagepub.com/content/55/6/1064
Published by: http://www.sagepublications.com
On behalf of:
Human Factors and Ergonomics Society
Additional services and information for Human Factors: The Journal of the Human Factors and Ergonomics Society can be found at: Email Alerts: http://hfs.sagepub.com/cgi/alerts Subscriptions: http://hfs.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Dec 19, 2013 Downloaded from hfs.sagepub.com Bibliotheque de l'Universite Laval on28, June 7,2013 2014 OnlineFirst Versionat of Record - Mar
What is This?
482326 2013
HFS55610.1177/0018720813482326Month XXXX - Human FactorsEffect of Accommodation on Stability Ball Sitting
Should We Be More on the Ball? The Efficacy of Accommodation Training on Lumbar Spine Posture, Muscle Activity, and Perceived Discomfort During Stability Ball Sitting Jennie A. Jackson, University of Gävle, Gävle, Sweden, and Uppsala University, Uppsala, Sweden, Priyanka Banerjee-Guénette, University of Waterloo, Waterloo, Canada, Diane E. Gregory, Wilfrid Laurier University, Waterloo, Canada, and Jack P. Callaghan University of Waterloo, Waterloo, Canada Objective: The aim of this study was to evaluate the efficacy of a 9-day accommodation protocol on reducing perceived discomfort while sitting on a stability ball (SB); trunk muscle activity levels and lumbar spinal postures were also considered. Background: Previous studies have compared SB sitting with office chair sitting with few observed differences in muscle activity or posture; however, greater discomfort during SB sitting has been reported. These findings may indicate an accommodation period is necessary to acclimate to SB sitting. Method: For this study, 6 males and 6 females completed two separate, 2-hr sitting sessions on an SB. Half the participants completed a 9-day accommodation period between the visits, whereas the other half did not use an SB during the time. On both occasions, self-reported perceived discomfort ratings were collected along with erector spinae and abdominal muscle activity and lumbar spinal postures. Results: Discomfort ratings were reduced in female participants following the accommodation; no effects on muscle activation or lumbar spine postures were observed. Conclusion: Accommodation training may reduce perceived low-back discomfort in females. Trunk muscle activity and lumbar spine postures during seated office work on an SB did not differ between groups; however, greater sample power was required to conclusively address these variables. Application: Regarding whether to use an SB in place of a standard office chair, this study indicates that females electing to use an SB can decrease discomfort by following an accommodation protocol; no evidence was found to indicate that SB chair use will improve trunk strength or posture, even following an accommodation period. Keywords: sitting, low-back pain, spine biomechanics, office work, ergonomics, chair design, discomfort Address correspondence to Jack P. Callaghan, Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, Waterloo, ON N2L 3G1, Canada; e-mail: [email protected]. HUMAN FACTORS Vol. 55, No. 6, December 2013, pp. 1064–1076 DOI: 10.1177/0018720813482326 Copyright © 2013, Human Factors and Ergonomics Society.
Introduction The use of a stability ball (SB) as a piece of exercise equipment has become increasingly common in recent years. This device (also known as a Swiss, gym, exercise, balance, or physio ball) is often used in both healthy persons and those with low-back pain (LBP) to enhance core stability and improve balance and control. In the last decade, the device has also been promoted as an alternative to an office chair, primarily for those who suffer sittingrelated LBP. Manufacturers of SB seating solutions have claimed that using the device as an office chair will promote “active sitting” (Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a, n.d.b) involving constant small adjustments (Posture Perfect Solutions, 2010c) and body movements (Ball Chair Headquarters, 2012b), release muscle tension (Posture Perfect Solutions, 2010c; Sit-A-Round, n.d.a), strengthen core muscles (Ball Chair Headquarters, 2012a; Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a), and build core muscle endurance (Posture Perfect Solutions, 2010c). SB sitting is also said to improve seated spinal posture (Ball Chair Headquarters, 2012c; Posture Perfect Solutions, 2010b; Sit-A-Round, n.d.a) and relieve spinal compression (Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a). Finally, SB sitting is also said to reduce back pain (Ball Chair Headquarters, 2012c; Posture Perfect Solutions, 2010a; Sit-A-Round, n.d.a) and even prevent back injuries (Sit-A-Round, n.d.a). Established links have been shown both between increased trunk strength and reduced LBP (Bentsen, Lindgarde, & Manthorpe, 1997; Handa, Yamamoto, Tani, Kawakami, & Take-
