Introduction Methods Subjects Apparatus Procedures Data Analysis Results Typing performance Survey Results Interim Conclusions References

1.0 INTRODUCTION

A growing body of evidence indicates wrist intracarpal pressure increases whenever the hands deviate sufficiently from a neutral position into vertical extension/flexion hand movements and/or with lateral radial/ulnar deviation of the hands (Rempel et al., 1994; Rempel and Horie, 1994; Rempel et al., 1997). Hand positions of moderate extension (>20°) or flexion (>20°) may prevent the free flow of fluids into the palm (Gelberman et al., 1984). Animal studies show that increases in CTP above 30 mm Hg can disrupt blood flow and impair median nerve function. (Dahlin., 1991). Indeed, nerve conduction velocity changes start to occur with CTP above 40 mm Hg, and complete blocking of nerve signals if this CTP is sustained for an 8 hour period (Hargens et al., 1979). CTP exceeds 40 mm Hg whenever the wrists are either flexed or extended beyond 20° (Gelberman et al., 1984, Rempel et al., 1997). Ergonomists universally agree that maintaining the hands in this neutral zone of motion is desirable to reduce the risks of a hand-wrist injury.

Use of a conventional computer keyboard on a desk can increase the amount of time the hands spend in wrist extension and ulnar deviation during typing. When placed on a desktop or on a flat, lowered keyboard tray, extreme excursions of wrist extension beyond 20° can readily be observed for many users during typing, as users curl their fingers with hands in dorsiflexion. Consequently, a variety of alternative keyboard designs have been developed. Most of these keyboards are simple geometric variants on the conventional horizontal keyboard design, and test results have been mixed. Indeed, However, Swanson et al. (1997) report that simple, split keyboard designs may have only a minimal impact on comfort, self-reported fatigue and productivity.

Chen et al. (1994) tested wrist posture and typing speed for four keyboards: flat conventional, Apple adjustableTM, Kinesis, and Comfort. All keyboards were used on an articulating keyboard tray. Median wrist extension angles for the four keyboards were 21.8°, 16.9°, 11.7°, and 18.4° respectively. Median values for wrist ulnar deviation were 14.7°, 15.5°, 4.5°, and 13.5° respectively. Typing speed was comparable for the traditional, Apple adjustable, and Comfort keyboards (between 40 - 45 words per minute), but typing was considerably slower using the Kinesis keyboard (23 - 30 words per minute). Smith and Cronin (1993) reported similar results for the Kinesis keyboard. Honan et al.(1995) compared postural differences when typing on 3 different keyboards (Apple Extended, and the Microsoft Natural keyboard with and without the leveler extended ). She found that wrist posture was improved with the Microsoft Natural keyboard without the Leveler . Hedge and Shaw (1996) tested a chair-mounted split keyboard (Floating Arms) and showed that the complete splitting of the keyboard reduced ulnar deviation, and this keyboard only reduced wrist extension when it was tilted downwards at 15° slope.

None of the alternative keyboard designs tested to date have addressed the issue of wrist pronation. This posture may be important because in pronation the ulna and radial bones cross and this reduces the volume of the channels through which the median and ulnar nerves pass, which may increase compression risks, causing what has been termed the "double-crush" syndrome. The maximum volume of the channels occurs when the hands are mid-pronate where the ulna and radial bones are uncrossed.

Recently, a prototype, vertically split-keyboard design has been developed. This revolutionary design offers the promise of improving wrist posture during typing and also eliminating forearm pronation during typing. The present study is a laboratory experiment designed to test the postural effects of typing on this vertical split-keyboard design compared with typing on a conventional keyboard.

2.0 METHODS

2.1 Location

The study was conducted at Cornell University’s Human Factors Laboratory in Ithaca, New York during the months of July and August 1998. The average environmental conditions of the experimental room was 75.2º F (SD= 2.3) and 45.5% relative humidity (SD=5.1).

2.2 Study Design

The study was a randomized block experimental design consisting of five conditions (Table 1). Every subject was exposed to each of the five conditions consecutively. To control for block effects and increase generalizability of the results, a Latin Square Design was used to randomize the order of the conditions over the twelve subjects.

