Short-term effects of workstation exercises on musculoskeletal discomfort and postural changes in seated video display unit workers. (Research Report).

Two thirds of the North American workforce sit for all or part of their day, and most work with video display units (VDUs). (1) Extended work with VDUs is associated with 2 factors predictive of musculoskeletal problems: infrequent postural changes (2) and the presence of discomfort while sitting. (3) Posture, postural changes, and discomfort while sitting are all, in our opinion, interrelated. According to Teiger, (4) any sitting position achieved by VDU operators represents the resolution of ongoing conflicts between internal constraints (eg, need for comfort), external constraints (eg, worker/workstation fit), the task (eg, need for precision), and work organization factors (eg, performance expectations). Those same constraints and needs also dictate the frequency of postural changes in seared workers. (4,5) We refer to these changes as "in-chair movements" (ICMs). Sometimes the resolution of the conflict favors limited ICM, and sometimes it favors frequent ICM. (6,7)

Infrequent ICMs have long been considered to be a primary risk factor in the development of work-related musculoskeletal disorders (WRMDs). (8) Frequent ICMs have long been considered to be related to sitting discomfort. (6,7) Presumably, there is some balance between those 2 ends of the spectrum, with a pattern of ICM that does not lead to discomfort or risk of injury. Little evidence, however, exists. (7,9) A growing body of literature (6,7,10) has shown that, when first seated, working subjects are comfortable and move little. Their discomfort, however, increases over time and appears to lead to increases in ICM. (6,7,10) One challenge for researchers would be to determine whether promoting ICM when a VDU operator sits would limit the discomfort over time, thereby reducing the risk of musculoskeletal problems. (2,3)

Work with VDUs has changed from simple data-entry tasks to interactive task. For example, one rapidly expanding area of VDU work is sales and service via interactive, electronic means (ie, telephone and Internet connections). Sales and service work, typically conducted in call centers, is characterized by external workload pacing in which the pace of incoming calls is controlled by a computer, not by the operator. (11) External pacing conditions limit a VDU operator's ability to take self-selected breaks away from his or her workstation, (12) especially breaks to exercise. (13) Several exercise programs have been introduced for VDU operators. (14-20)

These exercise programs are designed to reduce musculoskeletal discomfort and increase ICM in VDU operators. Instruction for these exercises is now available in electronic format (www.tifaq.org/software.html), but only a few exercise programs have been field tested for their effectiveness in reducing discomfort. (14-17) None have been examined with VDU operators who work under paced workload conditions. The claim that workstation exercises increase ICM also has not been substantiated. Finally, despite the difficulties encountered in returning patients with musculoskeletal problems to seated work, no literature was located regarding workstation exercises. These issues are addressed in our study.

Workstation Exercises: Interventions and Critique

Research on the effects of workstation exercises for seated VDU operators has led to variable results. (14-20) Saltzman (17) showed reduced muscle stiffness and discomfort in VDU operators who took frequent short stretching breaks. Our review of the literature * (Tab. 1) shows quite another result. Most researchers reported that workstation exercises appear to be ineffective in reducing musculoskeletal discomfort. Each of these studies, however, has what we would consider limitations. Most ergonomists consider that there are 2 fundamental study design elements needed to measure discomfort. (7,9,21) There should be comparison of preintervention and postintervention discomfort, as well as measurement of discomfort over time. (20) In no cases were both components evaluated in the exercise studies we reviewed. The exercise programs may have failed to reduce discomfort because, with one exception, (16) they were developed by specialists (eg, exercise physiologists, physical therapists) based on the perceived demands of the VDU operators' tasks, rather than on specific health problems identified by the VDU operators (eg, musculoskeletal discomfort). In addition, adherence to exercise programs was taken into account only in the 3 most recent studies. (15-17) Consequently, not all exercise (or rest) breaks may have been taken. We question whether poor adherence may have diminished the effect of the exercises in these intervention studies. In-chair movement was not considered in any study that we reviewed.

Evaluation of Workstation Exercises

Researchers studying workstation exercises, in our opinion, have ignored 2 fundamental issues: electronic pacing and ICM. Silverstein et al (12) noted that, for workers being paced, potential gains from an exercise program may not occur because of productivity demands when workers return to their tasks. Paced workers face conflicting demands: the desire to perform exercises versus the need to maintain productivity. (15) To date, the effectiveness of workstation exercises under paced conditions has not been studied. The second issue is that physical therapists, (22) biomechanists, (23) and ergonomists (1,24) recognize the value of frequent ICMs in promoting musculoskeletal health. Despite that recognition, the effect of workstation exercises on increasing ICM has yet to be measured.

The first purpose of our study was to test the hypothesis that electronically paced directory assistance operators doing regular workstation exercises on a short-term basis (ie, 3- to 5-day shifts spread over 5-10 days) would, over the 2 hour test period: (1) increase their ICM, (2) not experience increases in musculoskeletal discomfort over time, and (3) have decreased whole-body musculoskeletal discomfort. The second purpose of our study, in order to assist our interpretation of the results, was to evaluate subjects' responses to the revised exercise program via an exercise satisfaction questionnaire.

Method

Subjects

Eleven people (3 male, 8 female) who had been employed as directory assistance operators for an average of 78 months (SE=18.6, range=20.5-190) participated on a volunteer basis. The subjects ranged in age from 22 to 41 years ([bar]X=29.5, SE=1.9). Their mean height was 172.9 cm (SE=2.1, range=162.5-185.4), and and mean weight was 69.6 kg (SE=3.9, range=49-84.4). All directory assistance operators in the company were notified of the study by use of posters and letters. The study was open to part-time and full-time operators and to men and women. All subjects had to have at least 3 months of experience in directory assistance operations. Subjects were excluded if, at any time during the 3 months leading up to the study, they experienced any musculoskeletal problems that required the intervention of a health care practitioner (25) or general health problems, such as hemorrhoids, that affected sitting comfort. Subjects wearing corrective eyewear were accepted into the study on the provision that their eyewear did not change during trials. For those participants who met the preceding criteria, only those directory assistance operators scheduled to work 5- to 7-day shifts during specified 10-day periods were accepted into the study. All volunteers who met all the preceding inclusion and exclusion criteria were accepted into the study.