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
masa, 2000), and between improved posture and reduced LBP (Harrison, Harrison, Croft, Harrison, & Troyanovich, 1999; Williams, Hawley, McKenzie, & van Wijmen, 1991). Given SB manufacturers’ claims, we were motivated to investigate the scientific merit of SB use in lieu of an office chair during computer work. In addition to the potential benefits of increased trunk strength and improved posture, several previously proposed mechanisms for LBP development may additionally explain what motivates SB manufacturer claims. During sitting, the lumbar spine assumes a more flexed (kyphotic) posture than in upright standing (Keegan, 1953; Lord, Small, Dinsay, & Watkins, 1997). A more kyphotic (or less lordotic) lumbar posture has several implications for the occurrence of LBP. First, flexed postures displace the nucleus pulposus posteriorly (Alexander, Hancock, Agouris, Smith, & MacSween, 2007), which can place substantial tensile stress on the posterior aspect of the intervertebral disc (Adams & Hutton, 1985; Adams, Mcnally, Chinn, & Dolan, 1994), potentially contributing to disc herniation. Next, static loads during sitting have also been proposed as a cause of LBP (Callaghan & McGill, 2001). When there is a lack of motion in the lumbar spine, the provision of sufficient nutrition to the intervertebral discs is compromised (Holm & Nachemson, 1983). Statically maintained contractions have also been linked to muscle fatigue attributable to a lack of rest periods (Jonsson, 1978; Veiersted, Westgaard, & Andersen, 1990). In short, if using an SB in lieu of an office chair leads to more dynamic sitting, both in terms of muscle activation levels and spinal motions, then SB usage may reduce, or even eliminate, previously proposed risk factors for LBP. Low-back disorders can be viewed as a progression of events beginning with the application of a physical load, followed by the development of discomfort, then physical symptoms, and finally, the development of a disorder (injury or illness; Ferguson & Marras, 1997). Monitoring trunk postures and trunk muscle activity levels is feasible for (indirectly) evaluating applied loads to the lumbar spine, and discomfort surveys are commonly used as an outcome measure to evaluate relative discomfort levels in different
1065
conditions. In this paradigm, decreased discomfort would be associated with decreased likelihood of progression along the disorder pathway and vice versa. To date, two case studies, three quantitative laboratory studies assessing biomechanical variables, and one quantitative laboratory study assessing energy expenditure have been published comparing SB with office chair sitting. From the two case studies, a total of three patients reported a reduction in LBP as a result of sitting on an SB, both in the absence of additional treatments (Merritt, 2001) and in concert with chiropractic care and exercise therapies (Merritt & Merritt, 2007), thus providing anecdotal evidence in support of SB manufacturer claims. From an empirical perspective, however, there has been little evidence supporting the proposed physical changes or the effectiveness of SB use in the treatment or prevention of LBP. Gregory, Dunk, and Callaghan (2006) explored differences in participants sitting on an SB and sitting on a standard office chair while performing computer workstation tasks for 1 hr. Participants reported higher whole-body discomfort while using the SB; however, only minimal differences in biomechanical responses were found. Increased activation levels of the thoracic erector spinae muscles were shown while participants were sitting on the ball, although this difference was relatively small (less than 1% maximum voluntary contraction [MVC]). An increase in anterior pelvic tilt was also shown in participants during SB sitting; however, no differences were found in lumbar spine postures (Gregory et al., 2006). McGill, Kavcic, and Harvey (2006) also reported increased discomfort among participants sitting on an SB compared with sitting on a stool while watching a movie for 30 min; however, they did not find any biomechanical differences between stool and SB sitting in terms of muscle activations, lumbar spine postures, lumbar spine stability, or lumbar spine compression. Most recently, Kingma and van Dieen (2009) investigated differences in participants sitting on an SB and sitting on a standard office chair while performing a 1-hr typing task. Unlike in previous studies, increased trunk motion and increased
