Measuring Adjustability's Benefits
Being able and encouraged to change work posture more often can result in less fatigue, a two-year ergonomic study found.
THE emphasis on designing workplaces to accommodate individuals having a variety of physical differences has increased due to the knowledge and public awareness surrounding ergonomics. As both manufacturers and consumers understand, work environments that are more comfortable to use and adjustable to individual needs can reduce discomfort, the potential for fatigue, and cumulative trauma, as well as increase job productivity.
Accessible Designs/Adjustable Systems, Inc. (AD/AS) of Boise, Idaho, requested the ergonomics consulting services of The Ohio State University's Institute for Ergonomics to assess one aspect of the adjustability issue in a laboratory environment, using standardized quantitative analysis methods.
The goal of this study was to record fatigue and discomfort measurements of subjects doing the same physical work using a variety of work table configurations. The experiment's hypothesis was that using height-adjustable laboratory benches, or using the benches in conjunction with other accommodations (i.e., stools, chairs, footrests), would produce a measurable decrease in reported body discomfort, perceived exertion, or measured muscle fatigue, compared with a non-adjustable table.
Twelve healthy subjects (six male and six female) were tested for this study. Their characteristics are shown in Table 1. All were right-handed. Every attempt was made to recruit individuals so the data set represented the middle 90 percent range of standing heights for males and females found in a typical population (Kroemer 1997). These target standing heights were:
The sample collected for this experiment closely matched these criteria.
The standard and adjustable tables were provided by AD/AS, as were the sit-stand stool and other equipment used for the testing task. The adjustable chair was provided by the Institute for Ergonomics. Muscle fatigue was measured in two ways. The first was using a surface electromyography (EMG) system, provided by the institute, on the left and right erector spinae muscles. These muscles are located on either side of the lumbar spine, near the skin surface. A downward shift in the frequency spectrum of a muscle over time indicates the muscle is fatiguing. The second method was the use of a Near Infrared Spectroscopy (NIRS) system, also provided by the institute, on the trapezius muscle. This system uses light to measure blood volume and blood oxygenation of a muscle, which is another indicator of muscle fatigue.
Body part discomfort for 10 areas was measured using a one-to-seven scale, where one indicated "no discomfort" and seven indicated a high discomfort level, or "the worst pain you can think of." Whole-body perceived physical exertion (Borg 1990) was measured on a 6-20 scale.
The experimental task performed by subjects was designed to simulate the repetitive and precise activities of laboratory technicians as required in their typical jobs. Subjects were asked to transfer granular table salt from larger (two-quart) containers to six smaller beakers using a ladle. The amount of salt to be placed in each beaker varied. The positions of these containers were marked on the work tables; subjects were allowed to move them within these areas, but not outside of them, during the task.
The independent variable measured in this study was table working condition. Three conditions were studied:
Condition 1--Performing the experimental task at a standard height table whose working surface was at 36" from the floor.
Condition 2--Performing the task at a table whose height was adjustable from 27" to 38" from the floor.
Condition 3--Doing the work using the adjustable table but also having available use of the sit-stand stool, the adjustable chair, and a footrest.
The dependent variables collected were fatigue of the erector spinae and trapezius muscles; subjective ratings of body part discomfort; and the perceived exertion required of the task condition. Subjects arriving for each testing condition first were asked to give their informed, written consent. They were asked whether they were experiencing physical discomfort to any of their body parts and, if so, this was recorded on the data collection sheet. Then they were shown the experimental set-up and explained the requirements of the task. For the sessions in which the height-adjustable table was used (Condition 2) or when it was used in conjunction with work accommodations (Condition 3), each of these was explained and their adjustability features were demonstrated. Subjects also were asked to try out each accommodation, to become familiar with how the task would be performed using that accommodation, if they so desired.
The rate at which subjects worked was left to their discretion. However, they were asked to view the task as if it were their daily job and that their wage was determined by how many cycles of beaker-filling they could complete on an hourly basis. Their trade-off was that the amount of salt spillage would result in a reduction in their "wage."
