Abstract
The computer mouse is the primary input tool for a video display workstation. Mouse use reportedly occurs during 30%–80% of computer operation (Dennerlein & Johnson, 2006). Repetitive motions, awkward posture, and sustained muscle contractions during mouse use have been associated with carpal tunnel syndrome (Andersen et al., 2003; Carlson et al., 2010; Gerritsen et al., 2002; Hamilton, Jacobs, & Orsmond, 2005; IJmker et al., 2007; Keith et al., 2009; Kryger et al., 2003; Manente et al., 2001; Nobuta, Sato, Nakagawa, Hatori, & Itoi, 2008; Piazzini et al., 2007; Trujillo & Zeng, 2006). Muscle loads at wrist extension increase when users perform double clicking and dragging tasks with the mouse.
Cumulative muscle load plays a major role in carpal tunnel syndrome. Occupational therapists often recommend wrist splints to minimize carpal tunnel pressure, reduce pain, and support clients’ participation in their occupations. A well-designed splint and a fitted mouse are the keys to the successful treatment of carpal tunnel syndrome. One of the fruitful areas of splint design research has focused on the optimal angle of the wrist. Several studies have suggested the benefit of a neutral wrist position for patients with carpal tunnel syndrome (Carlson et al., 2010; Gerritsen et al., 2002; Keith et al., 2009; Manente et al., 2001; Nobuta et al., 2008; Piazzini et al., 2007). However, few studies have focused on the effect of forearm design in mouse control tasks, and limited research has focused on mouse design for patients with wrist splints.
Background
Effect of Cutaneous Feedback on Computer Tasks
When using a computer workstation, people use their vision to track the movements of the cursor and their proprioceptive and cutaneous feedback to sense the moving status of the mouse. The importance of cutaneous feedback is higher in more sophisticated tasks (Ebied, Kemp, & Frostick, 2004). Volar-based wrist orthoses, either custom made or commercial, increase muscle activities of upper limbs compared with no orthoses (Ferrigno, Cliquet, Magna, & Zoppi Filho, 2009). Volar-based splints provide less cutaneous feedback at the wrist (Brenda, Lohman, & Shultz-Johnson, 2001), although mice with cutaneous feedback can enhance performance in computer tasks (Akamatsu, MacKenzie, & Hasbroucq, 1995; de Korte, de Kraker, Bongers, & van Lingen, 2008; de Korte, Huysmans, de Jong, van de Ven, & Ruijsendaal, 2012). Our first hypothesis was that a volar-based splint, compared with a dorsal-based splint, would reduce the cutaneous feedback at the wrist and thus interfere with computer task efficiency.
Effect of Hump Position of Mouse on Computer Performance
For this study, wrist angle was kept in the extension position during mouse operation. Dragging and pointing tasks require the fingers to exert force, thus increasing wrist pressure significantly (Keir, Bach, & Rempel, 1999). The size, shape, weight, and material of the mouse change the grasp pattern of the user. Ergonomic mouse design can prevent awkward posture and minimize the risk of musculoskeletal injuries (Adapathya & Leonard, 2008). Much work has been done on the effects of size and shape of the mouse. Oude Hengel, Houwink, Odell, van Dieën, and Dennerlein (2008) suggested that bigger mice promote more neutral wrist postures and greater comfort. They also found that people with bigger hands used a larger wrist extension posture in mouse tasks.
Chen and Leung (2007) noted that muscle activities of the wrist extensor, trapezius, and pronator teres muscles increased as the slant angle of the mouse decreased. Houwink, Oude Hengel, Odell, and Dennerlien (2009) concluded that an alternative mouse that promoted less pronation posture in holding the mouse in turn lowered the risk of repetitive motion injuries. Because a wrist splint occupies the proximal end of the palm, users find it difficult to hold a mouse when wearing a wrist splint. Compared with a rear-hump mouse, a front-hump mouse creates more space at the proximal end and might allow better performance (Figure 1). Our second hypothesis, then, was that a front-hump mouse would be more suitable for patients with wrist splints. In this study, the hump position was predefined as the highest point as seen from the side view of a mouse. If the highest point of a mouse was at the front, the mouse was called a front-hump mouse ; on the other hand, if the highest point of a mouse was at the back, it was called a rear-hump mouse (Figure 2).
Participant tends to hold the front-hump mouse with distal fingers and leave more room for the palmar transverse bar of the splint than when holding the rear-hump mouse.