Table 1. Summary of Experimental Design
CONDITION 1	CONDITION 2	CONDITION 3	CONDITION 4	CONDITION 5
Chair 1: Criterion	Chair 1: Criterion	Chair 2: Ergomax	Chair 2: Ergomax	Chair 2: Ergomax
Vertical Keyboard	Conventional keyboard	Vertical Keyboard	Conventional keyboard	Vertical Keyboard Arm Rests

 

Figure 1: Photographs of the five experiemental conditions tested in the study
Chair 2: Ergomax Conventional Keyboard	Chair 1: Criterion Conventional Keyboard
Chair 2: Ergomax Vertical Keyboard	Chair 1: Criterion Vertical Keyboard
Chair 2: Ergomax Vertical Keyboard Arm Rests

2.2.1 Research Measures

The study was designed to measure the effects of a vertical split keyboard as compared to the effects of a conventional horizontal keyboard.

The objective measures assessed were dynamic wrist postures, typing performance, and upper body posture. The dynamic wrist postures of the flexion/extension and radial/ulnar movements were measured with the use of the Green Leaf Medical Wrist System and the accompanying monitoring software, MAS 6.0 for the Macintosh computer. Typing performance was assessed in words per minute (wpm) and percent accuracy with the use of the software Typing Tutor 6 for the PC computer. Upper body posture was measured in relative angles with the use of a four-angle video-motion analysis.

The subjective measured assessed were self-report ratings of discomfort, fatigue and workstations preferences, and a non-biased, overt behavioral analysis of upper body posture. The self-report ratings of discomfort and fatigue were measured with a questionnaire (Appendix A). Subjects were asked to rate their perception of discomfort and fatigue for 18 segments of the body on a four-point scale. Questionnaires were completed immediately before the first trial and immediately after each condition. A questionnaire on overall workstation preference was completed immediately after the last trial (Appendix B). The non-biased, overt behavior analysis of upper body posture was assessed.

2.3 Subjects

Twelve female, experienced touch-typists were recruited from the Cornell University community and private sector to participate in the study as paid volunteers. Subjects were predominately employed as professional administrative assistants or graduate students. All subjects spoke English fluently and were right-handed. No subject had a fingernail length that might have reduced performance of keyboarding exercises. The ages of the subjects ranged from 24 to 53 years with a mean age of 31.1 (SD=9.3) years. The number of years of keyboarding experienced reported by the subjects ranged from 6 to 30 years with a mean of 14.3 (SD=6.5) years. The standing height of the subjects ranged from 60.5 to 68.5 inches with a mean of 64.8 inches (SD=2.9). No subjects reported to have ever been treated by a physician for a cumulative trauma disorder. All subjects were screened for typing performance to ensure that they could type at least 45 wpm on a conventional, horizontal QWERTY keyboard.

2.4 Apparatus

Subjects were tested on the same workstation with identical equipment and layout. The following is a list of equipment used in experimental room.

Workstation:

• Standard height work table (height = 73 cm; width = 121 cm; depth = 75 cm)

• Adjustable height articulating keyboard tray

• Gateway, 166Mhz, P5-166 Personal Computer

• Gateway, Vivitron17 inch monitor

• Gateway conventional horizontal QWERTY keyboard (model number 219600)

• QWERTY vertical split-keyboard prototype (henceforth called the "vertical keyboard") designed by Ergonomic-Interface Keyboard Systems, Inc.

• Steelcase Criterion adjustable chair

• American Ergonomics Ergomax 2000 adjustable chair with attached, adjustable, opposable arm rests

• Adjustable foot rest

• Adjustable palm/wrist support

• Floor fan

          Monitoring and support equipment:

• Green Leaf Medical Wrist System which included monitoring gloves and data recorder.

• Macintosh laptop which recorded measurements for the Wrist System

• Typing Tutor 6 software package for the PC

• Four cameras:
  • Right angle– Panasonic VHS AG-450 SVHS movie camera
  • Left angle – Panasonic VHS AG-185 VHS movie camera
  • Rear angle – Panasonic AG-188 VHS movie camera
  • Above angle – Sony Hi8 Handy Cam connected to a VHS VCR
    
    
• Two 500 Watt spot lights positioned at 171 cm in height

2.5 Procedures

Subjects completed all five conditions in one 3 to 4 hour experimental session on one day. Subjects were tested independently by the same experimenter. All twelve subjects were tested within three weeks time. A chronological summary of the steps used to implement this experimental design for each subject is as follows:

• Subject preparation. Subjects were asked to wear short-sleeved tops to allow for a clear view of upper-arm and shoulder posture. Subjects were asked to trim fingernails to a reasonable length and remove all hand and wrist jewelry, and pin up long hair.