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1066 December 2013 - Human Factors Table 1: Mean (± SD) Values for Age, Height, Mass, and Body Mass Index (BMI) for Male and Female Participants Sex
Age (years)
Height (m)
Mass (kg)
BMI (kg·m–2)
Male (n = 6) Female (n = 6)
24.5 ± 3.7 23.2 ± 3.9
1.82 ± 0.04 1.66 ± 0.09
86.4 ± 7.4 62.3 ± 7.6
25.8 ± 1.9 22.8 ± 2.3
variation in trunk muscle activity were observed during ball sitting. An increase in lumbar electromyography (EMG) data was also observed, along with increased spinal shrinkage for ball trials compared with office chair trials, that is, an increase in factors that could be associated with increased risk of LBP. Increased energy expenditure during clerical work performed while sitting on an SB compared with an office chair has also been reported (Beers, Roemmich, Epstein, & Horvath, 2008). This finding would be expected with increased lumbar (L3) muscle activity, as was reported by Kingma and van Dieen (2009), but could also result from increased activity in other deep muscles for which involvement was not measured in any of the three biomechanical papers. Absent from all prior empirical studies is the inclusion of a recommended accommodation training period to accompany an individual’s introduction to an SB. For example, Posture Perfect Solutions (2007) recommends a 4-week “accommodation” period to become accustomed to sitting on the SB for an entire 8-hr working day and to gradually build up strength. This protocol progressively increases the daily time spent sitting on the ball until a “full working day level of fitness is achieved” (Posture Perfect Solutions, 2007). Consider that SBs were originally introduced to take advantage of the increased stress on the neuromuscular system offered by the unstable surface training (Behm, Drinkwater, Willardson, & Cowley, 2010). Furthermore, intermuscular coordination for a specific task must be developed through practice. The provision of an accommodation period to allow users to adjust to sitting on an SB may therefore be an important factor in the efficacy of an SB as an office chair. The lack of user accommodation in prior experimental evaluations may explain why changes in posture and muscle activity have not, to date, been documented.
To facilitate a lab-based study comparable in duration to previous sitting studies and representative of a typical working-day continuous exposure, a 2-hr sitting task was selected. According to the Posture Perfect accommodation protocol, 9 days of training are required to develop a base fitness for a 2-hour period of ball sitting (Posture Perfect Solutions, 2007). Thus, the purpose of this intervention study was to evaluate the effect of a 9-day, progressive accommodation protocol on user discomfort scores during a 2-hr office work task performed on an SB. Our primary hypothesis was that a progressive accommodation protocol would reduce the magnitude of self-reported discomfort development across the working time. Our secondary aim was to investigate the effect of the accommodation training on trunk muscle activity and lumbar spine postural changes. Method Participants
The study was completed by 6 males and 6 females (Table 1). Potential participants were excluded if they had experienced LBP in the preceding 12 months, had undergone spinal surgery, or had a known spinal disorder. All participants were regular computer users and right-handed mousers, and none had previously used an SB. All participants reviewed and signed information and consent forms that had been approved by the University of Waterloo’s Office of Research Ethics. A priori power calculations indicated that substantially larger group sizes (n = 34) would be required to ensure a power level of .8 at an alpha level of .05. Such group sizes were outside the scope of the current study; however, this initial pilot work was conducted to explore discomfort variables while providing preliminary descriptions for EMG and kinematic variables.
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
1067
Table 2: Recommended Accommodation Training Protocol for Use With the Evolution Chair™ Week
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
1 2 3 4
0:15 1:45 3:30 5:15
0:20 2:00 3:45 5:30
0:30 2:15 4:00 6:00
0:40 2:30 4:15 3:00
1:00 2:45 4:30 6:30
1:15 3:00 4:45 7:00
1:30 3:15 5:00 8:00
Note. Bold font indicates the portion of the accommodation protocol used in the present study.
