Tag-exercise creator: towards end-user development for ... - CiteSeerX

Tag-Exercise Creator: Towards End-user Development for Tangible Interaction in Rehabilitation Training Ananda Hochstenbach-Waelen, Annick A.A. Timmermans, Henk A.M. Seelen Adelante Centre of Expertise in Rehabilitation and Audiology Zandbergsweg 111, 6432 CC Hoensbroek {A.Hochstenbach,A.Timmermans, H.Seelen}@adelante-zorggroep.nl

Daniel Tetteroo, Panos Markopoulos Department of Industrial Design User Centred Engineering Group Eindhoven University of Technology Den Dolech 2, 5612 AZ Eindhoven {D.Tetteroo,P.Markopoulos}@tue.nl

ABSTRACT

to the market. Contrary to traditional interactive software that relies on rather standardized input/output devices, such standardization has not yet been achieved for tangible interactive systems [11]. This hampers the development of TEIT by exposing application developers to low level hardware and software issues, which are typically the tasks of specialized technology providers. For such companies, typically small and medium sized, application domain specific development can represent an excessive threshold. One approach to overcome this threshold is to enable domain experts to contribute to the development of relevant solutions, acting as end-user developers. This research explores how to enable this venture in the domain of rehabilitation technology, where the need arises for rehabilitation therapists to create training content tailored to different pathologies and even to specific patients. Specifically we examine how end-user development (EUD) can enable therapists to create solutions for rehabilitation training after stroke.

Tangible and embodied interactive technology (TEIT) consists of tightly coupled physical devices and software, which is less the case with mainstream platforms like personal computers, smartphones, etc. Currently TEIT is manufactured by small- and medium-sized niche technology providers for whom application domain specific development can represent an excessive threshold. Enduser development by domain specialists emerges as an avenue to mitigate this issue. This research has set out to enable therapists to create solutions for rehabilitation training, through the development of the Tag-Exercise Creator (TEC). This paper motivates the use of tangible interactive systems for this problem domain, and describes the design, implementation and initial evaluation of TEC. Our study indicates that tools like TEC can enable domain experts to perform EUD tasks and create training content. Improvements and extensions to TEC are under way to enable a field trial of the system where the feasibility of EUD as a professional practice will be evaluated.

Technology For Stroke Rehabilitation

Longitudinal studies of recovery after stroke have shown that less than 50% of the stroke survivors regain a functional arm [1]. For the remaining patients permanent sensory and/or motor disability of the hand constitutes a major problem, since they experience difficulties to use the hand in activities of daily living [10]; this in turn greatly affects their social participation. Stroke incidence is on the rise [15], so a general increase in therapy demand is anticipated. Rehabilitation technology may play an important role to 1) reduce therapist time needed to provide upper limb (skill) training and 2) enable rehabilitation training at home, extending and complementing current training approaches.

Author Keywords

End-User Development, rehabilitation technology, tangible interaction, upper extremity, neurology. ACM Classification Keywords

D.2.6 Programming Environments: Programmer Workbench, D.2.13 Reusable Software: Domain engineering, H.5.2 Information Interfaces and Presentation: User Interfaces INTRODUCTION

Almost two decades since the introduction of graspable/tangible interaction [4], related technologies are gradually making the transition from research laboratories

TEIT may be a valuable aid to support rehabilitation training, especially where this involves manipulation of real life objects, which is key to (re)learn and optimize everyday life actions [13]. Moreover, to exploit the patient’s potential for motor learning, training content has to be varied with a large number of exercises that are tailored to the needs of a particular patient [2].

Permission to make digital or hard copies of all or part of this work for LEAVE BLANKuse THE LAST without 2.5 cmfee (1”) OF THE LEFTare personal or classroom is granted provided that copies THEorFIRST PAGE FOR and THE not made COLUMN or distributed ON for profit commercial advantage that copies bear this notice and theCOPYRIGHT full citation on the first page. To copy otherwise, NOTICE. or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. EICS’12, June 25–26, 2012, Copenhagen, Denmark. Copyright 2012 ACM 978-1-4503-1168-7/12/06...$10.00.