Using statistical results [dagger] from previous subjects (N=8), (10) we calculated that a sample size of 15 was needed to achieve a power of .95. [double dagger] This was not achieved because of insufficient time prior to major restructuring that took place in the Canadian telecommunications industry. These same changes precluded the planned long-term (6-week) follow-up phase of the study. Prior to recruiting subjects, the employer and representatives from the directory assistance operators' union signed a participation agreement. Prior to participating in the study, subjects gave their signed informed consent. The research was approved by the Human Ethics Review Committee of Dalhousie University's Faculty of Graduate Studies using Research With Captive Populations guidelines from the Medical Research Council of Canada (now the Canadian Institute of Health Research).

Task Description

The VDU task used in this study (ie, directory assistance at a local telephone company) was chosen because the task, workload, environment, and furniture were standardized. (10) The operators' tasks were centered on a VDU and a headset and required no movements other than use of a keyboard, screen viewing, and speaking to customers via a headset. The operators primarily responded to requests for directory assistance and, to a lesser extent, handled intercepts on disconnected or out-of-service numbers.

Equipment Used for Measurement of Perceived Discomfort

Perceived discomfort was rated using the Body Part Discomfort Scale (BPDS), (21) a 6-point visual analog scale (0="no discomfort," 5="intolerable discomfort"). Discomfort was localized to body regions (eg, pelvis, shoulder) using the Localized Musculoskeletal Discomfort (LMD) body map. (26) The BPDS is based on the assumptions that discomfort is influenced by working posture, by the effort required to complete a task, and by the duration of the task. In their validation study, Corlett and Bishop (27) measured discomfort in spot welders (N=17) on multiple occasions over a 1-month period. Subjects worked under a range of conditions for posture (ie, suitable to unsuitable workstation height), foot pedal effort (ie, suitable to unsuitable force required), and time (up to 3 hours). Results showed that as work conditions varied (ie, suitable versus unsuitable) or as time passed, the level of discomfort varied. The results also showed that each level of discomfort had "a recognizable distribution about a mean value" (no statistics reported). For example, the combination of a workstation set too high, a foot pedal with high resistance, and working time greater than 2 hours results in an average BPDS rating in the "intolerable discomfort" range. Reliability of BPDS ratings of discomfort has not been reported. (7,21) In terms of the LMD body map, van der Grinten (26) showed that subjects (N=16) had the most reliable results for localizing uncomfortable body parts (Pearson r>.75, P<.05) when the LMD body map had only 19 body parts and individual body part ratings (eg, buttock) were combined to create regional scores (eg, buttock + thigh + knee + lower leg).

We analyzed whole-body scores for perceived discomfort (ie, the sum of the pelvis, buttock, back, neck, shoulder, and arm scores averaged over the number of uncomfortable body parts) obtained using the BPDS. Prior reliability studies, (13) in which the intraclass correlation coefficient (ICC [2,1])(28) was used, showed that whole-body scores were more reliable (ICC [2,1] >.90) than individual body part scores (eg, [ICC.sub.pelvis]=.63, [ICC.sub.neck]=.67) for measuring perceived discomfort.

Equipment Used for Measurement of In-Chair Movement

In-chair movement was measured by tracking center-of-pressure (COP) changes at the buttock-chair interface as subjects sat on a VERG (Vision Engineering Research Group) pressure-sensing mat. [section] The VERG mat is a 15 x 15 array of force-sensing resistors embedded in a 2-mm-thick rubber mat. Previous research showed that, as a result of the mat's thin profile, it readily conformed to chair seat contours and did not influence chair comfort. (13) In a series of trials with 12 subjects, Al-Eisa et al (29) demonstrated intertrial reliability of data obtained with the VERG mat for detecting mean pressure peaks (ICC [2,1]=.83-.97).

Validation. The COP is calculated by summing the pressure moments about the defined origin (0,0). Movement of the subject's COP is tracked in 2 planes simultaneously: forward and backward and left and right. We operationally defined in-chair movement as any movement of the chair occupant that changes the position of the COP. We based our operational definition on the following 3 assumptions about ICM and COP: that the seat supported most of the subject's body weight, that the predominant form of loading was compression, and that the acceleration components of seated postures were minimal. (10)

Data collection and reliability. Online data were collected continuously and stored in 24 five-minute blocks. Post-collection software was used to calculate the total COP distance traveled in the 2 horizontal dimensions for each block. Our ICM data analysis protocol was based on the results of a previous reliability study conducted on the job with 8 directory assistance operators. (10) We chose to collect ICM data as the means of 3 consecutive 5-minute blocks because the intertrial reliability of data obtained for 3 blocks (ICC [2,3]=.90-.96) was greater than the reliability of data obtained with any single 5-minute block (ICC [2,1]=.76-.89). Subjects were shown how the mat is used to collect interface pressure data, but they were masked to the fact that ICM would be derived from the interface pressure data by tracking the COP. Greater detail regarding the methods has been reported elsewhere. (10)

Dataspan exercises. A Dataspan Ergonomic Skills Training Program (30) was purchased by the telephone company 4 years prior to our study. The Appendix contains details of how we revised and implemented this program. At the end of the study, subjects were given an exercise satisfaction questionnaire about the revised Dataspan program. The questionnaire was based on the exercise critique guidelines described by Lee et al. (31) The questionnaire was completed and returned by mail. Subjects were asked to: (1) comment on the ease or difficulty of performing the exercises, doing the exercises without disrupting work, remembering to do the exercises, and doing the exercises without feeling conspicuous, (2) comment on the number and variety of exercises, (3) identify any problems encountered with particular exercises, and (4) provide general comments and suggestions.

