The ability to lie down while operating a computer can increase the amount of time that a person with low back pain is able to work. There have been many solutions proposed and accommodations made for computer operators who have difficulty working from an upright posture. However, there has been very little written to describe the impact of these postures on the performance and comfort of the computer operator. This study measured the impact of five different postures on typing performance and user comfort. Although the postures were significantly different, ranging from upright to fully supine, the typing performance did not vary significantly between postures. However, there was a significant difference in some measures of body part discomfort and overall comfort.
Low Back Pain, typing performance, comfort, supine computer workstation
Many occupations today require workers to be at a computer for several hours at a time. Traditional computer workstations are designed for a predominantly upright posture. Issues with low back pain (LBP) often prompt people to seek alternative postures when working on the computer. Some with chronic LBP will take periodic breaks from their workstation and lie flat on the floor to relieve their discomfort (1) . Systems have been developed to allow people to access their computer from a reclined or supine posture (2) . However, very little has been published regarding the impact of supine or significantly reclined postures on the performance of computer-related tasks. One paper by Monroe (3) looked at the impact of a reclined posture and split keyboard on comfort and typing performance. Monroe found that there was no significant difference in typing performance with the straight keyboard when moving from an upright to a reclined posture. However, the test did not look at more extreme postures such as lying fully supine or sitting with a significant seat pan tilt. Therefore, the purpose of this study was to determine what effect these alternative postures have on typing performance and comfort.
The test was designed to be a single-factor repeated measures test in which a series of typing tests would be completed in each of five different randomly assigned working postures. Dependent variables included typing speed (gross words per minute - GWPM), typing accuracy, body part discomfort (BPD) rating, and an overall comfort (OC) rating. The research null hypotheses for this study were as follows:
Attempts were made to identify participants who experience chronic, recurring LBP or discomfort. All subjects were required to average a minimum of 10 hours per week working on a computer and to be touch typists.
A Position Dynamics Optima tilt/recline wheelchair seating system was mounted to a fixed base. A keyboard tray was attached to the armrests in order to maintain keyboard position relative to the user during the tilt and recline motions. Vertical pads were added to each of the armrests to provide support to the upper arms in more severely reclined or tilted positions. The headrest was modified to increase vertical adjustability as well as anterior/posterior adjustability. A separate moveable frame was constructed to allow a flat-screen computer monitor to be positioned over the head of the user whether in an upright or a supine position. The monitor adjustment allowed for the angle to be varied between 90 degrees (upright) and 0 degrees (face down). For each posture, the monitor was adjusted to maintain a consistent view angle and view distance relative to the participant.
Participants completed three five-minute typing trials and three BPD assessments in each of five postures. Subjects also rated their OC for each posture, using a modified Shackel comfort scale (4) , which ranks OC on a scale from 0 ("I feel completely relaxed") to 10 ("I feel unbearable pain"). The BPD was rated using a Borg pain scale (5) , which ranks the level of discomfort on a scale from 0 ("nothing at all") to 11 ("Maximal"). The postures provided a range of adjustment from upright seating to fully supine positioning. Table 2 shows the seat angle, back angle, and seat-to-back angle for each of the postures.
| Posture | Seat Angle (measured from front horizontal plane) | Back Angle (measured from rear horizontal plane) | Seat to Back Angle |
|---|---|---|---|
| Upright | 5 |
75 |
100 |
| Tilted | 40 |
40 |
100 |
| Reclined | 5 |
50 |
125 |
| Supine #1 | 40 |
15 |
125 |
| Supine #2 | 5 |
15 |
160 |
The typing tests were completed using the Typing Master Pro software version 6.30 [2005] (Demo version). The subjects were encouraged to strive for accuracy without worrying about speed. The randomly assigned texts for the typing tests were selected from a collection of Grimm's Fairy Tales. The readability of the text was determined by using Microsoft Word 2003 spelling and grammar check to calculate the Flesch-Kincaid Grade Levels. The grade levels ranged between 4.7 and 6.8 with an average Flesch-Kincaid Grade Level of 5.8.
There were 13 subjects recruited for this test. Table 1 shows the breakdown of the ages between those who self-identified as having "chronic, recurring episodes of LBP or discomfort" and those who did not.
| n | Average | Range | |
|---|---|---|---|
| LBP | 7 |
29.1 |
19-51 |
| No LBP | 6 |
23.7 |
18-29 |
| TOTAL | 13 |
26.6 |
18-51 |
There was no difference in GWPM (mean = 50.3, SD = 19.17) or accuracy (mean = 96.17, SD = 2.66) between participants who self-identified as having problems with LBP and those who did not. A within-subjects repeated measures analysis examining differences in means between the different postures was run using Huynh-Feldt correction. This analysis found no statistically significant effects on GWPM F(4,43)=2.517 or accuracy F(4,42)=1.218, p>0.05. However, there were significant increases in GWPM F(2,22)=8.097, p<0.01 based on trial number for the different postures.
As a first level analysis for the BPD data, calculations were made to determine the BPD frequency (BPDF), the BPD severity (BPDS), and the product of those two values (BPDFS). The BPDF represents the number of non-zero entries for all of the body parts, for a given trial. The BPDS represents the average of the non-zero entries for a given trial. The BPDFS represents the sum of non-zero entries for a given trial. These measures are consistent with previous work conducted by Liao and Drury (6) . Using the same within-subjects repeated measures analysis with Huynh-Feldt correction, means between the different postures revealed statistically significant effects on BPDF F(3,23)=5.489, p<0.01; BPDFS F(3,27)=5.943, p<0.01; and OC F(4,43)=5.240, p<0.01. Trial number significantly impacted BPDF F(1,10)=6.663, p<0.05. A summary of the F values for relevant tests is shown in table 3.
| Variable | Posture | Trial |
|---|---|---|
| BPD - Frequency (BPDF) | 5.489** | 6.663* |
| BPD - Severity (BPDS) | 2.023 | 1.675 |
| BPD - Sum (BPDFS) | 5.943** | 5.143 |
| Overall Comfort (OC) | 5.240** | -- |
| Typing Speed (GWPM) | 2.517 | 8.097** |
| Typing Accuracy | 1.218 | 2.216 |
| * indicates statistical significance p<0.05 with Huynh-Feldt correction | ||
| ** indicates statistical significance p<0.01 with Huynh-Feldt correction | ||
The lack of significant differences in typing performance based on posture supports the first null hypothesis that changes in posture do not impact typing performance, even with the more extreme positions. This suggests that working from a reclined or supine position will not reduce typing productivity. The impact on typing speed due to trial number indicates that the participants' typing improves as they become accustomed to the new posture. Further testing should be conducted to investigate this issue. The statistically significant impact of posture on BPDF, BPDFS, and OC supports the rejection the second null hypothesis, which states that changes in posture do not impact comfort. Therefore, further efforts will be made to investigate the impact of these postures on specific body parts and to improve the test fixture to maximize comfort for participants with and without LBP.
This project was funded through the RERC on Workplace Accommodations, which is supported by Grant H133E020720 of the National Institute on Disability and Rehabilitation Research of the U.S. Department of Education. The authors also wish to acknowledge Professor D.L. Ashliman for his efforts in making Grimm's Fairy Tales available in electronic format from the following website: www.pitt.edu/~dash/grimmtales.html
Scott Haynes, MBME
490 10 th Street, NW,
Atlanta, GA 30332
Phone: (404) 894-9156,
FAX: (404) 894-9320;
scott.haynes@coa.gatech.edu
This should be in the right column.