A successful research prototype that supports task-oriented training for neurological patients is the Philips

293

Rehabilitation Exerciser. This system combines the TagTiles board (Serious Toys BV, Den Bosch, NL) with a sensor-enhanced feedback system [12]. A clinical trial of this system showed promising results regarding improvement of arm-hand performance in stroke patients [12]. However, this system also revealed some of the limits of the classical technologist driven approach to innovation in therapy: Participants in the clinical trial were confronted with the lack of an extended library of exercises to continuously challenge them to train and help them keep training beyond the duration of the trial (8 weeks). The need was identified for therapists to add exercises that are customized to the training needs of patients (e.g. supporting different skills and exercise variability), as well as to their proficiency level (e.g. changing the level of difficulty).

In the present paper the design and usability evaluation of the first TEC prototype is described. Therapists can use the TEC for setting up a training programme of arm-hand exercises for stroke patients on the TagTiles board. The complete training system, i.e. the combination of the TagTiles board and the TEC, is called the Tag-Trainer. TAG-TRAINER

The Tag-Trainer was designed to serve three main goals:

End-User Development

Currently, adding and modifying exercises requires software development knowledge, as such exercises need to be programmed into the TagTiles board. Although due effort has been made to facilitate non expert developers to program the TagTiles board through the ESPranto Software Development Kit (SDK) [16], this still presents a prohibitive threshold for therapists, who do not have the skills required to create exercises with such an SDK [7]. While ESPranto facilitates the otherwise daunting task of programming TEIT, till now mostly the task of specialized laboratories, it still requires an investment of effort and a motivation to engage in development that is not consistent with the priorities and mindset of the professional therapist [7]. Therapists are primarily motivated and paid to provide care rather than to create software. Their busy schedules can be prohibitive for extensive software development, and in all cases EUD [9] practices should be contained rather than divert therapists away from their primary activities. In addition, therapists are lacking time to (learn to) program, since time for providing therapy is prioritised and therapy time is limited.



The TEC should facilitate therapists to quickly and easily set up patient-tailored exercises for training armhand skills of stroke patients on the TagTiles board.



Stroke patients should be able to train with the Tag-Trainer themselves.



Therapists should be able to use the TEC as a tool to disseminate and share exercises with other rehabilitation experts in the WikiTherapist online community1.

The Tag-Trainer consists of two main parts: 1.

A simple, yet versatile domain specific tool to create new and adjust existing exercises on the TagTiles board would be helpful for therapists. Therefore a software tool, called the Tag-Exercise Creator (TEC), was developed with which a therapist can both make use of, or edit programmed exercises (already available in a library of exercises), as well as add new ones that fit the patient’s personal needs and proficiency level. Our longer term ambition is to enable therapists to engage in practices related to software creation and sharing, which are well known and practised in the open source software community but are still far removed from the problems and challenges facing health workers nowadays: End-users (e.g. therapists in stroke rehabilitation) should be able to create and share exercises for rehabilitation training. An online community is being developed to enable therapists to use this community as a tool for the dissemination, sharing and co-creation of software for therapy (e.g. the exercises programmed with the TEC).

The TagTiles board (see Figure 1) is a programmable interactive board that consists of a checkerboard area (24 x 24 cm) of 12 by 12 squares. The board can produce sound and each square on the board can provide output through different coloured LED lights in each square. The board is able to detect where and which RFID tags are placed on the board. In this way everyday life objects (to be used for skill training) that are tagged can be recognized by the board and displacements of tagged objects may be tracked. The combination of the board and the tagged everyday life objects is to be used by stroke patients for rehabilitation training of the upper limb. A preproduction version of the board has already been used to create training solutions for children with Cerebral Palsy [3,8] and patients with stroke [12].