• Subject Briefing. Subjects were briefed on relevant information of what to expect from the monitoring devices and the methods of testing. A clear description of the body segments referred to in the questionnaires was provided.

• Initial measures. Subject’s standing height without shoes was measured against a flat wall with a T-square and recorded. Both hands were traced on paper with a pen. The same pen was used for all subjects. Subjects completed the pre-experiment questionnaire to indicate feelings of discomfort and fatigue.

• Workstation preparation. Subjects were then fit to one of the five conditions specified by the pre-randomized trials. Adjustments to keyboard, chair, footrest, and monitor height were made to ensure subject comfort and that postures were within ergonomic guidelines for workstation configuration. The following describes the procedure used for adjusting every component of the workstation. The chair was adjusted to the subject so that the thighs were parallel to the floor, feet were flat on the floor when knees were at a 90º bend (a footrest was used if when needed), and lumbar support was in a comfortable position. Subjects adjusted the tilt of the chair to their comfortable keyboarding posture. A reclined posture was not encouraged. The chair was centered to the worktable. The keyboard tray was adjusted to the subject so that the initial typing posture would consist of the subject’s elbow angle at approximately 90º and the wrists in a relatively neutral posture. The tilt of the keyboard tray was measured with a level and maintained as close to level as equipment would allow. Tilt-down of the conventional keyboard was not allowed to provide a consistent platform for both keyboards. The wrist rests of both keyboards were adjusted to subject’s comfort. The horizontal wrist rests was removed from the keyboard tray when testing the vertical keyboard. Both keyboards were centered to the subject and workstation. The conventional keyboard was centered to the middle of the alphanumeric keys (excluding the use of the number keypad). The vertical keyboard was adjusted for width. The monitor height and distance was adjusted so that the center of the screen was approximately 15-208 below the subject’s line of sight. When used, the forearm rests were adjusted for subject’s comfort and neutral initial typing wrist posture. Adjustments were fine-tuned through subject’s feedback of comfort and the experimenter’s visual inspection of posture. Monitor height, arm rest height, keyboard height, and seated elbow height were recorded and held constant for both chairs. Measurements were recorded for both the vertical and conventional keyboards.

• Wrist System Calibration. Monitoring gloves were fit to the subject’s hands and calibrated using the manufacturer’s recommended method. The subjects did not remove the gloves after this calibration exercise.

• Video-motion preparation. Reflective markers were fastened to bony landmarks located on the shoulders and elbows. The gloves also had markers on them.

• Experimentation. Subjects sequentially completed each of the five 15 minute typing tasks presented in a random task-condition order. During typing subjects wore the monitoring gloves and were videotaped from four angles (left, right, rear, and above). Fifteen minutes of typing was recorded for each condition. Wrist posture was measured at 10 Hz.

• Questionnaires. Immediately following each trial, the subjects completed a questionnaire on their feelings of discomfort and fatigue for 18 body segments. At the end of the fifth trial, subjects completed an overall workstation preference questionnaire.

• Breaks. Subjects were given approximately 10 minutes of break between trials

2.6 Data Analysis

2.6.1 Wrist angles.

Data points for both radial/ulnar and flexion/extension angles for both the right and left hands from the Green Leaf Wristmaster were recorded at 10 Hz on a Macintosh Laptop with MAS version 1.06 for each condition of every subject. The data files, 60 in total, were exported and used with Microsoft Excel ’97 and SPSS version 7.5 for the PC. The experimenter examined each data file and subjective judgements were made concerning the quality of data of data points. Portions of the data that were thought to be non-representative of each subject’s keyboarding movements (such as the first and last minute of testing and movements such as scratching the face, etc.) were deleted from the analysis. The number of data points for each file used in the analysis ranged from 6,602 (11 minutes) to 8,581 (14.3 minutes) data points with a mean of 8,095 (SD=431). These edited data sets were the foundation for the following analysis.

