Evaluation of four cursor control devices during a ...

ARTICLE IN PRESS Applied Ergonomics xxx (2009) 1–11

Contents lists available at ScienceDirect

Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo

Evaluation of four cursor control devices during a target acquisition task for laparoscopic tool control S.R. Herring, A.E. Trejo, M.S. Hallbeck* Innovative Design and Ergonomic Analysis Laboratory, Industrial and Management Systems Engineering, University of Nebraska-Lincoln, NE 68588, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 July 2007 Received in revised form 1 April 2009 Accepted 4 April 2009

Current laparoscopic surgery instruments create awkward postures which produce fatigue and pressure points in surgeons. In order to alleviate some of this discomfort a new laparoscopic tool had been developed with the inclusion of an articulating end-effector manipulated by a trackball. The current study was developed to access the performance of four input devices which could replace the manual trackball in a powered laparoscopic tool. A simple Fitts’ law task was conducted and the devices’ performance was evaluated with both subjective and objective measures. This article makes three main contributions to the scientific community. First, it provides a comparison of four control devices (TouchPad, Mouse Button Module, MiniJoystick Module and MicroJoystick) for use in a powered laparoscopic tool. Second, it provides an understanding of how the non-traditional measure of target re-entry can be utilized to compare control devices and how this relates to the more traditional measures of throughput and error rate. Finally, it contributes to the understanding of how a user’s familiarity with a control device could affect the subjective and objective performance of the device. The main results indicate that the TouchPad and MicroJoystick are the best candidate-devices for use in a powered laparoscopic tool. The article also provides support for utilizing the new measure target re-entry when comparing control performance. Although studied in the application of laparoscopic surgery, the results can be generalized for the design of any hand-held device in which the speed and accuracy of the control device is critical. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Laparoscopic surgery Ergonomic design Human computer interaction Control design TouchPad MicroJoystick Minimal invasive surgery Laparoscopic tool Fitts’ law

1. Introduction Minimally-invasive surgery (MIS) or laparoscopic surgery is an advanced surgical technique that is performed with the assistance of a video (endoscopic) camera and several thin tools that resemble children’s scissors attached to a long thin shaft. During MIS procedures, small incisions (up to half an inch) are made in the body and plastic tubes (ports or trocars) are placed in these incisions to allow tools to be inserted into the abdomen (Berguer, 1998). Since MIS requires only small incisions, it reduces trauma to the body, shortens recovery time and reduces the risk of infection (Cuschieri, 1995). Although MIS has revolutionized the medical field, surgeons are limited by the abilities of their instruments since they cannot directly touch or see the target inside of the body. Surgeons are required to perform a variety of tasks in MIS procedures such as grasping, dissecting, cauterizing, and suturing with various hand tools which greatly influences the finger, hand, * Corresponding author at: 175 Nebraska Hall, Lincoln, NE 68588-0518, USA. Tel.: þ1 402 472 2394. E-mail address: [email protected] (M.S. Hallbeck).

wrist and arm posture adopted during surgery (Trejo et al., 2007) (Fig. 1A). In fact, one of the leading causes of surgeon postoperation pain or numbness is the non-neutral postures adopted during MIS procedures (Graves et al., 1994; Crombie and Graves, 1996; Berguer et al., 1998; Berguer et al., 1999; van Veelan and Meijer, 1999; Emam et al., 2001; Done´ et al., 2004a; van Veelan et al., 2004) (Fig. 2C). In addition, the current scissor-like handle design (Figs. 1A and 2) has also been shown to increase surgeon fatigue, discomfort and paresthesias in the fingers (Berguer, 1998; DiMartino et al., 2004). In an effort to remedy the deficiency of current non-powered laparoscopic instruments, a new tool was developed using usercentered design principles (DiMartino et al., 2004; Done´ et al., 2004b; Judkins et al., 2004; Trejo et al., 2005; Hallbeck and Oleynikov, 2006) (Fig. 1B). The IntuitoolÔ includes an ergonomically designed handle and a redesigned grasper actuation mechanism in order to create a more comfortable and intuitive handle/tool interface. In addition, an articulating end-effector (controlled by a trackball) was also added allowing surgeons to comfortably articulate the device by manipulating the end-effector with a trackball (Fig. 1B). This addition allows surgeons to make fine

0003-6870/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apergo.2009.04.001

Please cite this article in press as: Herring, S.R., et al., Evaluation of four cursor control devices during a target acquisition task for laparoscopic tool control, Applied Ergonomics (2009), doi:10.1016/j.apergo.2009.04.001

ARTICLE IN PRESS 2

S.R. Herring et al. / Applied Ergonomics xxx (2009) 1–11

Fig. 1. (A) The current laparoscopic tool. (B) The IntuitoolÔ prototype with articulating end-effector.

motor movements with the tool without altering their posture (DiMartino et al., 2004). In order to determine the effectiveness of the newly developed handle/grasper mechanism and end-effector, an evaluation of the IntuitoolÔ prototype was conducted (Trejo et al., 2005). Fiftyeight percent of responding laparoscopic surgeons believed the IntuitoolÔ would relieve hand/wrist pain and 53% believed the tool would reduce hand/wrist stiffness, confirming that the redesign using user-centered design principles was successful. In addition, 92% of respondents believed the addition of the articulating tip would be useful. These data show that the ergonomic handle design, the grasper mechanism and the end-effector articulation are all keys to a well-designed non-powered (‘‘cold’’), laparoscopic tool. The next step in the design process is to combine multiple laparoscopic instrument functions, such as grasping and cauterizing, into one unified tool. In order to accomplish this feat, an electrically powered (‘‘hot’’) tool must be developed. However, since no powered laparoscopic tools are equipped with an articulating end-effector, current control devices on the market must be adequately tested to see how they perform in this particular operation. In particular, we must ensure that the performance of the device used to control the end-effector is not affected by the distance traveled or the width of the target (which can range from an artery to a thin tissue) because the overall utility of the device would be significantly affected. There are many benefits associated with utilizing a powered control device for this operation. Specifically, the use of a powered control device could reduce the pain incurred during surgery by transferring the power source for the end-effector from the thumb or index finger to the powered device. However, before we can draw such a conclusion, we first must compare the performance of the control devices. The best available instrument for comparing pointing device performance is Fitts’ law. Fitts’ law is an information processing

