USER INVOLVEMENT OF PATIENTS WITH EATING

DISORDER

The Design Process from User needs to Prototype

Marie Sjölinder, Kristina Höök, Jarmo Laaksolahti

SICS, Box 1263, SE-164 29 Kista, Sweden

Liselotte Jansson, Cecilia Bergh, Per Södersten, Modjtaba Zandian

AB Mando, Box 4006, SE-141 04 Huddinge, Sweden

Keywords: Design, User needs, Interactive Treatments.

Abstract: The aim of this work was to investigate different methods for involving patients with eating disorders in the

design process. According to the specific needs of this group of patients four main design goals were set up

for the system. These were: Providing an artificial case manager able of answering questions and of

interacting with the patient; Enhancing motivation to get well; Providing distraction from the disease;

Providing an appealing environment that the patients find interesting enough to use for a longer period of

time. To meet these requirements several methods for involving end-users were used including “Cultural

Probes”, “Repertory Grid Technique” and ”Wizard of Oz”. This paper describes our experiences of using

these methods when developing a system for patients with eating disorders. The methods and the tasks used

in the different studies also turned out to be a distraction away from the disease, which shows the potential

of making the development process a part of the actual usage. The patients’ enthusiasm regarding the tasks

in the studies also showed the importance of using methods suitable with respect to the symptoms of the

disease and tasks that the specific group of patients find interesting and meaningful.

1 INTRODUCTION

There is a great need for treatment of eating

disorders with lasting effects. For a long period of

time the recovery rate among patients has been low.

For example, a patient with anorexia has less than 50

percent chance of recovery within 10 years

(Södersten, Bergh & Zandian, 2006). Eating

disorders are also quite common, The National

Institute of Mental Health (NIMH) estimates that 5

to 10% of the U.S. population has an eating disorder

of one sort or another. In addition to the suffering on

a personal level, the existing treatments are very

time consuming and expensive. These patients suffer

from a disease that has taken over their entire life,

and all their daily activities are centred on eating and

dieting. They are also ambiguous towards getting

cured since they have strong fear of gaining weight.

The fear leads to a constant desire to have certain

questions answered in terms of assurances related to

weight and eating. This need of assurance and

information could be provided for in new

complementary ways, in addition to traditional

treatment. The use of computers and the Internet

along with traditional treatment is a growing

phenomenon that could provide extended help to

people suffering from different psychological

disorders. These electronic environments could also

provide complementary benefits in terms of giving

the patients help and information around the clock.

They could also contribute with pleasant experiences

in terms of providing rich and enjoyable design that

the patients could explore or find distraction within.

Computer supported treatment has previously

been used within, for example, cognitive therapy

(Klingberg et al., 2002; Carlbring, 2004). One

further way to extend this therapy and to provide

treatment can be through the use of rich interactive

environments inspired by applications related to

entertainment, such as computer games. These

environments consist of, for example, 3D-

animations and artificial characters. Artificial

olinder M., H

ok K., Laaksolahti J., Jansson L., Bergh C., S

odersten P. and Zandian M. (2010).

USER INVOLVEMENT OF PATIENTS WITH EATING DISORDER – The Design Process from User needs to Prototype.

In Proceedings of the Multi-Conference on Innovative Developments in ICT, pages 85-92

 SciTePress

characters can also have motivational effects since

they create a sense of confidence (Berry et al., 2004;

Coyle & Matthews, 2004). DeRosis et al. (2005)

have explored the use of artificial characters within

the domain of eating disorders. Their system

consists of an artificial counsellor that through a

dialogue with patients gives advice about healthy

eating habits. However, their efforts have mostly

been geared towards creating an emotional

presentation of information in order to increase the

sense of believability in the artificial character.

The work described in this article has been a co-

operation between a research institute for computer

science (SICS, Swedish Institute of Computer

Science) and a clinic that treats patients suffering

from eating disorders (the Mandometer clinic in

Stockholm). The aim has been to develop an

interactive environment for patients with eating

disorders. The interactive environment and the

functionalities are based on the treatment at AB

Mando. This treatment has its main focus on re-

educating patients how to eat properly and on

motivational talks held on a regular basis with a

personal case manager. The method has been

successful and the probability of recovering at the

clinic is 75% (Bergh et al., 2002).