Once the briefing was completed, subjects were fitted with electrodes (used to measure EMG activity) on their erector spinae muscles and the NIRS on their right trapezius muscle. A baseline muscle activity reading was collected as subjects stood erect, with their backs against a vertical rack, holding steady a weight at arms' length for five seconds. The weight was 25 pounds for males and 15 pounds for females. Following this procedure, the subject was asked to walk to the table used for that condition and asked to begin working.
Subjects worked for a one-hour time period. Every 15 minutes, they were asked to rate their discomfort for each body part and their perceived exertion required of that particular condition. For the two conditions in which adjustments could be made, the time was recorded when subjects changed the table height, the height of the chair or sit/stand, or their position (e.g., from sitting to standing). At the end of the 60-minute testing period, subjects were immediately positioned against the vertical rack, and the weights were placed on their wrists so that a final reading could be taken of their erector spinae muscle activity. This was performed to determine whether there was a shift in the frequency spectrum for this muscle. A downward shift in the median frequency is an indication of muscle fatigue.
Subjects were tested three times, once for each experimental condition.
To reduce the potential for fatigue, only one testing session was conducted per day for that subject. Also, to reduce bias effects, the order in which subjects performed the task in each of these three conditions was randomized. Each subject was compensated $35 for participating in the study.
Muscle activity data gathered from the EMG and NIRS data collection systems were analyzed using standard, commonly used software programs. All data (subject information, discomfort and perceived exertion scores, and EMG and NIRS data) then were put into one database. The quality of the data was checked to ensure its accuracy and meaningfulness.
Descriptive statistics were computed on subjects' age, height, weight, and chosen task cycle rate. T-tests were used to determine whether significant differences existed in these variables between the males and females tested.
Other statistical tests (analysis of variance) were used to determine whether the discomfort scores and ratings of perceived exertion differed significantly across the 15-minute intervals. Similar tests were used to determine whether the median frequency values (from the EMG data on the erector
spinae muscles) differed at the end of the one-hour work period compared with that taken before the experiment began. The data were analyzed as a function of the working condition of the task (Condition 1, Condition 2, and Condition 3), the frequency with which available adjustability options were utilized, and the height of the subject.
Subject differences--Statistical testing found these groups of males and females did not differ significantly in terms of their age or the number of cycles chosen to perform the experimental task. However, as expected, males were found to weigh significantly more and were significantly taller than their female counterparts.
Evidence of fatigue--In developing the experimental protocol, it was desired that a representative task be used that also presented the potential for fatigue over the one-hour testing period. The NIRS data were primarily used as an objective, qualitative indication if this were indeed the case. In roughly 70 percent of experimental runs across all conditions, subjects exhibited some upper body/shoulder fatigue. In analyzing these data by condition, the results show that all of the Condition 1 tests produced some upper body/shoulder fatigue, while only 63 percent of the tests for Condition 2 and Condition 3 produced some upper body/shoulder fatigue in subjects.
Data variability--It was found from these data that considerable variability existed in the dependent measures studied. Subjects also varied considerably in how often they changed their work area when given the opportunity (i.e., Conditions 2 and 3). Given the relatively small sample size of 12 subjects having a considerable range in anthropometric dimensions, many of the statistical analyses were found to be not significant. However, several general trends were observed in the data, and these are highlighted below.
Erector Spinae Differences. These data were analyzed to determine whether there were low back fatigue effects exhibited in the erector spinae muscles for subjects across the three experimental conditions. For the right erector spinae muscle, more subjects had these frequency shifts in Conditions 2 and 3 than was found in Condition 1. However, for the left erector spinae, the opposite trend was observed; more subjects experienced these levels of muscle fatigue when doing work in Condition 1. This suggests each side of the erector spinae muscle is "working" or being activated in different ways as subjects are doing work either sitting or standing at these tables.
Because all subjects were right-handed and using the ladle in this hand, biomechanical principles suggest muscles on the left side of the body (including the erector spinae) would increase their activation to maintain a balance for body. That fewer subjects had left erector spinae fatigue in Conditions 2 and 3 suggests a possible improvement in using these workstation designs and allowing for adjustability. A possible explanation for why the frequency shift for the right erector spinae muscles was lower in Condition 1 is because subjects who had no opportunity in this situation for workplace adjustments may have shifted their stance more frequently when doing work, thus easing the demands on this muscle.