model of human psychomotor behavior which involves a pointing task in which researchers record the time it takes for a participant to position a cursor and select a target (Fitts’, 1954). Card et al. (1978) was the first study to study cursor control devices using Fitts’ law. Subsequent studies focused on comparing the performance of different control devices such as a mouse, trackball, isometric joystick, step keys, text keys and TouchPads (ex. MacKenzie et al., 1991). These studies used a discrete task, where participants began with the pointing device at the starting position and moved to a target. This is similar to the skills required for laparoscopic surgery as the cursor movement can be used to simulate the movement of the articulating end-effector in the body (Fig. 3). The most common evaluation measurements for Fitts’-type tasks are speed (movement time) and accuracy (error rate, the percentage of selections with the pointer outside the target). In fact, the ISO standard (ISO 9241; ISO ISO/TC 159/SC4/WG3, 1998), ‘‘Requirements for non-keyboard input devices’’, proposes only one performance measurement, throughput (in bits per second, bps), that is a composite measure derived from both the speed and accuracy in responses. These measures are typically analyzed over a variety of task or device conditions. More recently, however, researchers have developed new accuracy measures in order to elicit differences among devices in precision pointing tasks (MacKenzie et al., 2001). Unlike the standard measurements (movement time and error rate), which are based on a single measurement per trial, these proposed measurements capture the behavior during a trial. A recent exploratory study utilizing these measures found that of the seven measures developed (target reentry, task axis crossing, movement direction change, orthogonal direction change, movement variability, movement error, and movement offset) only two of them made a significant contribution to the prediction of throughput – target re-entry and movement offset (MacKenzie et al., 2001). Of these two, target re-entry explained about 41% of the variance.

Fig. 2. (A) The hand position incurred using the current laparoscopic tool. (B) The pressure point encountered when using the current laparoscopic tool. (C) The awkward hand postures incurred when using the current tool.


ARTICLE IN PRESS S.R. Herring et al. / Applied Ergonomics xxx (2009) 1–11

3

Fig. 3. (A) Experimental setup and upper extremity posture during task. (B) The movement of the end-effector with respect to the movement of the cursor on the monitor.

The aim of the present investigation is three-fold. First, the study aims to compare the performance of four electronic control devices by evaluating them in a laparoscopic surgery mock-up using a discrete Fitts’ task to simulate the movement of the laparoscopic end-effector inside the body. Second, the study tries to provide new insights into the relationship between the user’s familiarity with a control device and their subjective evaluation of its performance (ease of use and comfort) as well as the objective performance measures (error rate, target re-entry and throughput). Finally, the study aims to provide a better understanding of the new measure target re-entry as utilized in control device evaluation. Although studied in the application of laparoscopic surgery the results can be generalized for the design of any control mechanism imbedded in a hand-held device in which the speed and accuracy of the reply is critical (i.e. video game control design). 2. Methods

cm and 60.2 (6.2) kg for females and 22.2 (2.5) years, 177.5 (11.6) cm and 83.1 (12.7) kg for males, respectively. Students were used in the experiment because they have the same age range and basic background as novice surgeons. Since the control mechanism is novel, no surgeon or medical student has used or seen any tool of its kind, making students a good fit for the experiment. Finally, literature shows no gender differences in laparoscopic surgical skills, visual–spatial performance, tracking performance and applied visual–spatial performance (Madan et al., 2005; Kass et al., 1998; Stamper et al., 1997; Mayes and Jahonda, 1988). In addition, Kass et al. (1998) found that a brief training session is all that is required to increase mental rotation performance of women to the level of men on a specific task. Since a training session was employed in this study and because previous studies have found no gender differences in the tasks involved in this study, the experimenters were able to use an unequal number of female and male participants.

2.1. Participants

2.2. Apparatus

Seventeen (13 males and four females) right-handed students, with no previous experience in laparoscopic surgery techniques, participated in this study. Participants self reported no history of arm injury or illness. All participants were regular users of either a PC or laptop. All participants used input devices such as those employed in this study on a regular basis. The average (standard deviation) age, height, and weight were 22.8 (2.5) years, 157.7 (17.8)

To simulate laparoscopic surgery conditions, an adjustable surgical bed with a simulated abdomen was used. The simulated laparoscopic surgery set-up was similar to an insufflated 95th percentile male abdomen (40 cm diameter with neoprene ‘‘skin’’). The set-up used was similar to previous studies (Trejo et al., 2005b; Trejo et al., 2007). The simulated abdomen contained a 17 inch LCD monitor, which was connected to a personal computer. The monitor

Fig. 4. From left to right: (1) four input devices used, from left to right, MicroJoystick, Mouse button module, TouchPad, and MiniJoystick module; (2) hand position; (3) hand position and the actuation button used for the experiment (located under index finger).