Accommodation Protocol
Participants were recruited in matched pairs based on height, weight, and sex. One individual from each pair was randomly assigned to the control group and the other to the accommodation group. All participants completed two data collection sessions interspaced by 9 days. During the interim period, accommodation group participants followed the first 9 days of the 28-day protocol recommended for users of the Evolution Chair™ (Posture Perfect Solutions, 2007) to reach a total sitting duration of 2 hr, the length of the data collection protocol (Table 2, shaded portion). Accommodation group participants were each supplied with an SB, were taught how to customize ball and base height to their stature, and were asked to complete the accommodation protocol at home. Following each day’s at-home accommodation session, participants were asked to rate their discomfort using a 100-mm visual analog scale (VAS) and state whether they would have preferred to sit on a traditional office chair or on the SB and were given the chance to provide additional comments. Participants in the control group had no exposure to the SB between the two laboratory visits. Instrumentation
Muscle activation levels and three-dimensional lumbar spine kinematics were collected in 14-min blocks throughout the 2-hr sitting protocol (eight blocks in total). VAS scores for upper-back, lower-back, buttocks, and overall discomfort were collected prior to commencing the seated office task (baseline) and after each 14-min block. One minute was allotted between blocks for participants to complete this survey; during this time, participants remained seated.
Surface EMG was collected bilaterally from two abdominal and two erector spinae muscles: external oblique, 15 cm lateral to the umbilicus; internal oblique, below external oblique and superior to the inguinal ligament; thoracic erector spinae, 5 cm lateral to T9 vertebrae; and lumbar erector spinae, 3 cm lateral to L3 vertebrae. A reference electrode was placed over the right iliac crest. EMG was recorded using pairs of AgAgCl electrodes (Blue Sensor, Medicotest Incorporated, Ølystykke, Denmark) with an interelectrode distance of approximately 3 cm. Raw EMG signals were band-pass filtered (10 to 1000 Hz) and differentially amplified using a Bortec Amplifier (Model AMT-8, Bortec Biomedical Limited, Calgary, Alberta, Canada; common-mode rejection ratio 115 dB at 60 Hz, 10 GΩ input impedance). Lumbar spine kinematic data were recorded with the use of active markers and Optotrak Certus position sensors (Northern Digital Inc., Waterloo, Ontario, Canada). Rigid plates with three noncollinear markers were applied to the skin over the spinous process of vertebra T12 and over the sacrum to facilitate the calculation of relative, three-dimensional lumbar spine angles. Time-synched EMG and kinematic data were collected with the use of Toolbench software (Northern Digital, Inc., Waterloo, Ontario, Canada). EMG data were collected at 2,048 samples per second and kinematic data at 32 Hz. Normalization
Isometric MVCs were performed for each of the instrumented muscle groups according to the procedures described by McGill (1991). A minimum of two MVC trials were performed for each muscle; an additional trial was performed if the participant or experimenter felt
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1068 December 2013 - Human Factors
Figure 1. (A) Stability ball (SB) with accompanying four-point plastic base on casters. (B) Experimental setup for 2-hr work task performed on an SB. Ball inflation and caster risers were set to ensure the angle between the thigh and the trunk was between 110° to 120°, as per the manufacturer’s suggestion.
a maximum effort had not been achieved. Adequate rest was given between MVC trials to prevent fatigue. An upright standing trial was collected as a neutral lumbar spine angle calibration. Maximum lumbar range of motion (ROMmax) trials were collected about all three axes from a standing posture with knees in extension, arms at the sides, and head and neck in a neutral posture with respect to the torso. Lumbar spine angle data during all seated tasks were expressed as percentages of the ROMmax collected during these trials. Data Collection Protocol
During both data collection sessions, participants sat on an Evolution Chair, which consisted of an SB on a four-point base equipped with casters (Figure 1A). The SB was fit to each participant according to the manufacturers’ specifications, ensuring a starting hip posture of 110° to 120° flexion with both feet planted on the ground (Figure 1B). The experiment was conducted without shoes to standardize across participants and to match the anticipated accommodation training conditions: It was expected most participants would remove their shoes at