2.6.2 Comparison of mean wrist angles.

The mean for both radial/ulnar and flexion/extension wrist angles were calculated independently for every subject, condition, and hand. The SPSS General Linear Model Simple Factorial (GLMSF) procedure was used to test for significant main and interaction effects in mean wrist angle data at alpha = 0.05. Trial assignment was randomized among subjects and examination of the residuals suggests that the order of the trials did not require a repeated-measures analysis model. Two analyses were conducted: one to compare the differences in mean wrists angles for the variables keyboard and chair, and another to compare the differences between mean wrist angles with and without the use of forearm rests (henceforth referred to simply as 'arm rest'). To determine the differences in mean wrist angles for keyboard and chair, data from conditions 1-4 were used. Using GLMSF analysis, keyboard, hand, and chair were assigned as fixed factors, and subject was assigned as a random factor for the dependent variables of radial/ulnar and flexion/extension wrist angles. To determine the differences in mean wrist angles with and without the use of forearm rests, data from conditions 3 and 5 were used, both conditions in which the vertical split-keyboard was used. In this GLMSF analysis, hand and forearm rests were assigned as fixed factors, the variable subject was assigned as a random factor for the dependent variables of radial/ulnar and flexion/extension wrist angles. All post-hoc analyses of interactions were performed using Tukey HSD with an alpha=.05.

2.6.3 Plotting the 2-D and 3-D graphs.

SPSS was used to randomly select 3,000 data points from every subject for all five conditions. Two-dimensional scatter plots of radial/ulnar by flexion/extension wrist angles were created for right and left hand for all five conditions using the pooled data sets (36,000 data points). Two-dimensional scatterplots were created using Plot 1.2 (Fortner Research) running a custom written colorizing macro. Each 2-D scatter plot is colorized according to the radial distance of a movement from a true geometric neutral point (i.e. zero degrees extension/flexion and zero degrees radial/ulnar deviation). As the radial distance increases from the 0,0 origin (violet) so the colors increase in wavelength, moving through the spectrum until at a radial distance of 20º all points are colored red. This 20º limit corresponds with the outer limits of a neutral zone for hand movements while typing. The amount of red in each graph and the distance of the red points from the neutral zone gives a sense of the amount of risk associated with using that keyboard with that hand under the specified test conditions.

Following these 2-D analyses, the joint frequencies of these data sets (i.e. the frequency of occurrence of all combinations of radial/ulnar by flexion/extension wrist angles), were computed separately for right and left hand for all five conditions and used as the datsets for a series of three-dimensional surface plots. The surface plots were created in Transform 3.3.1 (Fortner Research) running a custom written colorizing macro. The color scheme of the 3-D plots is identical to the 2-D scatterplots. Frequency, or number of observations noted at each unique combination of radial/ulnar and flexion/extension movements, is denoted as "N".

2.6.4 Risk zone analysis.

To quantitatively estimate risks, the original edited data sets were used to divide the recorded radial/ulnar and flexion/extension wrist angles into zones. The ranges of wrist angles contained by each zone are as follows:

• Zone 1: – 10.5º to 10.5º
• Zone 2: – 5.5º to -10.6º and 10.6º to 15.5º
• Zone 3: – 5.6º to -20.5º and 15.6º to 20.5º
• Zone 4: <–20.6º and >20.6º

Data sets of the percent of radial/ulnar or flexion/extension wrist angles of each hand that occurred in each of the four zones in each condition for each subject (total of 240 data files) were created. Following this, a composite data set of the percentage of movements in each of the four zones for all subject x hand x conditions was created. Two multivariate analyses were then conducted on these data: one to compare the differences in the percentage of wrists angles in each zone for the variables keyboard, chair, and hand and another to compare the differences between the percentage wrist angles in each zone with and without the use of arm rests. A GLMSF analysis was used as described in the comparison of mean wrist angle analysis.

2.6.5 Survey analysis

The subject’s self-reports of fatigue and discomfort were analyzed by aggregating the four point scale (never, slight, moderate, severe) into a dichotomous scale (never/slight and moderate/severe). The results were compiled in two tables to present the number of subjects reporting moderate/severe discomfort or fatigue after each condition. Preferences were also analyzed. Final analysis of all of the survey data is in process.

3. 0 RESULTS

3.1 Comparison of mean wrist angles

3.1.1 Analysis of the effects of keyboard and chair.

Flexion/extension wrist angles.

The analysis of between-subjects effects indicated that there were two significant main effects on the mean flexion/extension wrist angles: keyboard (p=.001) and hand (p=.006). There was a significant effect of the interaction of keyboard and hand on mean flexion/extension wrist angles (p=.023). The effect of chair was not significant.

Post hoc analysis of the interaction showed that the flexion/extension mean wrist angles of the vertical keyboard were significantly lower for both the right and left hands than those of the conventional keyboard (p<.000 and p=.001 respectively).