The purpose of this work has been to develop a

prototype that could be a part of treatment of eating

disorders. The design was based on a 3D-

environment consisting of artificial characters. The

game-like approach was chosen since we wanted to

make the environment appealing to the intended user

group. If a system like this should be useful it is

important that the patients find motivation and joy in

using it and return to it on a regular basis.

From a research perspective the aim with the

system developed was to investigate how interactive

environments like this could be used in the treatment

of eating disorders, and how an interactive

environment and the use of artificial characters

could be used with the aim to answer questions

about a disease, contribute to distraction away from

the disease and to enhance motivation to get well.

Another aim with this work was to explore how

to apply methods for user-centred design in

situations where the user group is suffering from a

disease affecting their participation in the process.

User-centred design is well established, however,

the methods had to be adjusted, and other methods

had to be used as well, when designing for user

groups with disease related behaviours. Certain

demands are placed on the methods used, for

example, it is important to use methods that the

patients find interesting when lack of motivation is a

part of the disease.

2 DESIGN GOALS

Based on the specific needs of patients with eating

disorders four main goals were set up for the system.

During the entire process the design goals were

taken into consideration to ensure that they were met

by the final version of the prototype.

1) Providing an Artificial Case Manager: The

most important aspect of the system was that it

should contain an artificial case manger able to

answer questions in a trustworthy way. Since the

patients have constant worries about their weight

and what they should eat it was also important that

the case manager could disrupt ruminating questions

and thinking.

2) Enhancing Motivation: Motivational aspects

could further be affected by mixed feelings towards

getting well due to the unfounded fear of the patients

that getting well also will mean that they will

become fat. The use of computers could, due to

illness, also be affected by the decreased motivation

to participate in different every-day tasks. The

challenges for a system like this were to both find

ways to motivate the patients to use the system, and

to include aspects and functionalities that within

themselves could provide an overall motivation to

fight the disease, to get well and to look at life in

more positive way.

3) Providing Distraction: The life for this group

of patients has become very narrow due to the

disease and everything has become focused around

eating and dieting. One way of supporting these

patients is to provide them with tasks and things to

do that serves as distraction away from thoughts

related to the disease. The challenge related to this

design goal was to find functionalities to include in

the system that the patients found interesting enough

to actually serve the purpose of distraction.

4) Providing an Appealing Environment: In

order to make a system like this capable of providing

support and actually work as a complement to the

treatment it will, of course, be crucial that the

patients find the system appealing and that they use

it repeatedly over a period of time. To meet this goal

with this group of patients, mainly young girls with

reduced motivation, the environment has to be

entertaining and fun to be in and to use. The

environment should therefore have traits similar to

other interactive environments that these patients

normally use, such as games and on-line

communities.

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

3 DESIGN PROCESS

The problem description that served as a starting

point for this work was given in a broad sense. The

design challenges consisted of narrowing this down

to functionalities and design demands that actually

could be implemented, but at the same time meet all

the design goals mentioned earlier. One aim with

this work was to investigate and study a set of

methods, which situations to apply them in and how;

and also how these methods should be applied with

respect to this particular user group. Based on the

design goals an iterative design process was

conducted, and the entire work was conducted in

close cooperation with the patients that also

provided us with feedback on the methods that we

used. The design methods used in the project

included or was inspired by: Cultural probes (Gaver

et al., 1999), Repertory Grid Technique (Fallman &

Waterworth, 2005) and Wizard of Oz (Dahlbäck,

Jönsson, & Ahrenberg, 1993). Other design methods

were also used within the project; due to limitations

in space only the largest studies are presented here.