home. We adjusted the height of the ball using height inserts between the base and the casters and/or by altering the amount of air in the ball within the specified ball diameter range of 50 to 55 cm. The ball diameter and use of any height inserts were recorded to ensure consistency between days and for use by accommodation group participants in setting up their SB. The height of the workstation was adjusted so the starting elbow angle was 90° when the participant sat in an upright posture with relaxed shoulders. The experimental task was a 2-hr, seated, standardized typing and mousing task. Participants copied text from a series of articles provided in hard copy by typing the text into a blank computer document. After each paragraph, participants were required to highlight each row of typed text in a different color using only the mouse; no keyboard shortcuts were permitted. A document holder was provided and was positioned on the right side of the monitor. Participants were required to keep both feet flat on the floor and were not permitted to stand up at any time during the 2-hr protocol. Between 14-min blocks, when subjective discomfort ratings were recorded, participants were permitted to have a
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
drink of water and readjust their position on the ball, if necessary. Data Analysis
EMG data were full wave rectified, low-pass filtered with the use of a Butterworth filter with 2.5 Hz cutoff frequency (Brereton & McGill, 1998) to produce linear enveloped EMG and normalized to the maximum values obtained during MVC testing. For each muscle (eight channels) and each 14-min working block (eight blocks in total), average EMG and total number of EMG gaps were calculated. A gap was defined as a drop in EMG activity below 0.5% MVC for longer than 0.2 s and with a maximum duration of 2 s (Veiersted et al., 1990). Kinematic data were dual-pass filtered with the use of a Butterworth filter with a 6 Hz cutoff frequency (Winter, 2009). For each 14-min working block, mean lumbar flexion angle and number of postural shifts (Gregory & Callaghan, 2008) were determined for each axis of motion: flexion-extension, lateral bend, and axial twist. Perceived discomfort ratings recorded on the VAS were measured to the nearest millimeter for each body area (upper back, lower back, buttocks, and for overall discomfort). Comments from the take-home surveys completed by accommodation group participants during the training period were tabulated. Statistical Analysis
To determine whether differences occurred in posture, muscle activation, or discomfort between experimental groups, between sexes, or over time, we conducted a four-way, mixed general linear model with between-subject factors group (control or accommodation) and sex (male or female) and within-subject factors day (Visit 1 or 2) and block (eight work blocks). To compensate for multiple comparisons, we set the alpha level to p ≤ .01. Results Perceived Discomfort
In general, participants tended to report increased discomfort during the course of both 2-hr seated work periods (Visit 1 and Visit 2) in all regions examined (range: p = .01 to p =
1069
.04); however, the relative magnitude between visits of the increase in discomfort varied by group (control vs. accommodation) and by sex. Specifically, female accommodation participants developed less discomfort on the second visit as compared with the first, whereas only minimal differences in the level of developed discomfort were observed between visits in the female control group. For example, in the low back, the female accommodation group showed, on average, a 25.3-mm smaller increase in discomfort scores on the second visit compared with the first, whereas the female control group demonstrated, on average, a 2.3-mm smaller increase in this score. Interestingly, the opposite was observed in males: The male accommodation group showed, on average, a 3.0-mm smaller increase in low-back discomfort on the second visit as compared with the first, whereas the male control group showed, on average, a 15.0mm smaller increase in this score (Figure 2). A significant three-way interaction (Sex × Group × Visit) was found for overall discomfort (p = .01), with a similar trend shown for the lower back (p = .03) and buttocks (p = .04). All discomfort rating variables showed similar interaction patterns: Both control and experimental groups showed less discomfort on Visit 2, but the magnitude of the decrease differed by group, and opposite patterns for groups were observed between sexes (see Figure 2). Post hoc calculations showed that the effect size for between-group difference in selfreported discomfort measures was moderate. For example, the effect size for overall discomfort was 0.53. A post hoc power calculation indicated low power (0.2) for the finding of a significant group difference in overall discomfort reporting. Given the small sample size employed, this finding was not unexpected. Comments from the take-home surveys completed by accommodation group participants during the accommodation training often indicated increased discomfort and/or fatigue with increased time spent on the ball, for example: “long periods hurt”; “good for 30 min, but low back started to bother me”; “a bit sore in the back after the first half hour.” In addition, participants voted whether they would have preferred to sit on the ball or a regular office chair
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1070 December 2013 - Human Factors
Figure 2. Low-back discomfort scores for females and males by test group. Error bars denote ±1 standard deviation. Block denotes sequential bouts of time across the 2-hr task (14 min per block). Visit 1 data shown in solid line; Visit 2 data shown in dashed line.