Test Procedure

Subjects were tested twice, with each test session lasting 2 hours. Test sessions were conducted exclusively during day shifts, with each subject's first and second tests taking place at the same time of day. Because there was an influence of the time of day on discomfort, (32) the design was counterbalanced with 6 subjects tested at the start of a morning shift and 5 subjects tested after lunch. Subjects were first tested before the exercises were begun. At the end of that test, the revised Dataspan program was shown to the subjects, including explanations as well as demonstrations and practice of the exercise and relaxation techniques. The second test took place after the operators had worked from 3- to 5-day shifts, during which they were expected to take one workstation exercise break every 30 minutes.

For the exercises, subjects were instructed to stretch until they felt a comfortable sensation of muscle tension and then to hold that position for 5 seconds. The 5-second hold, although shorter than some current recommendations for holding a position during stretching (ie, 15-30 seconds), (33) was a compromise. The 5-second hold was based on the operator's allowable time for an exercise break (ie, number of exercises x number of repetitions x 5-second hold). Because exercise break time meant lost productivity (ie, time away from taking calls), the exercise break time was negotiated among, and agreed upon, by management, union representatives, and the researchers. Subjects were inconspicuously observed for adherence to the exercise program during the actual exercise tests and at least once between the first and second test sessions. This was done in a manner designed to make the observer inconspicuous. One researcher (AF) was available on site to provide follow-up as requested by the subjects regarding workstation exercises in general. Subjects' total duration in the program ranged from 6 to 10 days.

The test protocol for both ICM and perceived discomfort data collection is illustrated in Figure 1. Subjects provided perceived discomfort ratings at the 30-, 65-, and 115-minute marks. For each perceived discomfort rating, subjects were asked to close their eyes and evaluate discomfort by briefly focusing on each body part being characterized in turn. In order to evaluate ICM trends that coincided with ratings of perceived discomfort, ICM data were analyzed as the mean of 3 blocks at the start of the test (minutes 5-20), at the end of hour 1 (minutes 50-65), and the end of hour 2 (minutes 100-115). Subjects signed off their computers to rate their perceived discomfort. During the exercise test, subjects were not told to take an exercise break. Henning et al (15) suggested that encouragement to do workstation exercises is not enough to get subjects to do exercises. Therefore, subjects were cued visually (at 35, 65, and 90 minute) when 30 minutes had passed. These 30-minute cues were outside the time period when ICM data would be analyzed (Fig. 1).

[FIGURE 1 OMITTED]

Test Design

The independent variables evaluated were workstation exercises (preexercise, exercise) and time (start, hour 1, hour 2) within each test. The effects of exercise and time on ICM were measured with an exercise (2) x time (3) repeated-measures analysis of variance (ANOVA). For analysis of perceived discomfort data, a between-subjects factor (time of day: morning start/afternoon start) was added to the 2 within-subjects factors (exercise and time). (34) For both analyses (ICM, perceived discomfort), in the presence of nonzero two-way interactions, the analysis of simple main effects was used to evaluate the effects of each independent variable at each level of the other independent variable (eg, the effect of exercises at hours 1 and 2). All analyses were conducted using SPSS, [parallel] with alpha set at .05.

Results

In-Chair Movement

Tests of simple (ie, interaction) effects of exercises revealed that movement was greater with exercises in the first period (5-20 minutes) and second period (50-65 minutes), by 68.7 cm (P=.02) and 44.2 cm (P=.03), respectively (Fig. 2). In the final 15 minutes, there as no difference between pre-exercise and exercise conditions. Testing of time alone showed that ICM increased by 86.8 cm (P<.01) over the 2-hour test period in the pre-exercise condition. However, because ICM was consistently high throughout the exercise test, ICM did not increase over time when exercises were used.

[FIGURE 2 OMITTED]

Perceived Discomfort

Whole-body perceived discomfort was less (P<.01) when subjects exercised at 30 minutes (0.38 unit), 65 minutes (0.41 unit), and 115 minutes (0.93 unit) (Fig. 3). Perceived discomfort increased over time from 0.92 to 1.95 units (P<.01), but only in the pre-exercise condition.

[FIGURE 3 OMITTED]

Exercise Satisfaction Questionnaire

Questionnaire results (Tab. 2) showed that, generally, the workstation exercises were considered satisfactory and easy to perform, made the directory assistance operators feel better at day's end, and did not make them feel conspicuous (ie, exposed to unwelcome view). Although 3 subjects reported some minor discomfort, the operators' main concern was difficulty remembering or finding time to perform the exercises, particularly when "calls waiting" was posted on their computer screen.

Discussion

We believe that our study is unique because we demonstrated increases in ICM after introduction of workstation exercises. Results also showed that, when using workstation exercises, operators rated their musculoskeletal discomfort lower compared with the pre-exercise conditions.

Effect of Exercises on In-Chair Movement

The results upheld the hypothesis that ICM would increase after short-term use of workstation exercises. Following short-term use of the workstation exercise program, operators showed increases in ICM during the first and second test periods. The increases in ICM may be attributable to the following rationale presented to the operators regarding the exercises. During introductory sessions, physiological problems related to poor postures and infrequent postural changes (eg, buttock pressure, decreased circulation) were discussed. The operators were told that to reduce those problems, they needed to move (ie, to perform workstation exercises). The increases also may be due to factors beyond our control, including job stressors not measured in this study (eg, external workload pacing). The effects of these external factors on ICM cannot be ruled out, because the design (pretest-posttest) did not include a return to baseline (ie, pretest) conditions, which would have allowed us to isolate the treatment effects. We acknowledge that the increase in movements may be due to the motion associated with exercises. We believe, however, that the effect would be minor because the 30-minute time cues were intentionally delayed until after the first and second periods of ICM data collection. Although subjects knew they could do the exercises at will, they rarely did them before the 30-minute cues.