The mean flexion/extension wrist angles for the right hand was 9.49º less for the vertical keyboard (.03º) than the conventional keyboard (9.52º). The mean flexion/extension wrist angles for the left hand was 4.61 º less for the vertical keyboard (8.40º) than the conventional keyboard (13.01º). The data indicated that the mean flexion/extension wrist angles for the right and left hands were significantly different from each other for both keyboards.

The mean wrist angles for the right hand while using the vertical keyboard was a significant 8.38º less than the mean wrist angles for the left hand on the same keyboard (p<.000). The mean wrist angles for the left hand while using the conventional keyboard was a significant 3.49º less than the mean wrist angles for the left hand on the same keyboard (p=.015).

The variability in the means of flexion/extension wrist angles between subjects was analyzed by using the standard deviations in an analysis of variance (GLMSF). There were no significant differences between the standard deviations for any of the relevant one-way, two-way, or three-way factors.

Radial/ulnar wrist angles.

The analysis of between-subjects effects indicated that keyboard was a significant main effects on the mean radial/ulnar wrist angles (p<.000). A significant interaction effect of the of keyboard and hand on mean radial/ulnar wrist angles was found (p=.026). The effects of chair and hand were not significant. Post Hoc analysis showed that the radial/ulnar mean wrist angles of the vertical keyboard were significantly lower for both the right and left hands than those of the conventional keyboard (p<.000 and p<.000 respectively).

The mean radial/ulnar wrist angles for the right hand was 8.60º less for the vertical keyboard (3.56º) than the conventional keyboard (12.17º). The mean radial/ulnar wrist angles for the left hand was 12.28º less for the vertical keyboard (6.46º) than the conventional keyboard (18.74º). The data indicated that the mean radial/ulnar wrist angles for the right and left hands were significantly different from each other for both keyboards.

The mean wrist angles for the right hand while using the vertical keyboard was a significant 2.90º less than the mean wrist angles for the left hand on the same keyboard (p=.001). The mean wrist angles for the left hand while using the conventional keyboard was a significant 6.57º less than the mean wrist angles for the left hand on the same keyboard (p<.000).

The variability in the means of radial/ulnar wrist angles between subjects was analyzed by using the standard deviations in an analysis of variance. There were no significant differences between the standard deviations of the means for any of the relevant one-way, two-way, or three-way factors.

3.12 Analysis of the effects of the use of forearm rests.

Flexion/extension.

The analysis of between-subjects effects indicated that there was a significant main effect of hand on the mean flexion/extension wrist angles (p<.000). The interaction of forearm rest and hand was marginally significant (p=.053). The main effect of forearm rest was not significant.

The variability in the means of flexion/extension wrist angles between subjects was analyzed by using the standard deviations in an analysis of variance. Only the main effect of hand was significant (p=.048). There were no other significant differences between the standard deviations for any of the other relevant one-way, two-way, or three-way factors.

Radial/ulnar.

The analysis of between-subjects effects indicated that there were no significant main effects or interactions of hand or forearm rest on mean radial/ulnar wrist angles.

The variability in the means of radial/ulnar wrist angles between subjects was analyzed by using the standard deviations in an analysis of variance. There were no significant differences between the standard deviations for any of the other relevant one-way, two-way, or three-way factors.

3.13 Graphical analysis of right hand wrist movements while using the Criterion chair.

The following 4 figures show the 2-D scatterplots and 3-D surface plots for the distributions of hand movements for the right hand using both a conventional and a vertical keyboard for subjects sitting in the Criterion chair without forearm rests. These graphs show that the distribution of right hand movements is considerably worse (i.e. more movements in the red zone) for the conventional keyboard than the vertical keyboard

Select 2-D scatterplots or 3-D surface plots for enlargement.
HORIZONTAL	VERTICAL

3.14 Graphical analysis of left hand wrist movements while using the Criterion chair.

The following 4 figures show the 2-D scatterplots and 3-D surface plots for the distributions of hand movements for the left hand using both a horizontal and a vertical keyboard for subjects sitting in the Criterion chair without forearm rests. These graphs show that the distribution of left hand movements is considerably worse (i.e. more movements in the red zone) for the conventional keyboard than the vertical keyboard.