for the duration of that day’s training session. Of a possible 53 votes (6 participants × 9 training days, with one missing vote), 62% of votes were in favor of their own office chair, whereas only 38% of votes were in favor of the SB. In terms of sex, males preferred the SB 50% of the time, whereas females preferred the ball only 30% of the time. In general, sitting preference did not change during the 9-day accommodation; that is, participants who preferred the ball on Day 1 also preferred the ball on Day 9 and similarly for those preferring the chair. Muscle Activation
Muscle activity levels during the 2-hr seated working task were, on average, low across groups and sexes: mean erector spinae EMG, 2.4% to 5.0% MVC (individual participant range = 0.4% to 13.9% MVC); mean abdominal EMG, 2.0% to 3.2% MVC (individual range = 0.5% to 8.3% MVC) (Table 3). No changes
were found in mean muscle activation levels between visits for any of the eight muscle sites for either the accommodation or the control group (p > .01). There were also no sex differences in the average EMG amplitudes (p > .01) at any of the recording sites. The three-way interactions between visit, experimental group, and block for the mean EMG amplitudes of the left thoracic erector spinae and left lumbar erector spinae muscles tended to differ between accommodation and control participants, indicating varying group responses in muscle activation levels during the course of the 2-hr seated task between visits (p = .04 and p = .05, respectively). The interaction was, however, driven by differing responses between experimental and control groups on the first visit and was not a result of the accommodation training (see Figure 3). No differences were found in the number of EMG gaps for each muscle between groups
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
1071
Table 3: Pooled Estimates by Sex (n = 6) of Mean (SD) Values for Main Outcome Variables Pre- and Postintervention Period Control Measure
Sex
1. Erector spinae EMG (% MVC)
Visit 2
Visit 1
L TES
2. Abdominal EMG (% MVC) 3. Lumbar angle (degrees) 4. No. of postural shifts
Visit 1
Accommodation
Male Female R TES Male Female L LES Male Female R LES Male Female L EO Male Female R EO Male Female L IO Male Female R IO Male Female Flexion Male Female Flexion Male Female Lateral bend Male Female Rotation Male Female
Visit 2
4.3 (2.3) 5.4 (2.0)
3.7 (1.9) 3.8 (0.7)
6.1 (3.8) 6.9 (4.2)
4.1 (2.1) 6.1 (2.6)
3.8 (2.6) 5.7 (1.5)
3.6 (2.3) 3.9 (1.6)
4.3 (2.1) 4.8 (0.9)
5.6 (2.4) 3.6 (0.7)
1.6 (0.9) 3.3 (1.2)
1.9 (1.0) 2.6 (1.3)
2.8 (0.8) 2.2 (0.9)
2.9 (0.7) 2.2 (1.0)
1.3 (0.6) 2.5 (1.0)
2.4 (1.0) 1.9 (0.8)
3.5 (1.5) 2.1 (0.7)
2.8 (0.8) 2.4 (0.8)
2.0 (0.8) 1.8 (0.6)
1.5 (0.1) 2.1 (1.1)
2.1 (0.7) 3.2 (0.6)
1.6 (0.4) 3.7 (1.6)
1.5 (0.3) 2.0 (0.7)
1.5 (0.1) 2.3 (0.9)
1.9 (1.4) 2.2 (0.9)
1.8 (0.9) 2.3 (0.6)
3.4 (2.1) 2.6 (1.7)
3.5 (3.4) 2.6 (1.9)
2.7 (0.7) 4.8 (0.8)
3.6 (1.3) 2.6 (1.3)
3.5 (2.4) 2.1 (0.5)
2.7 (2.0) 3.4 (2.4)
4.1 (1.0) 5.1 (0.7)
62.0 (17.1) 57.2 (16.9)
62.7 (9.0) 54.8 (22.5)
54.0 (26.4) 56.4 (11.8)
70.8 (39.8) 62.9 (21.2)
69.9 (54.4) 186.7 (286.5)
62.5 (48.3) 74.4 (63.5)
3.4 (1.0) 3.6 (2.2) 54.0 (24.6) 59.7 (12.3) 80.9 (69.6) 51.2 (36.3)
297.7 (118.0) 306.1 (175.1) 330.9 (130.4) 605.4 (401.1)
361.3 (192.1) 758.9 (663.7) 304.9 (143.7) 494.4 (321.5)
282.9 (86.4) 199.1 (59.0) 282.9 (96.0) 462.9 (362.9)
314.9 (130.7) 339.3 (101.1) 291.3 (131.6) 309.3 (128.5) (continued)
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
1072 December 2013 - Human Factors Table 3: (continued) Control Measure
Sex
Visit 1
Accommodation Visit 2
Visit 1
Lower-back
Male
0–30
90–120 Female 0–30 90–120 Overall Male 0–30 90–120 Female 0–30 90–120
Visit 2
2.3 (3.9)
1.2 (1.7)
2.6 (2.8)
1.3 (2.1)
16.0 (14.0)
5.3 (5.4)
5.3 (3.0)
3.4 (3.2)
2.0 (2.4) 5.3 (3.7)
0.7 (1.0) 2.3 (2.3)
8.3 (7.0) 29.0 (3.8)
1.7 (1.9) 4.6 (4.7)
4.4 (3.8) 39.6 (29.3)
2.9 (3.9) 12.7 (11.5)
1.8 (1.6) 7.6 (6.8)
1.7 (3.0) 6.0 (6.4)
2.0 (1.7) 11.8 (7.0)
0.8 (0.7) 6.8 (3.5)
12.2 (6.5) 38.2 (8.5)