ARTICLE IN PRESS 4


Fig. 5. The questionnaires used in the experiment.



was in a fixed position for all trials (0 from horizontal). A 19 inch LCD screen was mounted on an adjustable arm, in front of the simulated abdomen, to provide visual feedback to the participants via an endoscopic camera (Olympus WA50295L), inserted at the center of the simulated abdomen to view the 17 inch LCD monitor display. The camera was in a fixed position so participants could see all targets on the 19 inch screen. The experimental setting was carefully designed to simulate an operating room set-up (Fig. 3). The previously designed Intuitool handle (DiMartino et al., 2004; Trejo et al., 2005; Trejo et al., 2006) was replicated (for a total of four handles) using stereolithography and four electronic cursor control devices were mounted inside each handle (Fig. 4). The four electronic cursor control devices evaluated were an Easy CatÒ TouchPad (Cirque with GlidePointÒ Technology), a high-precision MicroJoystick (Interlink Electronics v. 0.2), and two MicroModuleÔ controllers, the MiniJoystick Module and the Mouse Button Module (Interlink Technologies v. 5.0). An actuation button was added underneath the handle in a position similar to the grasper mechanism lock location, so that each participant could indicate when they reached the target. Each handle was attached to the 95th percentile abdomen via a metal rod which simulated a tool shaft. The handles were presented at the participant’s elbow height for each pointing task. The speed of the input devices was standardized to the 7th position, on an 11 point slow to fast scale, using the Microsoft WindowsÔ operating system. End point data was also collected to determine how far the cursor was from the target. 2.3. Questionnaires Two questionnaires were used in this experiment. The first questionnaire was administered to determine if the participant used a computer, and if so, the type of pointing device the participant used, how many hours a week the participant used a computer (if applicable), and the familiarity of the participant with each of the pointing devices tested. The second questionnaire was adapted from the ISO 9241-0:2000(E) standard for assessment of comfort. An independent rating scale was used to assess the impression of each input being tested. The survey questions can be seen in Fig. 5. All questions, with the exception of the first 3, used a verbally anchored 5-point Likert scale. 2.4. Procedure After the participant was informed of the purpose of the investigation and the experimental methods, an informed consent was signed. They were then asked to wear a hypo-allergenic (Kimberly–Clark SAFESKIN purple nitrile powder-free) surgical/ exam glove. While holding one of the four handles (all mounted at

5

the same height) the surgical bed height was adjusted until a 90 included elbow angle was reached. The participant was then asked to hold the handle by placing the thumb in the middle of the device, the index finger on the button underneath the input device, placing the remaining fingers underneath the handle as shown in Fig. 4. The participants were required to keep this hand position during the entire trial but were allowed to take breaks if necessary. The participant’s body posture was to be relaxed, keeping the wrist in a neutral position, and they were asked to maintain a 90 elbow angle. The experimental set-up can be seen in Fig. 3. Prior to testing, the task was demonstrated, and a sequence of practice trials was given. In order to compare the devices, a simple pointing task was tested using multiple target widths and circle diameters (Fig. 6). The task was the simple multidirectional pointselect task from ISO 9241-9. Participants were instructed to move to each target as quickly and accurately as possible, trying to maintain a straight path from the starting point to each target. The cursor used in the experiment represented the tip of the articulating endeffector in a laparoscopic surgical tool; the tip of the end-effector would mimic the motion of the cursor as it moved from target to target (Fig. 3). Each participant was allowed to learn how to use the four control devices until they reported that they felt they had significant preparation time and felt they did not show any additional signs of improvement. Although participants were not tested for significant improvements at the end of the practice session, the participant had to demonstrate to the researchers their ability to successfully manipulate each control device (no errors and good control) for multiple consecutive trials before the practice session was ended. On average, the practice session took approximately 10 min; however, participants were given as much practice time as needed. After the participants reported that they were acquainted with the devices, they were asked to complete a seven question survey, questionnaire 1, on their computer usage and familiarity with the control devices being tested. This gave the participants approximately a 5 min break before the study began, which minimized any effects due to fatigue incurred during the practice trials. All participants performed multiple trials with different circle diameters and target widths to ensure the performance of the device was not affected by the distance or width of the target. There were 16 circular targets arranged in a circular layout. There were three diameters of the circle (2, 4, and 8 cm) and three widths of the targets (0.27, 0.54 and 1.07 cm). The 6 layout combinations (cm) were (circle diameter, target width): (2, 0.27), (4, 0.27), (4, 0.54), (8, 0.27), (8, 0.54) and (8, 1.07). The diameter of the circle layout and target were chosen so that the combinations could be seen in the field of view for laparoscopic surgery as viewed by an endoscopic camera and so that the targets represented a variety of index of

Fig. 6. (A) Sequence task showing a sequence of 15 target selections (adapted from MacKenzie et al., 2001). (B) Circle diameter and target width.


ARTICLE IN PRESS 6


Table 1 Index of difficulty for each target width/ target distance combination. Target width (cm)

Target distance (cm)

ID

0.27 0.27 0.54 0.27 0.54 1.07

2 4 4 8 8 8

3.07 3.98 3.07 4.93 3.98 3.08

IDe ¼ log2

D þ1 We

The term is the effective index of difficulty (in bits). It is calculated from D, the distance to the target, and, the effective target width (which is the width of the distribution of selection coordinates computed over a sequence of trials) (MacKenzie et al., 2001). We is calculated as:

We ¼ 4:133 SDx difficulty (ID) levels. The ID for each of the circle diameter and target width combinations can be seen in Table 1. The starting point for each trial was at the center of the top target (black cross) of the layout circle and the trial began when the participant moved the cursor. The next selection was on the opposite side of the layout circle (similar to MacKenzie et al., 2001) and was presented in blue to alleviate confusion between the target and the cursor. When the target and cursor were properly aligned, the cursor turned red. To record end point data a trigger button (Fig. 4) was placed underneath each handle, in the same location as the lock for the grasper jaws (Figs. 1 and 4). Once the participant had reached the target destination, the 2nd digit was used to activate the button which concluded each trial. When the trial was complete, the target became the starting point for the next subsequent trial (Fig. 6). The cursor control devices were presented in a random order to control learning effects associated with the sequential use of the devices. The experiment proceeded until all combinations of layout circles and target diameters were completed, six per control device. Once the participant completed all trials for a cursor control device, the participant was asked to complete the 13 question survey on the performance of the input device tested (questionnaire 2). This gave each subject approximately a 5 min break which minimized any effects due to fatigue incurred during the previous trial. The study took about 1 h. 2.5. Objective and subjective data 2.5.1. Objective data The dependent variables of target re-entry (TRE), error rate (ER), and throughput were calculated in order to compare devices. TRE is defined as the number of times the pointer enters the target region, leaves, and then re-enters the target region. If this behavior is recorded twice in a sequence of 10 trials, TRE is reported as 0.2 per trial (Fig. 7). ER is the number of selections with the pointer outside the target. It is calculated as a percentage for each trial condition. Throughput is a composite measure derived from both the speed and accuracy in the responses and is calculated according to the ISO 9241 standard entitled ‘‘Ergonomic design for office work with visual display terminals (VDTs).’’ Specifically throughput is:

Throughput ¼

IDe MT

(1)

(2)

(3)

where SDx is the standard deviation in the selection coordinates measured along the axis of approach to the target and 4.133 represents the 96% confidence interval and We reflects the spatial variability in the sequence of trials. The independent variables tested in this experiment are control device, index of difficulty, target width and circle diameter. The index of difficulty is a measure, in bits, of the user precision required in a task. It was calculated according to the ISO 9241-9 standard for selection, pointing and dragging tasks:

ID ¼ log2

dþw w

(4)

where d is the distance of the movement to the target; w is the target width along the approach axis for selection, pointing or dragging tasks, and perpendicular for tracing tasks. This resulted in a 4 4 3 3 factorial design with blocking on participants nested within gender. The factors and levels were as follows: -

-

Control device (TouchPad, MicroJoystick, MiniJoystick and Mouse Button Module) Index of difficulty: 3.07, 3.08, 3.98, and 4.93 Diameter of circle: 2, 4 and 8 cm Width of target: 0.27. 0.54 and 1.07 cm

The diameter of the circle and width of the target are important factors to consider in this experiment because of the repercussion on laparoscopic surgery. In particular, since the device used to control the articulating end-effector should not be affected by the distance traveled or the width of the target metrics, it is important to test a variety of circle diameters and target widths in order to adequately compare performance. Additionally, the metrics of TRE, error rate and throughput allow for an accurate comparison of how well the devices perform with the variation of circle diameter and target width. In addition, these factors are important because they will help address any limitations in using an electronic control device in a laparoscopic surgical tool over a manually manipulated control device. SAS (v. 9.1) was used to analyze the data. Tukey post-hoc tests were performed on significant effects and simple effects F-tests were performed on all significant interactions. An alpha level of 0.05 was used in the experiment. Finally, in order to determine if the new accuracy measure (TRE) had a causal relationship to throughput (as found previously in MacKenzie et al., 2001), an adjusted participant and control device partial correlation was performed between

where

Fig. 7. Example of a target re-entry. The target enters the target region, leaves, and then re-enters the target region.



7

throughput and TRE, as well as control device and index of difficulty using SPSS (v. 14.0). 2.5.2. Subjective data Ordinal data was collected throughout the questionnaires; therefore, a Wilcoxson signed rank test was used to analyze the hypothesis tests using Minitab 14 (Minitab, Inc). The level of significance for all statistical tests was 0.05. Questionnaire 1 was divided into two parts; the first three questions pertained to the participant’s computer usage and the second set of questions (4–7) dealt with the participant’s familiarity with the devices being studied. These questions were analyzed using descriptive statistics. The null hypothesis used for the analysis of Questionnaire 2 was dependent on the question being presented. The null hypothesis for the first six questions was that the user did not experience the problems listed (median ¼ 0) versus the alternative that the user experienced a slight or substantial problem (median > 0 or median < 0) for each queried problem. For the second grouping of questions, 7–11 (Fig. 5), the null hypothesis was that the user did not experienced fatigue (median ¼ 0) versus the alternative that the user experienced some fatigue (median < 0). The last two questions, 12 and 13, were analyzed with the null hypothesis that the device was neither comfortable nor uncomfortable or neither difficult nor easy to use (median ¼ 0) versus the alternative that the device was comfortable or easy to use (median < 0), respectively. Finally, the participant and control device partial correlation between familiarity with each device (questions 4–7 of questionnaire 1), general comfort and overall satisfaction (questions 12 and 13 of questionnaire 2) was computed using SPSS (v. 14.0). 3. Results 3.1. Objective data 3.1.1. Traditional measures As previously discussed, the traditional measures for evaluating a Fitts’-type task are throughput and error rate. Several factors significantly influenced these measures. The width of the target was a significant factor for both throughput (F3,32 ¼ 7.44, p ¼ 0.0022) and error rate (F3,32 ¼ 4.10, p ¼ 0.0259). Post-hoc tests revealed that the 0.27 cm target width had a significantly lower throughput than both the 1.07 cm and 0.54 cm target width and the

Fig. 8. Interaction of id*device for target re-entry with standard error.