Figure 1. TagTiles board by Serious Toys, NL.

1

294

See http://www.wikitherapist.nl

2.

A software tool: the Tag-Exercise Creator (TEC). The TEC has been developed for therapists as a tool to set up patient-tailored exercises for training armhand skills of stroke patients on the TagTiles board.

To keep visual feedback/instructions simple and uniform for stroke patients, blue was chosen as the default colour to represent a location on which tagged objects need to be placed, while green was chosen to confirm that an object is placed correctly. To make feedback to patients possible on the number of repetitions still to be completed, each cycle of step 1 until step 4 is considered as one repetition of a movement within an exercise.

Below we discuss the design process, implementation, and user evaluation of TEC. DESIGN PROCESS

Seven physiotherapists and occupational therapists working in rehabilitation at the Adelante Rehabilitation Centre (Hoensbroek, NL) were interviewed to obtain insight into criteria that technology should meet to be useful and usable in assisting arm-hand skills training in stroke patients [6]. In addition, information was obtained from previous research in the field of technology-assisted arm-hand therapy performed at Adelante [12-14]. The TEC is especially inspired by the iterative design and development of the Philips Rehabilitation Exerciser [12] which showed how stroke rehabilitation could be supported on the TagTiles board. Based on these sources, an initial set of exercises for arm-hand training were created on the board. These exercises were used to demonstrate to therapists the potential of the board and to solicit feedback regarding adaptations and modifications they would like to have or would likely need in order to use it in their practice. It was thought that this would help derive requirements for the TEC tool. Five therapists and two rehabilitation researchers provided input for designing the TEC in repeated sessions over a period of 6 months, gradually extending the capabilities of the system and adjusting its graphical interface. At the end of this process, the TEC tool was evaluated in a summative evaluation process involving final year physiotherapy students, as surrogates to practicing therapists. We discuss the various steps of this process in more detail below.

Therapist Requirements

The initial exercises were demonstrated to two experienced rehabilitation researchers and to five therapists working in stroke rehabilitation. After the demonstrations the therapists and researchers were interviewed on the usefulness of the exercises in training of stroke patients, on the usefulness of the system’s features (e.g. auditory instructions and feedback, light feedback, tagging objects, selection of detection plane size) and on desired adjustable options. Participants indicated that they were content with 

The visual and auditory feedback for individual subtasks and progress feedback for the number of exercises performed.



The possibility to tag and make use of real life objects in the exercises.



The simple and uniform use of colours as visual feedback/instructions.

In addition they stated, among others, that they need to be able to 

Adjust exercises, auditory feedback and instructions to individual patients.



Adjust the number of repetitions, duration (e.g. to determine the interval between picking up and putting down objects) and accuracy of object placement within an exercise.



Create one handed as well as bimanual tasks.



Prevent patients from simply sliding an object when they need to pick it up and displace it.



Make use of more TagTiles boards to enlarge the movement range.

Initial Exercise Design

To familiarize with the nature of the training software that TEC should help create, and to expose therapists to it, we developed and implemented an initial set of exercises on the board. We used the T-TOAT (Technology-supported Task Oriented Arm Training) method [12,14] to define eleven sub-tasks relating to the daily life activities ‘eating with knife and fork’ and ‘drinking from a cup’, e.g., grasping, picking up, reaching out, placing an object on the board. These subtasks were programmed on the board by providing auditory and visual stimuli to the patient. The software was implemented using the ESPranto language [16].

Workflow Concept

1.

A blue plane lights up on the board.

2.

A tagged object is placed on the blue plane.

Therapists will use the TEC to set up an exercise program (of several exercises) tailored to the needs of an individual patient. If an already programmed exercise for one patient is not suitable for another patient (e.g. because it requires a high level of accuracy), the therapist can decide to adjust the already existing exercise to the specific needs of the latter patient. Therapists may also decide to develop completely new exercises, e.g. to match their own therapy approach or the latest developments in their field.