3.8 (2.8) 6.6 (5.2)
Note. Variables in Sections 1 through 4 are pooled estimates across the 2-hr work task; variables in Section 5 are pooled across the first (0–30 min) and last (90–120 min) half hour of each data collection (mean across Blocks 0 to 2 and 6 to 8, respectively). Standard deviations shown in parentheses. EMG = electromyography; L = left; R = right; TES = thoracic erector spinae; LES = lumbar erector spinae; EO = external oblique; IO = internal oblique.
Figure 3. Left thoracic erector spinae mean muscle activation level for each 14-min data collection block during the 2-hr working task by group and visit (data are collapsed across sex). Error bars denote ±1 standard deviation.
Downloaded from hfs.sagepub.com at Bibliotheque de l'Universite Laval on June 7, 2014
Effect of Accommodation on Stability Ball Sitting
by visit for any of the eight muscles studied (p > .01). Kinematics
The mean lumbar flexion posture adopted across the 2-hr working period did not change significantly between visits for either the control (Visit 1, 59.6% ROMmax; Visit 2, 58.7% ROMmax) or the accommodation group (Visit 1, 55.2% ROMmax; Visit 2, 56.7% ROMmax). Furthermore, no postural differences were found between sexes, nor were interactions found at any level (p > .01; Table 3). The number of postural shifts (per block) was calculated to determine whether the pattern of lumbar spinal movements across the 2-hr work task differed by visit between groups. A trend toward larger numbers of lateral bending movements was found on Visit 2 compared with Visit 1 (p = .04); however, this difference was observed in both groups and in both men and women. No change was present in postural response patterns across the 2-hr task (i.e., across blocks) between visits. Discussion
In this study, we examined differences in self-reported discomfort levels during a 2-hr seated work task performed on an SB between a control group and a group undergoing an accommodation training protocol; differences between two sequential visits were compared between groups. Decreased overall discomfort scores resulted following the accommodation training, predominantly in female participants; there was also a trend toward decreased lowback and buttocks discomfort following the accommodation period. Kinematic and muscle activation levels were also investigated; however, these variables were examined in a pilot study sense as the group sizes used provided only minimal power for detecting differences in these variables. No empirical evidence was found to support the claim that accommodation training for sitting on an SB will affect trunk muscle activity and lumbar posture. Participants showed no change in the average amplitude of muscle activity between visits or in the muscle recruitment and rest patterns, as evaluated with a gaps analysis. No between-visit difference was
1073
found in mean lumbar flexion angle for either group (control or accommodation), indicating that an accommodation period did not alter the initial posture adopted on the SB. There was therefore no evidence to indicate that a short, progressive accommodation protocol altered muscular involvement or promoted dynamic sitting while working seated on an SB. Together, these findings indicate that initial perceived discomfort for initial users can be mitigated with an accommodation protocol, even in the absence of any postural or muscular changes. Since control group participants also showed reductions in discomfort between Visit 1 and Visit 2, it is unlikely that the accommodation training is the only factor contributing to the decline in discomfort; task familiarity may have also played a role. Authors of other studies who have previously examined SB sitting did not incorporate the effect of a strict accommodation training period on lumbar posture, muscle activity, or discomfort. Earlier studies by Gregory et al. (2006), McGill et al. (2006), and Kingma and van Dieen (2009) showed relatively low muscle activation levels during SB sitting, ranging from