0.27 cm target width had a significantly higher error rate than the target width of 1.07 cm. The effect of control device was clearly significant for throughput (F3,48 ¼ 6.31 p ¼ 0.0011) but not error rate (F3,48 ¼ 2.02, p ¼ 0.1231). The MicroJoystick had a significantly higher throughput (1.39 bps) than all other control devices tested. On the contrary, the effect of ID was significant for error rate (F3,48 ¼ 3.32, p ¼ 0.273) but not for throughput (F3,48 ¼ 2.50, p ¼ 0.0710). The ID of 4.93 (circle diameter of 8 cm, target width of 0.27 cm) had a statistically higher error rate than the ID of 3.08 (circle diameter 8 cm, target width 1.07 cm). All other ID values were not significantly different for error rate. The diameter of the circle was not a significant factor for throughput (F2,48 ¼ 2.20, p ¼ 0.1274) or error rate (F2,48 ¼ 4.10, p ¼ 0.0259) and there were no significant interactions. The means and standard deviation for all variables can be seen in Table 2. 3.1.2. Target re-entry The width of the target was also a significant factor for TRE (F3,32 ¼ 19.74, p < 0.0001). Tukey post-hoc tests revealed that all target widths were significantly different with the target width of 1.07 having the lowest TRE and the target width of 0.27 having the highest TRE.

Table 2 The mean and standard deviation for each factor. Factor

Variable

Error rate (%) Significance

Device MiniJoystick TouchPad Mouse Button MicroJoystick ID

0.1231 2.0 1.2 3.7 1.5 0.0273

3.07 3.08 3.98 4.94 Target width (cm)

Throughput (bps) Mean

SD

0). The accuracy of the control devices was tested to see if the participants felt the device performed accurately during the experiment. The participants reported that it was difficult to be accurate using the MiniJoystick Module (median > 0), it was neither difficult nor easy to be accurate when using the Mouse

Table 4 The mean and standard deviation for each subjective measure.

Actuation force

Smoothness

Mental effort

Physical effort

Accurate pointing

Operation speed

General comfort

Ease of use

Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD

MiniJoystick

TouchPad

Mouse Button Module

MicroJoystick

Too high 0.76 0.83 Just right 0.41 1.18 Too high 0.59 0.62 Too high 0.59 0.80 Difficult 0.71 1.26 Just right 0.35 1.06 OK 0.41 1.12 OK 0.35 0.93

Just right 0.06 1.09 Too rough 0.88 1.32 Just right 0.18 0.95 Just right 0.35 1.06 Easy 0.94 1.30 Too slow 1.00 1.12 OK 0.06 1.14 OK 0.24 1.30

Too high 0.71 0.69 Too rough 0.65 0.86 Just right 0.41 0.62 Just right 0.47 0.80 Neither 0.65 1.22 Just right 0.06 0.83 OK 0.18 1.19 OK 0.12 1.05

Just right 0.12 0.70 Too smooth 0.59 0.80 Just right 0.29 0.59 Just right 0.18 0.64 Easy 1.18 1.01 Too fast 0.47 0.62 Comfortable 1.18 0.81 Easy to use 1.71 0.47



9

Table 5 The mean and standard deviation for fatigue encountered in each area queried.

Finger fatigue

Wrist fatigue

Arm fatigue

Shoulder fatigue

Neck fatigue

Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD Representation Mean SD

MiniJoystick

TouchPad

Mouse Button Module

MicroJoystick

Slight 1.53 1.37 No to slight 1.06 1.34 No to slight 1.00 1.06 No to slight 1.00 1.00 No to slight 0.94 1.09



No to slight 1.24 1.03 No to slight 0.88 0.86 No to slight 0.82 0.88 No to slight 1.06 0.97 No to slight 0.88 1.05

Button Module (median ¼ 0) and it was easy to be accurate with the TouchPad and the MicroJoystick (median < 0). The operation speed of the device was explored to see if the control device was operating at the ideal speed. If the device was found to be too fast or too slow, the experimenters would know which way to manipulate the speed to make the control device more ideal for the user. The participants found that the Mouse Button and MiniJoystick Module had the correct operational speed (median ¼ 0) while the TouchPad was too slow (median < 0) and the MicroJoystick was too fast (median > 0). It was also extremely important to determine the amount of perceived fatigue encountered during the experiment as this was one of the variables to be minimized. All participants, on average, reported some fatigue in every area (finger, wrist, arm, shoulder and neck) (median > 0), see Table 5. In order to determine the intensity of the fatigue, additional analysis was conducted with the null hypothesis that there was moderate fatigue (median ¼ 2) versus the alternative that there was no fatigue to moderate fatigue (median < 2). The reported fatigue for the MicroJoystick indicates that participants experienced no fatigue to moderate fatigue in all areas (median < 2). The Mouse Button Module, TouchPad, and MiniJoystick Module all show fatigue ranging from none to moderate in the wrist, arm, shoulder, and neck as well (median < 2), but moderate fatigue in the finger (median ¼ 2). The final two questions pertained to the participants’ overall evaluation of the devices: the comfort of the device and the ease of use of the device. The Mouse Button Module, TouchPad and MiniJoystick Module’s results show satisfaction with the comfort of the device; the device was neither comfortable nor uncomfortable (median ¼ 0). The MicroJoystick, however, shows greater than average comfort (median > 0). For the final question, how easy the device was to use, the Mouse Button Module, TouchPad and MiniJoystick Module were all found to be satisfactory: neither difficult nor easy to use (median ¼ 0). However, the MicroJoystick was found to be easy to use (median > 0).