3.

The blue plane turns into a green plane after which the green plane fades away.

The TEC accommodates these three tasks in that it provides three options upon the start of the program:

4.

The object is removed from its position after which the next blue plane appears.

1.

Creating a new exercise

2.

Adjusting an existing exercise, or

The exercises consisted in a basic pattern of four steps:

295

3.

Creating a complete exercise programme for a patient.

two or more objects should be placed at the same time (e.g. picking up and putting down knife and fork). This block gives therapists the opportunity to program bimanual tasks.

The latter option makes it possible to select several exercises from the exercise database and start a training programme for an individual patient.

3) Sequence of blocks – this block provides an abstraction mechanism for repeating a sequence of blocks (i.e. a specific order of appearance of consecutive planes). A number of repetitions can be set for a sequence in order that therapists do not need to repeatedly add identical sequences.

Creating And Modifying Exercises

Figure 2 shows an overview of the TEC interface that allows the creation and modification of exercises. This section discusses in detail the design of this part of the application.

4) Waiting time – this building block can be used to postpone the lighting up of the next plane(s), which is necessary in e.g. strength training (e.g. holding objects in the air for a longer period of time). In addition it often is necessary to add a rest period for the patient to avoid fatigue.

Exercise set-up. As described before, the board can contribute to picking up, displacing, sliding and/or putting down everyday life objects during task training by tracking the tagged objects. Therapists should be able to set the sequence of target locations (as blue planes lighting up on the board), thereby facilitating the displacements. The basic pattern of four steps described above has been abstracted into ready-made building blocks. The following building blocks can be dragged into, and deleted from, the exercise structure to compose an exercise:

Block settings. The building blocks can be used to create new exercises, but also to adjust existing exercises to individual patients’ needs. Each block contains several settings that can be adjusted: the type of object to be placed on the blue plane, if and which auditory instruction should be provided, if feedback should be given or not, and how many times the block should be repeated. In addition, for each block a target plane can be selected on the schematic representation of the TagTiles board by clicking the individual squares. This offers the opportunity to select

1) Sequential block – the added block appears in sequence over time after the previous block(s). 2) Parallel block – two or more parallel planes will light up at the same time, which is necessary for exercises in which

Figure 2. Overview of the TEC application part that allows creation and modification of exercises.

296

bigger or smaller planes to determine the accuracy of object placements, since smaller planes require a higher degree of accuracy from the patient. Exercise settings. The ‘exercise settings’ offer the opportunity to select, among others, the number of repetitions of the complete exercise and whether sliding an object across the board is allowed or not. In addition, an exercise can be tested (virtually, on screen), started (physically, on the board) and saved as general or patienttailored exercise. TEC USABILITY EVALUATION

A small-scale (N=7) user study was performed to assess whether the tool (i.e. the TEC) is usable enough to be deployed in a larger scale field study. All participants had little or no programming experience and had at least a basic level of domain-specific knowledge on rehabilitation therapy for stroke patients.

Figure 3. Dimensions

Method

Questionnaire

scores

for

Cognitive

It is also interesting to note that, although currently the possibilities to re-use parts of exercises are very limited, participants rated the tool to feature a fairly high level of abstraction (Med.=5, IQR=3, see Figure 3). While the concept of ‘blocks’ provides an abstraction layer by integrating several steps of an exercise (lighting-up target plane, waiting for object placement, showing confirmation), the tool contains no abstraction that enables for example group wise operations on blocks themselves.

Participants received a brief explanation about the purpose of the Tag-Trainer and the role of the TEC within the Tag-Trainer. Afterwards, participants were given 5 minutes to familiarize themselves with the tool. As participants could not be expected to map directly general exercises for arm-hand training to tasks that could be done on the board, they were shown a video of a stroke patient executing one out of the initial set of exercises. Finally, the participants were given 3 time-constrained tasks expected to be typical in future use by therapists, i.e.: 1) modification of an existing exercise, 2) creation of a new exercise, and 3) compilation of a training program for an individual patient.