The uniqueness of our ICM measures, in our view, presents a challenge for interpretation. For example, the 44.2-cm difference in ICM between pre-exercise and exercise conditions during an average 5-minute period at the end of hour 1 was real. Clinical relevance is more difficult to establish. To do so, we first turn to previous calibration results. Fenety (13) showed that there was considerable variation in the amount of movement of seated subjects' COP during selected tasks and postural changes. Typical keyboard activities performed while taking telephone calls resulted in minimal COP movement (eg, <0.25 cm), whereas moving the trunk from upright to resting on the backrest resulted in a greater mean COP excursion of 2.5 cm (SD=0.5). Larger excursions occurred when subjects crossed (or uncrossed) their ankles ([bar]X=5.5 cm, SD=1.2) or when they moved from sitting upright to slouching ([bar]X=7.1 cm, SD=1.6). (13)

We believe the results of another study further clarify clinical relevance. Using a modified posture targeting techniques, (12) we randomly sampled the postures of 22 directory assistance operators in multiple 5-minute blocks over 5 working days. Fenety (13) showed that the operators primarily worked in rigid (ie, fixed) positions with occasional major postural changes, such as slouching. These major shifts in posture often occurred when the "calls waiting" signal was not present (ie, there was less time pressure). Considering the results from our calibration and posture targeting studies, it is likely that several larger ICM excursions took place in order to get that 44.2-cm difference in the 3-block average.

Finally, we advise caution regarding the interpretation of the ICM increases under exercise conditions because it is not known why seated subjects move at all. Using simultaneous measures of ICM and discomfort, some authors (6,7) have suggested that the 2 variables were correlated. The best evidence linking discomfort with the need to move, however, comes from Zhang and Helander. (35) Subjects associated sitting discomfort with feelings of constraint (eg, stiffness, cramping, restlessness, poor circulation), which suggests that ICM may serve to reduce those problems and, as a result, promote comfort. (35)

Effect of Exercises on Musculoskeletal Discomfort

Our hypotheses regarding musculoskeletal discomfort were supported by study findings. We found that short-term use of workstation exercises curbed the growth of discomfort over time. Like Swanson and Sauter, (18) we found that discomfort increased temporally in both test conditions, but in our study the effect was not significant when subjects were using workstation exercises. Although the workstation exercises did not eliminate discomfort, their greatest effect was to minimize the rapid growth of discomfort in the second hour, as shown in Figure 3. The results supported our second hypothesis that whole-body discomfort would be reduced under exercise conditions. In the only other study that evaluated exercise and pre-exercise conditions in VDU operators, Henning et al (14) found the only change after 3 weeks of exercises was a nonsignificant decrease in leg discomfort. Given the number of methodological differences between our study and that of Henning et al, (14) we believe the 2 studies are not comparable.

Implications of Workstation Exercises on Reducing Musculoskeletal Discomfort

Because directory assistance operators report high rates of musculoskeletal discomfort in regions such as the back (80%-84%) and neck (67%-72%), (36,37) our key finding that discomfort can be reduced by workstation exercises is important. Workstation exercises did not eliminate discomfort. Based on previous results, (14,18) more breaks are not likely the answer, either.

Research involving controlled clinical trials in the area of WRMDs is limited. Kuorinka and Forcier (38) acknowledged that the precision with which a given cause (ie, a work-related factor) results in a negative outcome (eg, WRMD) is low. We acknowledge that the likelihood that our intervention (ie, workstation exercises) would result in an improved outcome (ie, fewer musculoskeletal problems) is similarly low. Despite that limitation, one of the most common first symptoms associated with WRMDs is discomfort, (38) and we have demonstrated one means to reduce discomfort. A variety of other options have been suggested to reduce discomfort, including job rotation, (39) increased task variety, (9) and relaxation of muscles not involved directly in tasks. (40) Although it appears that the reduction of musculoskeletal discomfort will require a multifaceted, possibly expensive, approach, there will be benefits. Unfortunately, due to the short-term nature of our study, we cannot infer that reduced discomfort would result in cost savings or improved quality of worklife. There is, however, evidence that the introduction of stretching and rest breaks decreased the incidence of WRMDs over a 6-month period in 72 electronic-based sales and service representatives tracked over 30 months. (41)

General Considerations of Workstation Exercises

Method. Because our short-term use of a revised Dataspan exercise program resulted in improved comfort and increased ICM, we examined reasons for the program's success. First, we considered the program's content (eg, small number of exercises, brochure format) and administration (ie, exercises reviewed) as possible reasons. These changes may have benefited some, but not all, subjects, because over half of the directory assistance operators had difficulty remembering or finding time to perform the exercises due to time pressure. We considered other reasons. For example, revisions to the exercise program were based on specific musculoskeletal complaints made by the directory assistance operators during interviews. In addition, the operators may have benefited not just from the exercises, but also from breaks due to exercise periods. Zwahlen et al (42) have presented descriptive evidence that regular rest breaks are related to the temporal growth of musculoskeletal discomfort. In our study, the effects of taking a rest break, therefore, may be confounded with the effects of taking an exercise break. To date, no researcher has demonstrated any difference in perceived discomfort between rest breaks only and rest breaks plus exercises. (14,15,18)

Adherence. A common reason cited for the failure of long-term (>1 year) workplace exercise programs is that discomfort patterns are resistant to change. (12) We showed that changes occurred in less than a week. Long-term exercise programs, therefore, may fail for other reasons, such as nonadherence to the exercise program. Based on survey data, Silverstein et al (12) reported that a 1-year program to reduce discomfort in industrial workers was unsuccessful when two thirds of workers claimed participation that ranged from "some exercise" to "daily exercise." Participation (ie, adherence) can be monitored, but it generally requires human resources, such as team leaders. (30) Despite monitoring, the use of small groups, and the use of team leaders, the original Dataspan program at the telephone company where we conducted our study failed when introduced 4 years previously. Similarly, Silverstein et al (12) reported that, despite their subjects' reports of feeling better with exercises, their adherence to the exercise protocol declined. If, as it appears, exercise adherence requires repeated follow-ups and ongoing program modifications, the cost of workstation exercises may be high. Nonetheless, we believe those costs may be offset by the exercise benefits demonstrated in our study, namely the potential to decrease discomfort--a risk factor in the development of musculoskeletal problems. (3) Ultimately, to determine their ergonomic value, a cost-benefit analysis of workstation exercises would be required, including measures of productivity and long-term health effects.