Finally, no significant correlation was found between the familiarity with the device and the overall comfort, or overall satisfaction of each device. However, there was a correlation between the overall comfort of the device and the overall satisfaction with the device (p ¼ 0.000, r2 ¼ 0.558). This can be seen in Table 6. 4. Discussion MIS surgeries have been shown to cause significant post-operation discomfort to surgeons (Graves et al., 1994; Crombie and Graves, 1996; Berguer et al., 1998; Berguer et al., 1999; van Veelan and Meijer, 1999; Emam et al., 2001; Done´ et al., 2004a; van Veelan et al., 2004). In an effort to remedy the deficiency of current laparoscopic instruments, a new tool (the IntuitoolÔ) was developed. This tool includes a new handle shape, a new grasper mechanism and an articulating end-effector (DiMartino et al., 2004; Done´ et al., 2004b; Judkins et al., 2004; Trejo et al., 2005; Hallbeck and Oleynikov, 2006). The addition of the end-effector allows surgeons to make fine motor movements with the tool without altering their posture (DiMartino et al., 2004). The purpose of the current study was to further develop this tool by creating an electrically powered tool which has multiple functionalities (grasping, suturing, cauterizing, etc.). In laparoscopic surgery, it is important to perform operations as quickly and accurately as possible to alleviate injuries and increase the chance of a successful surgery. For this reason, the current study was developed to compare four different control devices for use in a powered laparoscopic tool. Since minimal operation time and errors are ideal, the variables throughput, error rate, and the new measure target re-entry were tested in order to accurately compare these devices. Since the device will be used to control the articulating end-effector of a laparoscopic tool we must ensure that the performance of the device is not influenced by the distance traveled or the width of the target as this could be detrimental to the

Table 6 Partial correlation between familiarity with each device, comfort rating and overall evaluation of the device. Control variables Participant and device

Familiarity Familiarity

Comfort

Overall

Correlation Significance (1-tailed) df Correlation Significance (1-tailed) df Correlation Significance (1-tailed) df

1.000 – 0 0.140 0.130 64 0.084 0.252 64

Comfort

Overall

0.140 0.130 64 1.000 – 0 0.558 0.000 64

0.084 0.252 64 0.558 0.000 64 1.000 – 0


ARTICLE IN PRESS 10


effectiveness of the devices. In order to combat this effect, multiple target widths and circle diameters were tested to ensure that the performance of the devices was not affected by these alterations. It was also important to determine the user’s satisfaction and opinion of each of the control devices in order to optimize the user’s perceived comfort. The results show that the MicroJoystick has the highest throughput (1.39 bps) and the user’s rated it to be both fast and accurate. The MicroJoystick, along with the TouchPad, also has the lowest error rate and fewest target re-entries (although not significant for error rate). Conversely, the MiniJoystick and Mouse Button had the highest ER and TRE and the participants ranked these devices as being ‘‘difficult’’ to use. This not only shows that the MicroJoystick and TouchPad had a superior performance (both qualitatively and quantitatively) to both modules but it also shows that the participants were rather accurate in their estimation of the accuracy and speed of the devices (as throughput is a composite measure of both variables and ER and TRE taking into account the accuracy of the control devices). The study also explored the possible correlation between the participant’s familiarity with each control device and both their subjective and objective evaluations. The first hypothesis was that the participant’s familiarity with the control device would impact their impression of how comfortable and/or easy a control device was to use. If such a correlation existed, it would suggest that designers should consider users previous knowledge of a control device before making a selection. However, the results show there was no significant correlation between the participant’s familiarity with the control devices tested and their comfort/ease of use rating. This means the user’s rating of comfort/ease of use of control mechanisms will not be influenced by the user’s familiarity with the control. The second hypothesis was that if the participants were familiar with a control device, their performance would be significantly better (i.e. low error rate and TRE and high throughput). However, the results show that although participants were least familiar with the MicroJoystick, this device had the highest throughput and the lowest (statistically significant) TRE (although statistically the same as the TouchPad). Additionally, even though the four control devices were not statistically different for error rate, the TouchPad and MicroJoystick had the lowest error rates as well. This finding could be attributed to the control devices being used in a different manner than their normal (familiar) utilization (imbedded inside a laptop or pointing device case). This result is very interesting and has great implications for the design of a powered hand-held tools operated by an electric control device, as it shows that user’s opinions and performance with a device is not affected by their familiarity with the device. Fatigue was also an important variable to consider when comparing control devices. Current laparoscopic instruments compel the user to hold awkward positions which cause fatigue, pressure points and nerve lesions (Graves et al., 1994; Crombie and Graves, 1996; Berguer et al., 1998; Berguer et al., 1999; van Veelan and Meijer, 1999; Emam et al., 2001; Done´ et al., 2004a; van Veelan et al., 2004). Trejo et al. (2005) found that 92% of responding surgeons believed the addition of the articulating tip would be useful and 44% believed that the new handle design would reduce numbness of the finger and/or the thumb. Therefore, it was important to test the current control devices, imbedded inside the IntuiToolÔ handle, for perceived fatigue. Although the subjective data show that all of the devices were relatively easy to use, all devices tested created some amount of subjective fatigue, in a study that took only 1 h. Even though participants reported the least amount of fatigue with the MicroJoystick, they still reported some fatigue in every area tested. The TouchPad also showed less than moderate fatigue in all areas with the exception of the finger.