Select 2-D scatterplots or 3-D surface plots for enlargement.
HORIZONTAL	VERTICAL

3.15 Graphical analysis of right hand wrist movements while using the Ergomax chair.

The following 4 figures show the 2-D scatterplots and 3-D surface plots for the distributions of hand movements for the right hand using both a horizontal and a vertical keyboard for subjects sitting in the ErgoMax chair without forearm rests. These graphs show that the distribution of right hand movements is considerably worse (i.e. more movements in the red zone) for the conventional keyboard than the vertical keyboard.

Select 2-D scatterplots or 3-D surface plots for enlargement.
HORIZONTAL	VERTICAL

3.16 Graphical analysis of left hand wrist movements while using the Ergomax chair.

The following 4 figures show the 2-D scatterplots and 3-D surface plots for the distributions of hand movements for the right and left hand using a vertical keyboard for subjects sitting in the ErgoMax chair with forearm rests.

Select 2-D scatterplots or 3-D surface plots for enlargement.
HORIZONTAL	VERTICAL

3.17 Risk zone analysis - Flexion/extension and keyboards.

The following graph shows a comparison of the overall distribution of the flexion/extension movement risks zones for the conventional horizontal and vertical keyboards. Significant differences between the size of the zones for each keyboard are indicated by 'p' values on the graph. Results show that 71% of hand flexion/extension typing movements were in the lowest risk zone when using the vertical keyboard compared with only 44% when using a conventional keyboard.

3.18 Risk zone analysis - Radial/ulnar deviation and keyboards.

The following graph shows a comparison of the overall distribution of the radial/ulnar deviation movement risks zones for the conventional horizontal and vertical keyboards. Significant differences between the size of the zones for each keyboard are indicated by 'p' values on the graph. Results show that 78% of wrist radial/ulnar deviation typing movements were in the lowest risk zone when using the vertical keyboard compared with only 25% when using a conventional keyboard.

3.19 Risk zone analysis - Flexion/extension, keyboards and hands

The following graph shows a comparison of the overall distribution of the flexion/extension movement risks zones for the right and left hands for the conventional horizontal and vertical keyboards. Significant differences between the size of the zones for each keyboard are indicated by 'p' values on the graph. Results show that 81% of right hand flexion/extension typing movements and 61% of left hand flexion/extension typing movements were in the lowest risk zone when using the vertical keyboard compared with only 49% of right hand flexion/extension typing movements and 39% of left hand flexion/extension typing movements when using a conventional keyboard.

3.20 Risk zone analysis - Radial/ulnar deviation, keyboards and hands.

The following graph shows a comparison of the overall distribution of the radial/ulnar deviation movement risks zones for the right and left hands for the conventional horizontal and vertical keyboards. Significant differences between the size of the zones for each keyboard are indicated by 'p' values on the graph. Results show that 81% of right hand radial/ulnar deviation typing movements and 76% of left hand adial/ulnar deviation typing movements were in the lowest risk zone when using the vertical keyboard compared with only 39% of right hand adial/ulnar deviation typing movements and 10% of left hand adial/ulnar deviation typing movements when using a conventional keyboard.

3.2 Typing performance

There was a small but statistically significant fall in the average typing speed of subjects of around 10% using the vertical keyboard. There was a very small but statistically significant decrease in typing accuracy of around 2% when subjects used the vertical keyboard.

3.3 Survey Results

Reports of moderate/severe musculoskeletal discomfort and fatigue (maximum of 18 points per subject in each condition) were minimal for all but one subject (see table below).

Fatigue    1 2 3 4 5 6 7 8 9 10 11 12   Initial 0 0 0 0 0 0 0 0 2 0 0 0 Horizontal Criterion 1 0 0 0 0 0 0 0 0 0 0 0 Horizontal Ergomax 0 0 3 3 2 0 0 0 0 0 0 0 Vertical Criterion 2 0 0 7 0 1 0 6 0 5 0 1 Vertical Ergomax 0 0 0 8 2 0 6 3 4 0 0 1 Vertical Ergomax w/ Armrest 0 0 0 12 0 0 0 0 0 0 0 - Discomfort   1 2 3 4 5 6 7 8 9 10 11 12   Initial 0 0 0 0 0 0 0 0 0 0 0 0 Horizontal Criterion 1 0 0 0 0 2 1 0 0 0 0 0 Horizontal Ergomax 0 0 0 2 0 0 0 0 0 0 0 0 Vertical Criterion 2 1 0 6 0 1 1 0 3 5 0 1 Vertical Ergomax 2 0 0 0 2 1 3 0 0 0 0 0 Vertical   - - - - - - - - - - - - -