Conflict between exercises and job performance. Answers to our questionnaire identified a potential problem that we contend is related more to ergonomics than to physical therapy. Seven of the 11 directory assistance operators reported difficulty in finding (or making) time to do the exercises. That is, although our results show that exercising reduced discomfort, the operators found it difficult to stop and do the exercises, particularly when they were under time pressure (eg, when calls were waiting). Those comments reinforce the contention of Henning et al (15) that poor adherence to exercise programs is most often due to a conflict between exercise demands and task demands, particularly when work is paced. The directory assistance operators in our study worked under electronic pacing and electronic performance monitoring. We believe that because of the pacing conditions, long-term adherence to the revised Dataspan program would likely decline unless some organizational changes occur. One solution to alleviate the conflicting demands of job performance and exercise breaks may be to shift the responsibility for "making time" for a break from the operator to management. The telephone company could accomplish this by programming online exercise prompts or by signaling breaks with light-emitting diodes mounted on each computer. (43) The change in responsibility would decrease the risk of forgeting the exercises and would signal that the break was sanctioned by management. Whether this would be beneficial warrants additional research.

Specific exercise problems. Minimal problems with the exercises were reported. Two operators (18%) reported wrist discomfort during wrist stretching. This figure is similar to the daily prevalence of wrist discomfort reported by telecommunication workers. (36) Considering that the extension-in-standing exercise was novel, the single complaint of discomfort with that exercise was lower than we expected. No problems were reported by subjects with respect to the duration of the muscle stretches. We used 5-second stretches to reduce the operators' time away from work and thereby encourage their adherence. Given the time constraints in directory assistance operations, the recommended 15- to 30-second holds for each muscle or muscle group stretched (33,44) would be unrealistic for workstation exercises. Therefore, in the future, we suggest that the short-duration workstation stretching exercises be continued by the directory assistance operators, but augmented with one daily set of prolonged (30-second hold) stretching exercises performed during "down" times (eg, not taking calls). We have no data, however, indicating this would be beneficial.

Study design. Our previous ICM reliability study (10) showed high between-subjects variability. This led us to not use: a control group. Instead, we thought it appropriate to use a repeated-measures design. Accordingly, we took a number of precautions to limit any Hawthorne effect. (45) First, subjects were masked to the fact that ICM data were being derived from the VERG mat recordings. Second, the entire directory assistance operations staff was invited to participate in some aspect of this study. For example, the majority of the company's 145 directory assistance operators took part in the exercise review and revision process, and all were given the opportunity to learn the revised exercise protocol. In addition, no preferences were given to the study volunteers with respect to breaks, salary, or shift assignments. Finally, at the start of this study, all directory assistance operators would likely have accommodated to the researchers who, at that point, had been present in the directory assistance operations sector on a regular basis for 14 months. Although repeated-measures designs are susceptible to learning effects, previous results showed no between-trial differences in ICM or perceived discomfort during repeated baseline tests. (13)

Generalizabilily. The results of this study cannot be generalized to long-term applications because adherence will likely vary across time. The sample size was small, and subjects were screened on the basis of musculoskeletal health. Subjects also represented only one occupational group, directory assistance operators, a group in which musculoskeletal discomfort rates exceed general rates. (36,37) The type of computer used by directory assistance operators also is different from that used by most VDU operators. Compared with data-entry clerks, directory assistance operators use fewer keystrokes per hour and are more likely to be electronically monitored and paced. (37) The directory assistance operators reported difficulties in finding time to perform exercises, a less likely scenario for unpaced VDU operators. (12) Although we believe these results are not generalizable to typical VDU users, we contend that they are applicable to a rapidly expanding sector of the workforce--a sector that includes 3% of Canadian workers. (46)

Conclusion

Based on our results, workstation exercises appear to be an appropriate tool in the prevention of discomfort and in the promotion of ICM. Given the rapid increase in office computerization and sitting-related problems, this is important. However, time constraints should be considered when prescribing workstation exercises for these electronically paced VDU operators.

Appendix.

Exercise Program Revision

Dataspan Exercises: Four years prior to the start of this study, the telephone company had introduced the Dataspan Ergonomic Skills Training Program, (30) which included a series of stretching, range-off movement, and eye-relaxation exercises. In the interim, the adherence rate for the exercises had dropped to nearly zero among the directory assistance operators. Six months prior to the start of the present study, we conducted a series of small-group discussions with 30 directory assistance operators who had been in the Dataspan program. There was consensus among the operators that the exercises made them "feel better." Despite that, they cited several reasons for exercise nonadherence, such as feeling conspicuous or silly, dislike of particular exercises (eg, seated pelvic tilt), difficulty remembering exercises, and having too little time to do the exercises. Of the 11 directory assistance operators in our study, 4 had been employed at the time the Dataspan exercises were originally introduced, but none were doing the Dataspan exercises.

Review of Dataspan Exercises: Examination of the Dataspan exercise program revealed limitations regarding the 3 body regions where directory assistance operators commonly report discomfort (upper neck, lower back, and buttocks). (36) First, the only exercise for the buttock region, the glute clench, would not likely decrease buttock pressure. Second, the Dataspan exercise specific to the upper neck, the cable stretch, encouraged an upright sitting posture, but did not require subjects to stretch the upper neck extensor muscles (eg, suboccipital muscles (a)). Third, the operators found the seated pelvic tilt--an exercise that temporarily changed lumbosacral posture-objectionable, conspicuous, and difficult to do in their chairs, which had built-in lumbosacral supports.

Exercise Revisions: The seated pelvic tilt and glute clench exercises were replaced with an extension-in-standing exercise, (b) which involved standing to decrease buttock pressure and spinal extension to reverse the effects of prolonged flexion on the spine. To complement the Dataspan neck range-of-motion exercises (head turn, head tip), a head-retraction-in-sitting exercise (c) was added to stretch the upper cervical extensors. In the reorganization, the exercises were divided into 5 groups by body region: low back and pelvis, eye, neck, shoulder, and extremities (wrist and ankle). These exercises were printed in pamphlet format and given to each subject for ready access. Except in the low back and pelvis region (extention-in-standing exercise), operators were given 2 or 3 exercises to choose from within each region for each workstation exercise break. Although the literature suggested 3 to 4 repetitions (d) and 15- to 30-second holds, (33,44) time constraints imposed by productivity standards dictated the protocol. Therefore, during the exercise breaks, subjects were asked to perform 2 repetitions of each exercise and to hold each stretch for 5 seconds only.