Mouse on the left is a front-hump mouse; mouse on the right is a rear-hump mouse. Arrows indicate the peak of the mouse as viewed from the side.
Anthropometric Index
Computer task performance is strongly affected by a good fit between hand size and input devices. Previous studies have used length of hand (Oude Hengel et al., 2008) and length of index finger (Johnson & Blackstone, 2007) to determine the size of the mouse for adults with smaller hands or for children. Adults with smaller hands had larger wrist extension and forearm supination movement while controlling mice than did adults with bigger hands (Oude Hengel et al., 2008). At the same time, adults with bigger hands required larger wrist extension movements and more muscle activation when using smaller mice. School-aged children abducted the shoulder and deviated the wrist to the ulnar side more than adults did in the use of standard mice (Johnson & Blackstone, 2007). Research has revealed that length of hand (Oude Hengel et al., 2008) or length of index finger (Johnson & Blackstone, 2007) can be good indexes for offering appropriate computer mice to participants without splints. Nevertheless, task demand may change if participants wear splints while operating mice.
We aimed to analyze the relationship between anthropometric measurements and computer task performance. Our research questions were as follows:
Do dorsal-based wrist splints and front-hump mice enhance computer task efficiency and comfort?
What kind of relationship exists between the anthropometry measurements and computer task efficiency?
Method
The study design was a 2 × 2 (Cutaneous Feedback × Hump Position) factorial experiment with repeated measures and two within-subject factors. We designated two levels for cutaneous feedback, namely, with feedback and without feedback. Similarly, for the hump position, we designated two levels, front and rear hump. The experimental procedure was approved by the ethics committee of Fu Jeng Catholic University, and we obtained informed consent from all participants. The following dependent variables were measured:
Efficiency: movement time (time needed to complete a set of mouse tasks)
Satisfaction: scores on the Quebec User Evaluation of Satisfaction With Assistive Technology—Taiwanese (QUEST–T; Mao et al., 2010) questionnaire
Anthropometric measurements: length of hand, length of palm, length of index finger, and width of hand.
Participants
Participants were recruited by advertisements placed around the campus of Fu Jeng Catholic University. The inclusive criteria were (1) age ≥18 yr; (2) right-hand dominant; (3) no symptoms of discomfort, pathology, or sequelae related to the right upper extremity; (4) no other physical dysfunctions that hindered normal right-hand function; and (5) experienced computer users: operating a computer at least 1 hr per day.
Satisfaction Survey
Electromyography (Chen & Leung, 2007; Dennerlein & Johnson, 2006; Gustafsson & Hagberg, 2003; Lee, Fleisher, McLoone, Kotani, & Dennerlein, 2007; Oude Hengel et al., 2008) and joint angles (Houwink et al., 2009; Ugbolue & Nicol, 2010; Wahlström, Svensson, Hagberg, & Johnson, 2000) are often used as indicators for measuring physiological loadings. Comfort and task performance are also closely related to musculoskeletal symptoms (Hagberg, Tornqvist, & Toomingas, 2002). ISO 9214 pointed out the importance of effectiveness, efficiency, and satisfaction in usability testing (Jokela, Iivari, Matero, & Karukka, 2003). Because the clinical effectiveness of wrist orthotics is well understood, we chose task performance and self-rated satisfaction as measurement outcomes (Carlson et al., 2010; Keith et al., 2009; Manente et al., 2001; Nobuta et al., 2008; Piazzini et al., 2007).
The QUEST is commonly used to assess user satisfaction with assistive technology (de Boer et al., 2008, 2009; Demers, Wessels, Weiss-Lambrou, Ska, & De Witte, 2001). The full inventory includes two subscales: satisfaction with device and satisfaction with service. The Device subscale consists of eight items, and the Service subscale consists of four items. Response choices for each item range from 5 (very satisfied) to 1 (not satisfied at all).
The Taiwanese version of the QUEST, the QUEST–T, was used in this study. The QUEST–T has good validity (two factors explained 53.42% of variance) and reliability (mean test–retest reliability = .62; Mao et al., 2010). The Device subscale (dimensions, weight, adjustments, safety, durability, ease of use, comfort, and effectiveness) was used in this study because the splint was provided directly by the research team. A sample item from the QUEST–T is, “When you are using the mouse, how satisfied are you with the dimensions (size, height, length, width) of your assistive device?”
Procedures
Anthropometric Data Collection.