There are several possible explanations for these findings. First, the participant responses could have been affected by the participants’ avoidance of extreme responses on questionnaires as participants tend to answer in the median of a scale (central tendency bias). Since the scale utilized was a 5-point Likert scale with the response ‘no fatigue’ on the extreme end of the scale, participants may have been biased with their replies because they wanted to avoid the extreme response. Another possible explanation could be that these two devices (MicroJoystick and TouchPad) were not subjectively deemed to be at the ideal operating speed despite their standardization; the TouchPad was shown to operate too slowly and the MicroJoystick was reported to operate too quickly. This implies that the devices could show further improvements in throughput and TRE if their speed were optimized. Future studies should further explore perceived fatigue by not only alter the scale utilized but also quantify the fatigue through other measures (i.e. EMG sensors) as well as determine an ideal operating speed for the candidate device. In addition to the device being tested, the index of difficulty can also play a crucial role in the target re-entry and error rate. The ID of 3.08 had the lowest error rate and TRE. Although it might seem counterintuitive that the ID of 3.08 would outperform the ID of 3.07, it can be logically explained. The ID of 3.08 was comprised of a circle diameter of 8 cm and a target width of 1.07 cm; this was the only ID were the target width of 1.07 cm was used. When looking at the results from the target width for both TRE and ER, it can be seen that the target width of 1.07 outperformed the target width of 0.54, although they were not statistically different for ER. Since the Fitts’ Law model is a function of both the distance to the target and the size of the target and since the target distance was not a significant factor for TRE or ER, it can be concluded that the difference in ID performance is due to the different target widths. This means that the larger target width allowed the participant to move the cursor around inside the target without leaving the area creating less target re-entries. It also allowed the participant to more easily get the cursor inside the target, reducing the ER. The second goal of this study was to provide insights into the potential benefits of using the new measure of target re-entry in hand-held tool evaluation. The current ISO standard (ISO 9241), requirements for non-keyboard input devices, proposes only one performance measurement, throughput. However, a recent study by MacKenzie et al. (2001) derived seven new accuracy measures of which the measure target re-entry explained about 41% of the variance in the data. MacKenzie et al. (2001) also found target reentry to be negatively correlated with throughput (p < 0.0001, r2 ¼ 0.82). This study found the same result with a lower correlation coefficient (p ¼ 0.000, r2 ¼ 0.313). Therefore, it can be concluded that as the throughout increases, target re-entry decreases, but weakly. Although this is only a weak relationship, this type of metric is extremely beneficial for control comparison as it allows for a better understand of the behavior during a trial. This is particularly important in the design of hand-held tools which require accurate end point precision such as in the specific application of laparoscopic tool design or in other applications such as game control design. 5. Conclusion The main goal of this investigation was to determine the best input device for a powered articulating laparoscopic surgery tool. User-centered design principles specify that designers should create instruments that are easy to use, reduce awkward postures and eliminate fatigue (reduce surgery time). The input devices that best achieve these goals are the TouchPad and MicroJoystick, based on both the subjective and objective usability measures employed



in this study. Although a small participant sample was used in this experiment (N ¼ 17), the results can be used to refine further investigations. The TouchPad and the MicroJoystick should be further investigated as an end-effector manipulation control or input device. In addition, future studies should also test the operational speed which could have an effect on the perceived fatigue of the participants. Finally, in order to show that the new device can be used in different ways without creating awkward postures, future studies should explore different positions of the hand when manipulating these devices. Additional findings from this study show that user’s preconceptions or familiarity with control devices have little effect on both the qualitative and quantitative evaluations of a device. The new measure of target re-entry also showed to be a useful evaluation tool. The study showed that there is a strong correlation between target re-entry and throughput. This new metric not only provides another performance measure, but it also allows ergonomic designers to understand the controls performance during a trial which is of particular importance when the controller’s accuracy and precision is essential to the utility of the device. Finally, although studied in the specific application of laparoscopic surgery, the results found here can be generalized to the design of any hand-held device in which the speed and accuracy of the reply is critical, such as in the design of neutron detectors or game controllers. Acknowledgments The authors would like to acknowledge the Nebraska Research Initiative for partial support of this study and Justin Rousek for his help in conducting this study. References Berguer, R., 1998. Surgical technology and the ergonomics of laparoscopic instruments. Surgical Endoscopy 12, 458–462. Berguer, R., Gerber, S., Kilpatrick, G., Beckley, D., 1998. An ergonomic comparison of in-line vs. pistol-grip handle configuration in a laparoscopic grasper. Surgical Endoscopy 12, 805–808. Berguer, R., Forkey, D.L., Smith, W.D., 1999. Ergonomic problems associated with laparoscopic surgery. Surgical Endoscopy 13 (5), 466–468. Card, S.K., English, W.K., Burr, B.J., 1978. Evaluation of mouse, rate-controlled isometric joystick, step keys, and text keys for text selection on a CRT. Ergonomics 21 (8), 601–613. Crombie, N.A.M., Graves, R.J., 1996. Ergonomics of keyhole surgical instruments – patient friendly, but surgeon unfriendly. In: Robertson, S. (Ed.), Contemporary Ergonomics. Taylor & Francis, London, pp. 385–390. Cuschieri, W.A., 1995. Minimal access surgery: tribulations and expectations. The American Journal of Surgery 169, 9–19. DiMartino, A., Done´, K., Judkins, T., Morse, J., Melander, J., Olenikov, D., Hallbeck, M.S., 2004. Ergonomic laparoscopic tool handle design. Proceedings of the Human Factors and Ergonomic Society 48th Annual Meeting, New Orleans, LA, pp. 1354–1358.