Implementation: The revised program was introduced on an individual operator basis in 45-minute sessions. At each session, the possible physiological benefits were introduced with 2 general themes: increased relaxation (muscle and eye) and decreased stiffness (joint and muscle). Reduction of musculoskeletal discomfort was not discussed as a benefit. Throughout the introductory session, the emphasis was placed on prevention such that the operators were to perform exercises before they felt stiff or before their eyes felt tired. Since the original Dataspan exercises had been introduced, new workstations had been added that afforded greater privacy. If operators still felt conspicuous, exercise alternatives were offered.

Time pressure concerns: It was more difficult to address the operators' concerns about time pressure, especially when the telephone system was at full capacity and the "call waiting" symbol was posted on-screen. Performing some exercises, such as palming (ie, covering the eyes to promote eye relaxation), required operators to sign off (ie, to withhold calls). This practice put the operators' desire to perform exercises in conflict with company performance standards that were electronically monitored. The first step was to ask subjects, where possible, to perform all exercises except palming and extension in standing while taking calls. Second, we reiterated to the directory assistance operators that the exercise program was initiated by management in full recognition that some of the revised Dataspan exercises required work stoppage.

Table 1.
Summary and Critique of Studies Examining Workstation Exercises (WEX)
for Video Display Unit (VDU) Operators With Respect to Effect on
Discomfort (a)

Study Features

                 Subjects/        Protocol and
Study            Task/Site        Asssessment            Results

Swanson and      37 F/VDU         NIOSH breaks:          Breaks only vs
  Sauter, (18)     operations/      (6, 3 min; 2,          breaks +
  1993             laboratory       10 rain, 32, 15 s)     WEX: NS
                                                           difference
                                                           in discom-
                                                           fort
                                  Protocol:              NS increase in
                                    --1/2 breaks only      discomfort
                                    --1/2 breaks + WEX     over time
                                  Discomfort ratings       (both con-
                                    8 x/d                  ditions)

Henning et       20 F/VDT         Protocol: 3/52         Breaks only vs
  al, (14) 1993    clerks/field     each of:               breaks +
                                    --baseline             WEX: NS
                                    --NIOSH breaks only    decrease in
                                    --NIOSH breaks +       leg discom-
                                     WEX                   fort
                                  Discomfort ratings
                                    3 x/d

Thompson,        85 M, F/data     Stretching 5 min,      Comfort was
  (19) 1990        entry/field      twice daily            improved
                                  Exit survey              (NS)

Lee and          4 F, 1 M/data    Light WEX 4 min,       NS increase in
  Oh, (20) 1985    and file         3 x/d                  arm dis-
                   entry/         Protocol: WEX vs no      comfort with
                   laboratory       WEX                    no WEX
                                  Subjective
                                    discomfort ratings

Klemetti et      TNP: 74 treated  Neck school 8x/mo      TNP-tx exer-
  al, (16) 1997    (TNP-tx), 77     (WEX training,         cised more
                   control          self-care, phy-        often
                   (TNP-C)/field    sical workouts)        (P=.04)
                                  Physical therapy       TNP-tx de-
                                    treatment (TNP-tx)     creased neck
                                  No treatment (TNP-C)     discomfort
                                  Questionnaire 6/12       (NS)
                                    postintervention

Saltzman,        25 F, 19 M/VDU   Software-based WEX,    Data entry
  (17) 1998        operations       3 exercises every      rated dis-
                   (data entry      30 min                 comfort
                   only vs mixed  Study duration:          Jower than
                   tasks)/field     15/52                  mixed task
                                  Online visual analog     group (no
                                    discomfort scale       significance
                                                           tests)

Henning et al,   Large site:      3 groups:              Large site: NS
  (15) 1997        65 F, 8 M;       --control              difference
                   small site:      --breaks only          in discom-
                   15 F, 4 M/VDU    --breaks + WEX         fort among 3
                   operations/    Discomfort ratings       test condi-
                   field            3 x/d                  tions
                   Note: "large"                         Small site:
                   and "small"                             leg/foot
                   were not                                discomfort
                   defined                                 decreased
                                                           with WEX as
                                                           trend (P<.l)

                 Critique

                                WEX Based on:
                 Assessment:    1. Perceived
                 1. Discomfort  Discomfort      Considered:
                 vs Condition   2. Exercise     1. Adherence
                 2. Discomfort  Specialist      2. In-Chair
Study            vs Time        Opinion         Movement

Swanson and      1. No          1. No           1. No
  Sauter, (18)   2. Yes         2. Yes          2. No
  1993

Henning et       1. No          1. No           1. No
  al, (14) 1993  2. No          2. Yes          2. No

Thompson,        1. No          1. No           1. No
  (19) 1990      2. No          2. Yes          2. No

Lee and          1. Yes         1. No           !. No
  Oh, (20) 1985  2. No          2. Yes          2. No

Klemetti et      1. Yes         1. Yes          1. Yes
  al, (16) 1997  2. No          2. Yes          2. No

Saltzman,        1. No          1. No           1. Yes
  (17) 1998      2. No          2. Yes          2. No

Henning et al,   1. Yes         1. No           1. Yes
  (15) 1997      2. No          2. Yes          2. No

(a) NIOSH=National Institute of Occupational Health and Safety,
F=female, M=male, NS=nonsignificant, VDT=video display terminal,
TNP=patients with "tension neck."
Table 2.
Exercise Satisfaction Questionnaire Results Reported by 11 Directory
Assistance Operators Regarding the Revised Dataspan Workstation
Exercises

                                              Percentage
Factors on Exercise Satisfaction             (Number) of
Questionnaire                             Subjects Reporting

Did not feel conspicuous performing
  exercises                                     90 (10)
Exercises were easy to perform                  90 (10)
Satisfied with the number and variety
  of exercises                                 100 (11)
Reported minor discomfort
  performing:
  Wrist flexion                                 18 (2)
  Extension in standing                         10 (1)
Difficulty rememberring to do
  exercises                                     55 (6)
Difficulty finding time to do exercises         64 (7)

(a) "Rocabado M. Biomechanical relationship of the cranial, cervical, and hyoid regions. J Craniomandib Pract. 1983;1:62-66.