Anthropometric data of each participant were obtained by using a digital clipper (Absolute Digimatic Caliper Series 500, Mitutoyo Corporation, Aurora, IL). Anthropometric data included (1) length of hand (distance between the distal wrist crease and the tip of the middle finger in extension), (2) length of palm (distance between the distal wrist crease and the proximal digital crease of the middle finger), (3) length of index finger (distance between the upper top of the proximal digital crease of the thumb and the finger tip of the index finger), and (4) width of hand (distance from the radial side the proximal palmar transverse crease to the ulnar side of the distal palmar transverse crease; Figure 3).
Anthropometric measurements. A = length of hand; B = length of palm; C = length of index finger; D = width of hand.
Mouse Trial Data Collection.
The test room used in this study was noise minimized, and participants faced the testing computer (U41J Series, ASUS, Taiwan) running Windows 7, with a 15-in. display running at 1,366 × 768 resolution) and a wall. The seat height was adjustable so that participants’ feet were flat on the floor with their knees and hips flexed at about 90°. In the beginning, participants were informed of the protocol and the purpose of this study. The order of the mice was then randomly assigned to participants. Before starting the standard pointing and selecting tasks, participants practiced five trials for each of the four conditions (2 mice × 2 feedback). Every condition contained 72 trials and was separated by 5-min breaks to reduce the cumulative effect of muscle fatigue. During breaks, participants completed a paper QUEST–T questionnaire. All procedures were performed by the same research assistant and supervised by a principal investigator.
Computer Mouse.
During the experiments, the participants used two types of mice. Type A was a Microsoft Wireless Mobile Mouse 1000, and Type B was a Microsoft Wireless Mobile Mouse 3500 (Microsoft Corporation, Redmond, WA). These two mice were compatible in length, width, height, and symmetry, except for the hump position. The length of both Type A and Type B mice was 95.21 mm. The width of the Type A mouse was 54.58 mm, and the width of the Type B mouse was 54.13 mm. The height of the Type A mouse was 34.72 mm, and the height of the Type B mouse was 70.78 mm. Type A was chosen as a representation of the front-hump position; Type B, of the rear-hump position.
Software.
Standardized pointing and selecting tasks were designed in Microsoft Word (Microsoft Corporation, Redmond, WA) using Microsoft’s Visual Basic program. The cursor started in a home box located at the center of the screen. Participants were allowed to move the cursor only by using the mouse. Each participant was asked to move the cursor to a circular target at the screen margin and to click on it. After it was clicked, the target disappeared; each disappearance counted as the end of a successful trial. Participants had to move the cursor back to the home box, and the target would show up again. The size of the target and the distance and angle from home box to target were varied across all 72 possible combinations of pixel sizes of 24, 48, or 64; pixel distances of 80, 160, and 320; and angles of 0°, 45°, 90°, 135°, 180°, 125°, 170°, and 315°. Targets randomly presented in any one of the 72 combinations. A complete set consisted of all 72 trials. Movement time (the time between leaving the home box and clicking on the circular target) of a complete set was recorded.
Splint.
Volar and dorsal wrist splints with thermoplastic materials were made by an occupational therapist with 14 yr of experience. Both types of splints had a fixed wrist angle at extension of 20°. Each type of splint was produced in large, medium, and small sizes. The dorsal wrist splint allowed participants to feel the cutaneous feedback at the wrist, whereas the volar wrist splint blocked the cutaneous feedback.
Data Analysis
The movement time data were examined for a normal distribution. Two-way repeated-measures analysis of variance (ANOVA) was used to analyze movement time data. The two main effects (hump position of mouse vs. cutaneous feedback) were both within-subject factors.
Questionnaire data were also analyzed with a normal distribution test. A one-way ANOVA was performed on the total scores of the questionnaire and the score of each item for the four conditions (cutaneous feedback–front hump, cutaneous feedback–rear hump, no cutaneous feedback–front hump, no cutaneous feedback–rear hump), and a least significant difference test was used for post hoc testing.
Relationships between the anthropometric measures and computer task efficiency were examined with Pearson product–moment correlation coefficients. Significance was noted for a probability of a false positive being less than 5% (p < .05).
Results
Thirty-two healthy participants enrolled in the study. One participant withdrew from the experiment because of a wrist sprain experienced 1 day before the appointment, and another participant dropped out because of technical problems. Therefore, the mouse trial data were partially incomplete. The final sample size was 30 and included 19 women and 11 men. Participants ranged in age from 18–27 yr, with a mean age of 21.5 yr.