Finally, another interesting result is that participants gave a low rating for the tool’s viscosity dimension, i.e. the effort required to make modifications with changes in one part of the program propagating to the rest (Med.=2, IQR=0.75, see Figure 3). An important goal of the tool is to enable therapists to quickly make changes in existing exercises in order to adapt them to individual patients. The low rating for viscosity is promising in this aspect, because it implies that the tool provides a quick way for tailoring training content.

Participants were asked to report their thoughts and experiences while performing the tasks (i.e. think-aloud protocol was applied). Finally they were presented a questionnaire with 34 statements about using TEC, based on the Cognitive Dimensions (CD) framework [5] and the Technology Acceptance Model (TAM) [17]. Participants were asked to indicate their agreement with each statement on a 7-point Likert-scale.

CONCLUSIONS AND FUTURE WORK

This research illustrates the importance of carefully designed EUD tools, especially for highly-specialized domains such as rehabilitation therapy. The relevance for EUD in this domain is clear: Therapist time needed to provide therapy to patients may be reduced, since therapists only need to set up the patient-tailored training programme after which patients can train by themselves. However, this is only possible if therapists are able to use and modify the technology by themselves in order to create patient-tailored therapy programmes. Previous attempts to provide end-user programming tools for TEITs have been too generic in nature. We developed a domain specific end-user programming tool (TEC) to enable therapists to create and modify exercises for arm-hand rehabilitation on the TagTiles board.

Results

All users were able to complete the tasks given to them, albeit with varying levels of success. In terms of TAMconstructs, users rated TEC overall moderately easy to use (Med.=5, IQR=1.5) for the modification and construction of rehabilitation exercises. Still, a number of usability issues were detected that greatly hindered the performance of some users. Results from the cognitive dimensions questionnaire show that although participants found it easy to modify exercises, TEC did not always prevent users from making errors (Med.=4, IQR=2.75, see Figure 3), mainly because the effects of certain operations are not always directly visible. One frequently observed user action was drawing multiple planes in a single exercise block. Only when a participant wanted to test or execute the exercise, it became clear that this behavior was not supported by the TEC.

The contribution of this paper is to illustrate the nature of the tools that are required to enable domain experts to perform EUD tasks. Furthermore, we show that certain aspects of

297

such a tool, like low viscosity, low error-proneness and high abstraction facilitate therapists in their EUD tasks. Once further improvements have been made on the usability of the tool, it will be deployed for a larger-scale field trial. This trial will focus on the acceptance of a tool like TEC, the therapists’ willingness to perform EUD tasks, and organisational antecedents of acceptance in the therapy domain. In parallel to this trial, a clinical study will be performed to validate the therapeutic value of the proposed system. Finally, we want to evaluate the viability of the used approach (therapists as end-user developers) when applied to other pathologies. In this way technology-assisted therapy could play an important role in decreasing the future therapy demand on therapists ACKNOWLEDGMENTS

2.

Demain, S., Wiles, R., Roberts, L. and McPherson, K. Recovery plateau following stroke: fact or fiction? Disabil Rehabil, 28, 13-14 (2006), 815-821.

3.

Dhillon, B., Goulati, A., Politis, I., Raczewska, A. and Markopoulos, P. A set of customizable games supporting therapy of children with cerebral palsy, in Proceedings of INTERACT 2011 (Lisbon, Portugal, 2011), Springer-Verlag, 360-361.

4.

Fitzmaurice, G.W., Ishii, H. and Buxton, W. Bricks: laying the foundations for graspable user interfaces, in Proceedings of CHI 1995 (Denver, Colorado, USA, 1995), ACM Press, 442-449.

5.

Green, T.R.G. and Blackwell, A.F. Cognitive dimensions of information artifacts: a tutorial, Tutorial presentation at BCS HCI 1998

6.