11

Done´, D., DiMartino, A., Judkins, T., Hallbeck, M.S., 2004a. Evaluation of laparoscopic tools for usability and comfort. Proceedings of the Human Factors and Ergonomic Society 48th Annual Meeting, New Orleans, LA, pp. 1359–1361. Done´, D., DiMartino, A., Judkins, T., Hallbeck, M.S. Oleynikov, D., 2004b. Evaluation of laparoscopic tools for functionality, usability and comfort. Proceedings of the Human Factors and Ergonomic Society 48th Annual Meeting, New Orleans, LA. Emam, T.A., Frank, T.D., Hanna, G.N., Cuschieri, A., 2001. Influence of handle design on surgeon’s upper limb movements, muscle recruitment, and fatigue during endoscopic suturing. Surgical Endoscopy 15, 667–672. Fitts, P.M., 1954. The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology 47, 381–391. Graves, R.J., Gardner, J.M., Anderson, M., Seaton, R., Ross, J., Porter, R., 1994. The Development of an Ergonomic Methodology for the Evaluation of Operating Table Systems. CSA final report. University of Aberdeen, Aberdeen. Hallbeck, M.S., Oleynikov, D., 2006. Hands on. Surgeon-centered laparoscopic tool design. Pikaar, R.N., Koningsveld, E.A.P., Settels, P.J.M. (Eds.), Triennial Congress of the International Ergonomics Association. Elsevier Ltd., pp. 1777-1781 (ISSN 0003-6870). ISO ISO/TC 159/SC4/WG3, 1998. Ergonomic requirements for office work with visual display terminals (VDT’s) – Part 9 – Requirements for non-keyboard input devices, International Organization for Standardization, May 25, 1998. Judkins, T.N., DiMartino, A., Done´, K., Hallbeck, M.S., Oleynikov, D., 2004. Effects of handle design and target location on wrist posture during aiming with a laparoscopic tool. Proceedings of the Human Factors and Ergonomic Society 48th Annual Meeting, New Orleans, LA, pp. 1464–1468. Kass, S.J., Ahlers, R.H., Dugger, M., 1998. Eliminating gender differences through practice in applied visual and spatial task. Human Performance 11 (4), 337–349. MacKenzie, I.S., Kauppinen, T., Silfverberg, M., 2001. Accuracy measures for evaluating computer pointing devices. Computer Human Interaction 3 (1), 9–16. Mackenzie, I.S., Sellen, A., Buxton, W., 1991. A comparison of cursor control devices in elemental pointing and dragging tasks. In: Proceedings of the ACM conference on human factors in computing systems, 161–166. Madan, A.K., Frantzides, C.T., Park, W.C., Tebbit, C.L., Kumari, N.V.A., O’Leary, P.J., 2005. Predicting Baseline Laparoscopic Surgery Skills. Surgical Endoscopy 19 (1), 101–104. Mayes, J.T., Jahonda, G., 1988. Patterns of visual–spatial performance and ‘spatial ability’: dissociation of ethnic and sex differences. British Journal of Psychology 79, 105–119. Stamper, D.A., Lund, D.J., Molchany, J.W., Stuck, B.E., 1997. Laboratory simulator and field pursuit tracking performance with females and males in the presence of laser glare. Aviation. Space and Environmental Medicine 68 (7), 580–587. Trejo, A.E., Jung, M.-C., Oleynikov, D., Hallbeck, M.S., 2005. Evaluation of a surgeoncentered laparoscopic tool design to conventional tools. Proceedings of the Human Factors and Ergonomics Society 49th Annual Meeting, 2005, Orlando, FL, pp. 1705–1709. Trejo, A.E., Done´, K.N., DiMartino, A.A., Oleynikov, D., Hallbeck, M.S., 2006. Articulating vs. conventional laparoscopic grasping tools – surgeons’ opinions. International Journal of Industrial Ergonomics 26, 25–35. Trejo, A.E., Jung, M.-C., Oleynikov, D., Hallbeck, M.S., 2007. Effect of handle design and target location on insertion and aim with a laparoscopic surgical tool. Applied Ergonomics 38 (6), 745–755. van Veelan, M.A., Jakimowicz, J.J., Kazemier, G., 2004. Improved physical ergonomics of laparoscopic surgery. Minimally Invasive Therapy Allied and Technologies 13, 161–166. van Veelan, M.A., Meijer, D.W., 1999. Ergonomics and design of laparoscopic instruments: results of a survey among laparoscopic surgeons. Journal of Laparoendoscopic and Advanced Surgical Techniques, Part A 9 (6), 481–489.


Evaluation of four cursor control devices during a ...

Evaluation of four cursor control devices during a ...

Suggest Documents

Comparison of four different intraosseous access devices during ...

Performance Evaluation of Four On-Site Drug-Testing Devices for ...

Laboratory Evaluation of Four Different Devices for ... - Semantic Scholar

Performance Evaluation of Four On-Site Drug-Testing Devices for ...

Cursor control by Kalman filter with a non-invasive ...

The Neural Control of 2D Cursor Movement - Division of Applied

Evaluation of Traffic Control Devices for Rural High-Speed ...

evaluation of indoor aerosol control devices and their ...

Estimation of Glance from EEG for Cursor Control

A Laboratory-Based Evaluation of Four Rapid

The Vocal Joystick: Evaluation of Voice-based Cursor ... - CiteSeerX

Improving Browsing Performance: A study of four input devices for ...

A prospective study during four consecutive seasons

Nose Tracking Cursor Control for the People with ...

CDM Conference 2012 Mission Palms Hotel - Cursor Control, Inc

A Neuromuscular Interface for Robotic Devices Control

Halo Effects during Internal Control Evaluation: The Influence of ...

Human neural cursor control 1000 days after array implant - CiteSeerX

Computer cursor control by motor cortical signals in humans with ...

Electromyogram-based Cursor Control System for Users with Motor ...

to move the cursor

Information conveyed through brain-control: cursor versus robot ...

An EEG-based BCI System for 2D Cursor Control

Body-machine interface for control of a screen cursor for a child with ...