Efficiency
A significant effect of type of cutaneous feedback was found on the movement time, F (1, 29) = 6.038, p = .020. The effect of type of mouse hump position on movement time was significant, F (1, 29) = 7.487, p = .010. No interaction effect between type of cutaneous feedback and hump position was found, F (1, 29) = 2.701, p = .111. The results showed that task efficiency (shorter movement time) was better with cutaneous feedback in the dorsal wrist-splint condition (mean [M] ± standard deviation [SD] = 63.43 ± 63.4 ms) than without cutaneous feedback in the volar wrist-splint condition (M ± SD = 67.26 ± 12.53 ms). At the same time, task efficiency was significantly better with rear-hump mice (M ± SD = 63.74 ± 10.33 ms) than with front-hump mice (M ± SD = 66.94 ± 10.48 ms; Figure 4).
Mean movement time of Type A mouse and Type B mouse for both types of cutaneous feedback. Type A mouse = front hump; Type B mouse = rear hump.
Satisfaction Survey
In general, total scores of the QUEST–T questionnaire were similar across the four experimental conditions, F(3, 116) = 0.402, p = 7.52. Item scores were also comparable for the four conditions and did not have significant differences. The p values were .084, F (3, 116) = 2.271; .767, F (3, 116) = 0.381; .559, F (3, 116) = 0.692; .132, F (3, 116) = 1.908; .923, F (3, 116) = 0.160; .726, F (3, 116) = 0.438; .364, F (3, 116) = 1.072, and .790, F (3, 116) = 0.349 for the dimensions, weight, adjustments, safety, durability, ease of use, comfort, and effectiveness, respectively.
Anthropometric Measurements
The result of the Pearson correlation test indicated a significant negative relationship between movement time and length of index finger, r = −.107, p < .001, whereas we found no significant relationship between movement time and the other anthropometric parameters (r = −.108, p = .469; r = −.026, p = .2996, and r = −.027, p = .285, for the length of hand, length of palm, and width of palm, respectively). The results disclosed that a longer index finger resulted in better task efficiency. In contrast, the length of hand, length of palm, and width of hand had no impact on task efficiency
Discussion
We examined two main questions: (1) Do dorsal-based wrist splints and front-hump mice enhance computer task efficiency and comfort? and (2) What kind of relationship exists between anthropometry measurements and computer task efficiency? Our evidence supports the hypothesis that dorsal-based wrist splints might provide better cutaneous feedback and, in turn, increase task efficiency. However, results indicating that a rear-hump mouse benefits task performance refute our original hypothesis. At the same time, we found no interaction between the type of splint and mouse hump position on task efficiency. Neither type of splint nor hump position influenced user satisfaction. With regard to the second question, the length of index finger clearly has a positive relation to task efficiency.
Satisfaction
We expected that a rear-hump mouse or volar-based splint might cause discomfort, but this hypothesis was not supported. First, both mice used in this study were relatively small (length = 9.5 cm) compared with a standard mouse (12.6 cm; Hughes & Johnson, 2012). Participants could hold the mouse at its distal end without losing the clicking movements. When participants held the mice that we provided, we found some room to accommodate the palmar transverse bar of the splint, so subjective discomfort may not have been as large as we expected. At the same time, satisfaction was context sensitive and could have been affected by other factors, such as subjective representativeness (Lawrence & Low, 1993). Future studies are needed to improve our understanding of user satisfaction in the context of splint application.
Second, the sensitivity of the QUEST–T questionnaire has not yet been examined. The differences in the satisfaction scores between the two groups might be too small to be detected by the questionnaire. We suggest applying these results with caution until the sensitivity of the QUEST–T questionnaire is provided.
Design Principles for the Mouse Hump Position
A possible mechanism by which the rear-hump mouse speeds task efficiency is that the postures participants with splints used to control mice were different from those adopted by users without splints. The prominent hill of the rear-hump mouse serves as a support base for the palm and allows the mouse to be fixed distally under the metacarpophalangeal joint rather than under the center of the palm. As a result, participants compensate by flexing their index fingers more to click or scroll; this motion decreases the conflict between the mouse and the splint. Oude Hengel et al. (2008) found that participants flexed the metacarpophalangeal, proximal interphalangeal, and distal interphalangeal joints more when using smaller mice; our results are consistent with the work of Oude Hengel’s team.