Hochstenbach-Waelen, A. and Seelen, H.A.M. Embracing change: practical and theoretical considerations for successful implementation of technology assisting upper limb training in stroke. J Neuroeng Rehabil (2012), Accepted for publication.

8.

Li, Y., Fontijn, W. and Markopoulos, P. A tangible tabletop game supporting therapy of children with Cerebral Palsy, in Proceedings of the 2nd International Conference on Fun and Games 2008 (Eindhoven University of Technology, 2008), Springer-Verlag, 182193.

9.

Lieberman, H., Paterno, F., Klann, M. and Wulf, V. End-User Development: An Emerging Paradigm, in End User Development, Springer, Dordrecht, 2006, 1-8.

11. Shaer, O. and Hornecker, E. Tangible User Interfaces: Past, Present and Future Directions. FnT in HCI, 3, Nr 1-2 (2010), 1-138. 12. Timmermans, A.A., Seelen, H.A., Geers, R.P., Saini, P.K., Winter, S., te Vrugt, J. and Kingma, H. Sensorbased arm skill training in chronic stroke patients: results on treatment outcome, patient motivation, and system usability. IEEE Trans Neural Syst Rehabil Eng, 18, 3 (2010), 284-292.

REFERENCES

Broeks, J.G., Lankhorst, G.J., Rumping, K. and Prevo, A.J. The long-term outcome of arm function after stroke: results of a follow-up study. Disabil Rehabil, 21, 8 (1999), 357-364.

Kierkegaard, P. and Markopoulos, P. From Top to Bottom: End User Development, Motivation, Creativity and Organisational Support. End User Development, Lecture Notes in Computer Science, 6654 (2011), 307312.

10. Nowak, D.A. The impact of stroke on the performance of grasping: usefulness of kinetic and kinematic motion analysis. Neurosci Biobehav Rev, 32, 8 (2008), 14391450.

We acknowledge the support of the Innovation-Oriented Research Programme ‘Integral Product Creation and Realization (IOP IPCR)’ of the Netherlands Ministry of Economic Affairs, Agriculture and Innovation. In addition, we thank Mark de Ronde (Hogeschool Zuyd, Heerlen, NL) for his contribution to the development of the TEC, the rehabilitation therapists of Adelante Rehabilitation Centre (Hoensbroek, NL) for sharing their valuable expertise, the physiotherapy students of Fontys (Eindhoven, NL) for participating in the usability evaluation, and Serious Toys BV (Den Bosch, NL) for providing software and advice during the creation of TEC. 1.

7.

13. Timmermans, A.A., Seelen, H.A., Willmann, R.D. and Kingma, H. Technology-assisted training of arm-hand skills in stroke: concepts on reacquisition of motor control and therapist guidelines for rehabilitation technology design. J Neuroeng Rehabil, 6 (2009), 1. 14. Timmermans, A.A.A., Geers, R.P.J., Franck, J.A., Dobbelsteijn, P., Spooren, A.I.F., Kingma, H. and Seelen, H.A.M., T-TOAT: A method of task-oriented arm training for stroke patients suitable for implementation of exercises in rehabilitation technology, in IEEE 11th International Conference on Rehabilitation Robotics, (Kyoto International Conference Center, Japan, 2009), 98-102. 15. Truelsen, T., Piechowski-Jozwiak, B., Bonita, R., Mathers, C., Bogousslavsky, J. and Boysen, G. Stroke incidence and prevalence in Europe: a review of available data. Eur J Neurol, 13, 6 (2006), 581-598. 16. Van Herk, R., Verhaegh, J. and Fontijn, W.F.J. ESPranto SDK: an adaptive programming environment for tangible applications, in Proceedings of CHI 2009 (New York, USA, 2009), ACM, 849-858. 17. Venkatesh, V. and Davis, F.D. A Theoretical Extension of the Technology Acceptance Model: Four Longitudinal Field Studies. Management Science, 46 ,2 (2000), 186-204.

298