(b) McKenzie RA. Treat Your Own Back. Waikanae, New Zealand: Spinal Publications Ltd; 1985.

(c) McKenzie RA. Treat Your Own Neck. Waikanae, New Zealand: Spinal Publications Ltd; 1983.

(d) Taylor DC, Dalton JD, Seaber AV, Garret WE. Viscoelastic properties of muscle-tendon units: the biomechanical effects of stretching. Am J Sports Med. 1990; 18:300-309.

* We conducted our review of the literature using the following databases: MEDLINE, CINAHI, NIOSHTIC, SPORT DISCUS, EI COMENDEX, Ergonomics Abstracts, and online using SavvySearch.com. We used the following key words in our literature search: exercises, physical therapy, posture, sitting, stretching, workplace, workstation, and VDU.

[dagger] Sample size was calculated using the standard deviation and smallest detectable difference for the dependent variable with the greatest variability (ie, ICM).

[double dagger] Actual power based on the sample (N=11) was .87 for ICM measurements and .91 for BPD scores.

[section] VISTA Medical, 120 Maryland St, Winnipeg, Manitoba, Canada R3G 1L1.

[parallel] SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

References

(1) Grieco A. Sitting posture: an old problem and a new one. Ergonomics. 1986;29:345-362.

(2) Hettinger T. Statistics on diseases in the Federal Republic of Germany with particular reference to diseases of the skeletal system. Ergonomics. 1985;28:17-20.

(3) Sauter SL. Predictors of strain in VDT users and traditional office workers. In: Grandjean E, ed. Ergonomics and Health in Modern Offices. London, United Kingdom: Taylor & Francis; 1984:129-135.

(4) Teiger C. Regulation of activity: an analytical tool for studying workload in perceptual motor tasks. Ergonomics. 1978;21:201--213.

(5) Branton P. The Comfort of Easy Chairs. Technical Report #22. Hertfordshire, United Kingdom: Furniture Industry Research Association; 1966:82.

(6) Grandjean E, Jenni M, Rhiner A. Eine indirekte methods zur erfassung des komfortgefuhls beim sitzen. Int Z Physiol Einschl Arbeitsphysiol. 1960;18:101-106.

(7) Bhatnager V, Drury C, Schiro S. Posture, postural discomfort and performance. Hum Factors. 1985;27:189-199.

(8) Occhipiniti E, Colombini D, Grieco A. A procedure for the formulation of synthetic risk indices in the assessment of fixed working postures. In: Queinnec Y, Daniellou F, eds. Designing for Everyone: Proceedings of the 11th Congress of the International Ergonomics Association. London, United Kingdom: Taylor & Francis; 1991:3-5.

(9) Winkel J. On the significance of physical activity in sedentary work. In: Knave B, Wideback PG, eds. Work With Display Units '86. Amsterdam, the Netherlands: Elsevier; 1987:229-236.

(10) Fenety A, Putnam C, Walker JM. Measuring in-chair movement: validity, reliability and implications for measuring discomfort. Appl Ergon. 2000;31:383-393.

(11) Hoekstra E, Hurrell J, Swanson N. Evaluation of work-related musculoskeletal disorders and job stress among teleservice centre representatives. Appl Occup Environ Hyg. 1995;10:812-817.

(12) Silverstein BA, Armstrong TJ, Longmate A, Woody D. Can in-plant exercise control musculoskeletal symptoms? Occup Med. 1988;30: 922-927.

(13) Fenety PA. Factors Influencing Sitting Comfort and In-chair Movement in the Electronic Office [microform thesis]. Ottawa, Ontario, Canada: National Library of Canada; 1996.

(14) Henning RA, Alteras-Webb SM, Jacques P, et al. Frequent, short rest breaks during computer work: the effects on productivity and well being in a field study. In: Luczak H, Cakir A, Cakir G, eds. Work With Display Units '92. Amsterdam, the Netherlands: Elsevier Press; 1993: 292-295.

(15) Henning RA, Jacques P, Kissel GV, et al. Frequent short rest breaks from computer work: effects on productivity and well-being at two field sites. Ergonomics. 1997;40:78-91.

(16) Klemetti M, Santavirta N, Sarvimaki A, Bjorvell H. Tension neck and evaluation of a physical training course among office workers in a bank corporation. J Adv Nurs. 1997;26:962-967.

(17) Saltzman A. User perception of the effectiveness of exercise mini-breaks. In: Proceedings of the Silicon Valley Ergonomics Conference and Exposition-ErgoCon '98. San Jose, Calif: San Jose State University; 1998:147-151.

(18) Swanson NG, Sauter SL. The effects of exercise on the health and performance of data entry operators. In: Luczak H, Cakir A, Cakir G, eds. Work With Display Units '92. Amsterdam, the Netherlands: Elsevier; 1993:288-291.

(19) Thompson DA. Effect of exercise breaks on musculoskeletal strain among data entry operators: a case study. In: Sauter SL, Dainoff MJ, Smith MJ, eds. Promoting Health and Productivity in the Computerized Office: Models of Successful Interventions. London, United Kingdom: Taylor & Francis; 1990:118-127.

(20) Lee KS, Oh YG. An alternative to reduce the physical stress of VDT operators. In: Proceedings of the Human Factors Society 29th Annual Meeting. Santa Monica, Calif: The Human Factors Society; 1985: 932-936.

(21) Corlett EN. The evaluation of industrial seating. In: Wilson JR, Corlett EN, eds. Evaluation of Human Work: A Practical Ergonomics Methodology. London, United Kingdom: Taylor & Francis; 1990: 500-515.