Moreover, we found that a larger flexion range of motion at the interphalangeal joints is a consequence of wrist immobilization. Several previous research projects have confirmed that wrist immobilization causes compensatory movements in adjacent joints, such as more shoulder flexion (King, Thomas, & Rice, 2003), shoulder abduction (Chan & Chapparo, 1999; King et al., 2003; Shu & Mirka, 2006), and trunk lateral flexion (Shu & Mirka, 2006) and less elbow flexion range of motion (Chan & Chapparo, 1999). Increased shoulder muscle activities were also used as a compensation for wearing wrist orthoses (Yoo, Jung, Jeon, & Lee, 2010). Although no research has focused on the compensatory movements at the fingers, it is reasonable to infer that compensatory movements exist at the distal joints.
Role of Cutaneous Feedback
This study found little interaction between the type of feedback and the hump position. Cutaneous feedback is crucial in mouse operation, regardless of which type of mouse is used. Therefore, a dorsal wrist splint can provide cutaneous feedback to maximize computer task performance, and it can meet the needs of stabilizing the wrists to minimize further wrist pressure. Our findings corroborate the research of Cockburn and Brewster (2005), who found that mice with 200-Hz vibration and with stickiness feedback can significantly improve targeting behaviors. Vitense, Jacko, and Emery (2003) also reported that in a visual-plus-haptic-feedback (mice vibrate for 0.3 s on target) condition, target highlight time and mental workload were significantly decreased. Although the cutaneous feedback in two studies was provided by the mouse itself, the role of cutaneous feedback remains robust.
Mouse operation is a delicate task that requires coordination of multiple sensory inputs. In this study, movement time was longer without cutaneous feedback. Ebied et al. (2004) also found that the function of cutaneous feedback was more important in a demanding handwriting task than in one of maintaining submaximal force.
Anthropometric Measurements
This study showed that length of index finger had a significant negative correlation with movement time. No significant correlations were found with the rest of the anthropometric measurements. Thus, when wearing a splint, participants with longer index fingers performed better on mouse operation tasks. Similarly, Johnson and Blackstone (2007) discovered that the average index finger of a 5-yr-old child was only one-third the length of the 50th-percentile adult’s finger. These anthropometrical differences led to greater wrist ulnar deviation and less extension in children than in adults in computer tasks. In contrast, other studies applied hand length as an anthropometric indicator to calculate the size of computer devices. Hughes and Johnson (2012) used hand length to calculate the length, width, height, and switch location of the mouse and suggested four different mouse sizes for users of different ages and genders. Similarly, Oude Hengel et al. (2008) used hand length to categorize users with smaller or bigger hands and found that users with bigger hands had more wrist extension and muscle activation when using a smaller mouse. Although hand length is a common anthropometric index, we propose that length of the index finger is more related to the computer task efficiency of participants with wrist splints.
Limitations and Future Research
Limitations of our study include that all the participants were healthy and right-hand dominant. In the future, adding participants with pain at the wrists or who are left-hand dominant will enlarge the scope of study. In addition, other confounding factors for design features, such as height and material of the mouse, may have affected the results. Finally, our survey on satisfaction did not yield significant differences. Future research could extend the wearing period of the splints and allow participants more time to examine their subjective feelings about the splints. At the same time, a custom-made splint for each participant could minimize potential ill-fitting and confounding factors in the future.
Conclusion
Musculoskeletal injuries are closely associated with mouse use. Extended mouse operation increases the damage to the wrist and hand tissues and causes pain at the wrists. Wrist splints are often recommended to clients with wrist pain to stabilize the joint and enhance participation in their daily occupations. However, wearing a splint and holding a mouse at the same time often cause clumsiness or inconvenience. In summary, dorsal wrist splints coupled with a rear-hump mouse provide sufficient cutaneous feedback and increase task efficiency. Meanwhile, we recommend that the length of the index finger be the prime anthropometric measurement related to computer task performance for participants with a splint.
Implications for Occupational Therapy Practice
The results of this study have the following implications for occupational therapy practice:
If computer tasks are a major daily client activity, clinicians should choose a dorsal wrist splint design and a rear-hump mouse to preserve the cutaneous feedback.
If a volar-based splint is desired, a rear-hump mouse offers a stable rest area for the fingers and facilitates coordination in mouse scrolling and clicking.
Footnotes
Acknowledgments
We acknowledge the study participants. We also thank Rong-Kwei Li, Ya-Fang Liu, and Wei-Cheng Shen for inspiring the research topic and reviewing the manuscript of this article.