Because there were no restrictions on how often and to what height subjects could adjust their work table (Condition 2) or use other accommodations (Condition 3), there was considerable variability across subjects. That is, some individuals made adjustments many times over their work hour, and some made none. Thus, the data were grouped into three categories: None/Condition 1, indicating no adjustments were made (16 of the 36 testing hours); Few, indicating only one or two changes were made over the course of the testing hour (13 of the 36 tests); and Many, indicating three or more adjustments were made (7 of the 36 tests).
For both the right and left erector spinae muscles, those who made no work adjustments or just a few adjustments produced positive median frequency shifts, indicating some fatigue was occurring in these muscles, compared with those who made many adjustments. In this latter group, the numbers were actually negative. This indicates the median frequency was actually higher at the end of the hour's testing than at the pre-work, baseline reading. It should be noted the median frequency shifts were less than 7 percent, and the means of these data were not statistically significant at a
=0.05 (because of the variability among subjects). However, these trends suggest that making more accommodations to one's work table, through changes in table height or the use of sit-stand stools and chairs, may reduce fatigue in those performing this type of work.
Body Discomfort Differences. Subjects rated their levels of physical discomfort at baseline (before each testing session began) and at 15-minute intervals during the hour of testing. The scale ranged from 1 (no discomfort) to 7 (maximum discomfort). Generally, the scores reported were low, indicating this work task did not cause even moderate levels of discomfort during the testing sessions, regardless of the condition under which the work was performed. There were indications that discomfort did increase over the hour of testing, though. For each of the three conditions, summed discomfort increased, though nowhere near the maximum amount of discomfort attainable. This was not surprising given the nature of the task. Steady (though small) increases in summed discomfort were observed for Conditions 1 and 2. This was not the case for Condition 3, in which summed discomfort decreased slightly or leveled off after 30 minutes.
There were, however, some statistically significant discomfort differences (at the a
=0.05 level) found across the 10 body areas surveyed, and these differences are discussed below.
Mean shoulder discomfort scores were significantly different between the three conditions only at the 15- and 30-minute testing intervals. At the 15-minute time period, the Condition 3 score (mean of 1.83) was found to be significantly higher than either of the mean scores for Condition 1 or Condition 2. At the 30-minute period, the mean Condition 3 score (2.42) was significantly higher than that mean for Condition 1.
There are several hypotheses for why these results were found. First, several subjects opted to sit during the Condition 3 task. As suggested by these discomfort results, this scooping/measuring task may have been overloading the shoulder joint, resulting in increased discomfort scores. It may not have been an appropriate task for seated technicians. Second, there were no differences at the 45- and 60-minute testing periods, indicating subjects may have adjusted their work tables or their postures to make themselves more comfortable. Third, all of these scores are low on the one-to-seven scale, suggesting these statistical differences may have little practical significance.
At the 30-minute testing period, the mean hand/wrist discomfort score for Condition 3 was significantly greater than that found for Condition 2. The possible explanations here are similar to those posed for the shoulder joint: more stress on the hand/wrist while sitting to perform this task, no significant difference between testing conditions later in the hour period, and low levels of discomfort on the scale used.
Mean discomfort scores for this body region were significantly higher for Condition 2, toward the end of the testing session (at minute 45 and minute 60). After 45 minutes, Condition 2's mean score of 1.33 was significantly higher than that found for Condition 3 only (at 1.00).
At the end of the hour (minute 60), Condition 2's mean score (1.42) was significantly larger than those means recorded for both Condition 1 and Condition 3 (1.08 and 1.00, respectively). It was not surprising Condition 3's mean scores were lower because sitting was allowed. The reason a difference was found between Conditions 1 and 2 at the end of the hour is unclear because both conditions required standing.
After the 30-, 45-, and 60-minute work times, the amount of mean discomfort to the subjects' ankles/feet was significantly higher under Condition 2. Discomfort scores were similar between Conditions 1 and 3. These results were counter to what was expected. However, the reason for this finding may involve the design of the two work tables. The table used in Condition 1 had a cabinet underneath, while there was no such obstruction with the table used for Condition 2. The Condition 1 cabinet may have caused subjects to move more, changing their posture and resulting in less fatigue to these body areas. It should be noted that similar results to these were found within the thigh/knee region, as well.