(22) Sahrmann SA. The Twenty-Ninth Mary McMillan Lecture: Moving precisely? Or taking the path of least resistance? Phys Ther. 1998;78: 1208-1218.

(23) vanDeursen DL, Lengsfeld M, Snijders CJ, et al. Mechanical effects of continuous passive motion on the lumbar spine in seating. J Biomech. 2000;33:695-699.

(24) Stranden E. Dynamic leg volume changes when sitting in a locked and free floating tilt office chair. Ergonomics. 2000;43:421-433.

(25) Kuorinka I, Jonsson B, Kilbom A, et al. Standardized Nordic Questionnaire for the analysis of musculoskeletal symptoms. Appl Ergon. 1987; 18:233-237.

(26) van der Grinten MP, Smitt P. Test-retest reliability of a practical method for measuring body part discomfort. In: Queinnec Y, Daniellou F, eds. Designing for Everyone: Proceedings of the 11th Congress of the International Ergonomics Association. London, United Kingdom: Taylor & Francis: 1991:54-56.

(27) Corlett EN, Bishop RP. A technique for assessing postural discomfort. Ergonomics. 1976;19:175-182.

(28) Stratford PW. Reliability: consistency or differentiating between subjects. Phys Ther. 1989;69:299-300.

(29) Al-Eisa E, Fenety A, Egan D, Crouse J. Measurement of sitting pressure distribution under the ischium: a reliability study. In: Winters J, ed. Proceedings of the RESNA 2000 Annual Conference. Orlando, Fla: RESNA Press; 2000:357-359.

(30) Joyce M, Peterson M. Dataspan Ergonomic Skills Training Handbook. Seattle, Wash: The Joyce Institute: 1985:5-22.

(31) Lee KS, Swanson N, Sauter S, et al. A review of physical exercises recommended for VDT operators. Appl Ergon. 1992;23:387-408.

(32) Schleifer LM, Amick BC. System response time and method of pay: stress effects in computer-based tasks, Int J Human-Computer Interaction. 1989;1:23-39.

(33) Shrier I, Gossal K. Myths and truths about stretching. Phys Sports Med. 2000;28:57-63.

(34) Zar JH. Biostatistical Analysis. 4th rev ed. Upper Saddle River, NJ: Prentice Hall; 1999.

(35) Zhang L, Helander M. Identifying factors of comfort and discomfort: a multidimensional approach. In: Kumar S, ed. Advances in Industrial Ergonomics and Safety IV. London, United Kingdom: Taylor & Francis; 1992:395-402.

(36) LeGrande DE. VDT repetitive motion health concerns in the US telecommunications industry. In: Smith MJ, Salvendy G, eds. Human-Computer Interaction: Applications and Case Studies. Proceedings of the 5th International Conference on Human-Computer Interaction. Amsterdam, the Netherlands: Elsevier; 1993:780-785.

(37) Smith MJ, Carayon P, Sander's KJ, et al. Employee stress and health complaints in jobs with and without electronic performance monitoring. Appl Ergon. 1992;23:117-127.

(38) Kuorinka I, Forcier L. Evidence for work relatedness for selected musculoskeletal disorders of the neck and limbs. In: Kuorinka I, Forcier L, eds. Work Related Musculoskeletal Disorder: A Reference Book for Prevention. London, United Kingdom: Taylor & Francis; 1995:17-137.

(39) Waersted M, Westgaard RH. Working hours as a risk factor in the development of musculoskeletal complaints. Ergonomics. 1991;34: 265-276.

(40) Edwards RHT. Hypotheses of peripheral and central mechanisms underlying occupational muscle pain and injury. Eur J Appl Physiol. 1988;57:275-281.

(41) Ferreira M, de Souza Conceicao G, Saldiva P. Work organization is significantly associated with upper extremities musculoskeletal disorders among employees engaged in interactive computer-telephone tasks of an international bank subsidiary in Sao Paulo, Brazil. Am J Indust Med. 1997;31:468-473.

(42) Zwahlen H, Hartmann AL, Rangarajulu SL. Effects of rest breaks in continuous VDT work on visual and musculoskeletal comfort/discomfort and on performance. In: Salvendy G, ed. Human-Computer Interaction: Proceedings of the first USA-Japan Conference on Human-Computer Interaction. Amsterdam, the Netherlands: Elsevier; 1984:315-319.

(43) Henning RA, Sauter SL, Salvendy G, Krieg EF. Effects of micro-breaks on performance and well-being in data entry work. In: Work With Computers: Organizational, Management, Stress and Health Aspects. Amsterdam, the Netherlands: Elsevier; 1989:377-380.

(44) Bandy WD, Irion JM. The effect of time on static stretch on the flexibility of the hamstring muscles. Phys Ther. 1994;74:845-850.

(45) Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. East Norwalk, Conn: Appleton & Lange; 1993:141.

(46) Nova Scotia Department of Finance. Nova Scotia Statistical Review 2000-Labour Force. Available at: www.gov.ns.ca/finance. Accessed: October 2001.

A Fenety, PT, PhD, is Assistant Professor and Undergraduate Coordinator, School of Physiotherapy, Dalhousie University, Forrest Bldg, 5869 University Ave, Halifax, Nova Scotia, Canada B3H 3J5 ([e-mail]). Address all correspondence to Dr Fenety.

JM Walker, PT, PhD, is Professor (Retired), School of Physiotherapy, Dalhousie University.

Both authors provided concept/research design and writing. Dr Fenety provided data collection and analysis, subjects, and project management. Dr Walker provided fund procurement, facilities/equipment, institutional liaisons, and consultation (including review of manuscript before submission).

This study was approved by the Human Ethics Review Committee of Dalhousie University's Faculty of Graduate Studies.

This study was funded, in part, by a Postgraduate Scholarship to Dr Fenety from the Natural Sciences and Engineering Research Council of Canada.

This study was presented, in part, at the 12th International Congress of the World Confederation for Physical Therapy; June 25-30, 1995; Washington, DC.

This article was submitted January 9, 2001, and was accepted January 5, 2002.