4.0 INTERIM CONCLUSIONS

The results from this laboratory study that have been analyzed to date clearly show that the use of a vertical split-keyboard design can substantially improve typing wrist postures for subjects, both in terms of flexion/extension and radial/ulnar deviation movements. These postural improvements occur immediately within the short-term typing tasks (15 minutes) that were used. These results were obtained for touch typists who were not experiencing any musculoskeletal problems, and this represents a 'best-case' scenario for typing

Even though the vertical split-keyboard is an unfamiliar design there was only a relatively small decrement in typing performance and a very small difference in error rates between keyboards. Given the short-term nature of this study it is predicted that touch typist subjects will quickly master the vertical split-keyboard because this preserves the familiar QWERTY layout.

Little systematic variation was seen between keyboards in the survey data for musculoskeletal discomfort and fatigue. This is perhaps not surprising given the short-term design of the study. However, in view of the improvements in hand posture during typing with the vertical split-keyboard, longer-term musculoskeletal benefits can be anticipated. A field study is needed to test for such effects.

As previously noted, the vertical split-keyboard is an unfamiliar design and therefore it is perhaps not surprising that, when asked about keyboard preference and comfort, the subjects, who were not experiencing musculoskeletal discomfort and pain, indicated a preference in favor of workstation configurations incorporating the familiar conventional 101 key keyboard.

Motion analyses of the videotape recordings is underway to determine the effects of the vertical split-keyboard design on detailed aspects of seated body posture.

5.0 REFERENCES

CHEN, C., BURASTERO, S., TITTIRANONDA, P., HOLLERBACH, K., SHIH, M. and DENHOY, R, 1994, Quantitative evaluation of 4 computer keyboards: wrist posture and typing performance, Proceedings of the Human Factors and Ergonomics Society 38th Annual Meeting, 2, (Human Factors and Ergonomics Society, Santa Monica), 1094-1098.

DAHLIN, L.B, 1991 Aspects on pathophysiology of nerve entrapments and nerve compression injuries, Neurosurgery Clinics of North America, 2, 21-29.

GELBERMAN, R.H., SZABO, R.M., and MORTENSON, W.W, 1984, Carpal tunnel pressures and wrist position in patients with Colle's fractures, Journal of Trauma, 24, 747-749.

HARGENS, A.R., ROMINE, J.S., SIPE, J.C., EVANS, K.L., MUBARAK, S.J., and AKESON, W.H, 1979, Peripheral nerve-conduction block by high muscle-compartment pressure, Journal of Bone and Joint Surgery, 61-A, 192-200. HEDGE, A. and SHAW, G.F. (1996) Effects of a chair-mounted split-keyboard on wrist posture, comfort and performance. Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting, 1, (Human Factors and Ergonomics Society Santa Monica),624-628.

HONAN, M., SERINA, E., TAL, R. and REMPEL, D, 1995, Wrist postures while typing on a standard and split keyboard, Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting, 1, (Human Factors and Ergonomics Society Santa Monica), 366-368.

HORIE, S., HARGENS, A. and REMPEL, D, 1993, Effect of keyboard wrist rest in preventing carpal tunnel syndrome, Proceedings of the American Public Health Association Annual Meeting, Asbtract.

REMPEL, D. and HORIE, S, 1994, Effect of wrist posture during typing on carpal tunnel pressure, in A. Grieco, G. Molteni, E. Occhipinti and B. Piccoli (eds.) Work With Display Units '94, Proceedings of the Fourth International Scientific Conference, 3, (University of Milan, Milan, Italy), C27-C28.

REMPEL, D., HORIE, S., and TAL, R, 1994, Carpal tunnel pressure changes during keying. Proceedings of the Marconi Keyboard Research Conference, (UC San Franscisco, Ergonomics Laboratory, Berkeley), 1-3.

SMITH, W.J. and CRONIN, D.T, 1993, Ergonomic test of the Kinesis keyboard, Proceedings of the Human Factors and Ergonomics Society 37th Annual Meeting, 1, (Human Factors Society, Santa Monica), 318-322.

SWANSON, N.G., GALINSKY, T.L., COLE, L.L., PAN, C.S. and SAUTER, S.L, 1997, The impact of keyboard design on comfort and productivity in a text-entry task. Applied Ergonomics, 28, 9-16