It also should be noted there were no statistically significant differences found between the three experimental testing conditions for discomfort to the neck, elbow/forearm, fingers, upper back, low back, or low leg. There were no differences found across the perceived exertion scores given by subjects, either between conditions or grouped by the amount of table/accommodation adjustments done.
Differences Due to Subject's Standing Height
The EMG and discomfort data also were analyzed as a function of the standing height of the subject. Individuals were grouped into two categories--Tall (5' 8" or more) and Short (less than 5' 8"). The results found there to be no differences in the frequency shifts of the left or right erector spinae muscles. However, there were significant differences in the discomfort scores at the a
=0.05 level. For some body areas, Tall subjects reported higher levels of discomfort, regardless of the condition in which the work was performed. Tall subjects reported significantly higher mean levels of shoulder discomfort (p=0.031) and ankle/feet discomfort (p=0.036), and their mean neck discomfort was marginally significant (p=0.066). Tall subjects reported higher discomfort scores at the less restrictive a
=0.10 level (p=0.104).
The objective of this study was to measure various responses to individuals working at a non-adjustable and at a height-adjustable table (the latter both with and without workplace accommodations such as stools, chairs, and footrests) and to determine whether differences existed in muscle fatigue or subjective reports of body discomfort. The hypothesis was that the more-adjustable conditions would result in lower levels of muscle fatigue and reported body part discomfort.
The results reported here partially supported this hypothesis. The more objectively recorded data, which assessed shoulder (trapezius muscle) and low back (erector spinae muscle) fatigue, found trends indicating more fatigue occurred in subjects under Condition 1. More specifically, a greater percentage of subjects showed trapezius muscle fatigue using the standard work table (Condition 1) than did those performing the same work with the height-adjustable table (Condition 2) or this same table with other accommodation (Condition 3). In the low back, a larger number of subjects exhibited a downward shift in the median frequency of the left erector spinae muscle (an indication of muscle fatigue) for the adjustable conditions, compared with Condition 1. However, the reverse was true for the right erector spinae muscle, indicating each of these muscles was being activating differently during this task.
The hypothesis was further supported by the fact that subjects who changed their workstation more often, thus making more changes to their posture while doing their task, had, on average, less erector spinae muscle fatigue than those who made fewer workstation/posture adjustments. This finding suggests that one's ability to change work posture more often, or the encouragement to make such changes, can result in less worker fatigue.
Discomfort scores, resulting from subjective feedback, did not clearly support the stated hypothesis. There is at least one explanation for why these findings did not coincide with the NIRS and EMG results. The task, though able to produce some fatigue, may not have been strenuous enough to produce differences that subjects could detect physically. The absolute differences were very small across the discomfort scale. To determine greater differences in discomfort, if they exist between these conditions, either a more physically demanding task would need to be studied or this same task would need to be observed under a longer time period, such as across an eight-hour work day.
There were several indications the task designed for this experiment did produce fatigue. However, the levels of this fatigue were not dramatic. It was not the intent of this study to produce a task so tiring that these effects would translate to measurable readings. Instead, the task was designed to replicate work tasks commonly performed with these tables.
Many of the results reported here showed trends in the data and were not all statistically significant at the a
=0.05 level (the level commonly used by researchers to determine whether differences exist between experimental conditions). This is believed the result of several factors. First, there was a large amount of inherent variability in the subject population. Both males and females were tested, and their body sizes (in terms of standing heights) varied across a wide range. The subjects also were free to work at their preferred pace and could make as many or as few workplace changes as desired (as in Conditions 2 and 3). This approach was preferred to a study design limited to just one gender, which forced subjects to work at a pre-determined pace and required them to make changes in a manner they may not have found to be desirable.
Although the results presented here do not have as much statistical power as desired, the ability to generalize these results to a larger population of individuals who may work with these tables is much larger and more relevant.
1. Borg G, (1990), Psychophysical scaling with applications in physical work and the perception of exertion, Scandinavian Journal of Work, Environment & Health, 16 (supplement 1):55-58.
2. Kroemer KHE, (1997), Engineering anthropometry, in Salvendy G, ed., Handbook of Human Factors and Ergonomics, New York: John Wiley & Sons, Inc., 219-232.
This article originally appeared in the April 2004 issue of Occupational Health & Safety.