3.1 Investigating User needs

by the use of Cultural Probes

The design goals of providing functionalities that

could provide motivation and distraction in an

appealing environment were addressed by a study

using the method “Cultural Probes” (Gaver et al.,

1999). With this method the participants are given

probe packages with items encouraging them to be

creative and playful when solving tasks in a self-

reporting manner. This is an engaging method to

involve users in the design process in areas where it

can be difficult to research by other means. We

chose this method since it would capture issues in

the patients’ lives that are positive and that will

contribute to their recovery. Instead of observing

them in their current atypical situation without social

interactions and suffering from maladaptive

behaviour. By using this method we aimed to get

designing ideas of how to encourage and motivating

the patients to recover.

Seven females patients participated in the study.

The patients were selected on the criterion of getting

a mixed group of patients from the clinic and that

the patients should be in-care patients or frequent

day-care visitors.

In the study the patients got a cardboard box

wrapped as a present. The box contained a disposal

camera, a blank diary, a glue stick, a pair of scissors,

a set of colour pens, pens, pencils, stickers, a

newspaper and a magazine (see figure 1).

Figure 1: The material in the box.

The patients were given one task every day for

four days. During the study, we changed the cameras

and gave the patients previously taken photographs

back printed on paper.

The different tasks were:

 Day 1: “Use the camera to get pictures of, and

the diary to describe things you enjoy or find

distracting.”

 Day 2: “Use the camera to get pictures of, and

the diary to describe things you miss or long

for.”

 Day 3: “Use the camera to get pictures of, and

the diary to describe things making you fight the

disease.”

 Day 4: “Choose some of your previously taken

pictures and paste them in the diary where you

think they best illustrate your previously written

text and possibly add some comments.”

In order to solve the tasks the patients were

encouraged to fill the diary by using the material in

the cardboard box and also to take pictures with the

cameras. They were also encouraged to solve the

tasks in their own personal way. The only obligation

was to paste the note describing the task of the day

as a headline, on a blank sheet in the diary and to

take at least five pictures with the camera to

illustrate the headline.

The patients had devoted large amounts of time

on the tasks and their diaries contained many

creative and colourful entries describing important

issues in their lives. The most frequently mentioned

items in the diaries, regardless of task, were the

family, friends, travelling, reading, sports, pets,

health and music.

Based on the results from the Cultural Probes

study a brainstorming session was held. The aim

with this session was to develop a number of design

suggestions for different functionalities that could

meet the design goals in terms of providing

motivation for recovery and distraction away from

the disease. However, it should be pointed out that

this method aims to capture the participants’

subjective experience of different issues and

USER INVOLVEMENT OF PATIENTS WITH EATING DISORDER - The Design Process from User Needs to Prototype

situations, which in turn provide input to design of

artefacts and interfaces.

Many ideas came up during the brainstorming

and only a few of them could be described or

implemented. One idea that was kept all the way

into the final version of the prototype was the need

for having more than one character. The diversity of

the design goals, in terms of providing answers and

convey trust, enhance motivation to get well and to

provide distraction away from thoughts related to the

disease, made it impossible to incorporate all these

in one character or functionality.

3.2 Appearance of the Characters

and the use of the RGT Method

To be able of providing a trustworthy artificial case

manager, and appealing artificial characters for

motivation and distraction a method inspired by

“Repertory Grid Technique” (Fallman &

Waterworth, 2005) was used. The method is based

on a selective process where different types are

compared and described according to their

similarities and dissimilarities. This method was

chosen since it provided the opportunity for the

patients to describe their subjective experiences

regarding how they perceived the characters.

As a starting point we produced a wide range of

different types of drawn characters to test on the

patients. A professional illustrator produced the

characters as black and white hand-drawn sketches.

Five patients were shown hand-drawn sketches

of 24 characters, divided into four categories, three

categories of human-like characters (women over

the age of 30, women 20-30 years old and men over

the age of 60), and one category of animal-like

characters. Beside asking the patients to describe the

characters, the patients were asked to rate how much

they liked each character (scale 1-9), and to which

extent they would prefer talking to the character

(scale 1-3).

The results regarding the human-like female

characters women over the age of 30 seemed to be

most appealing. However, the characters of this

category were getting the lowest score on “the want

to talk to rating”. Thus, it seemed as even though the

patients fancied the characters visualised as women

over 30 years of age, they preferred to talk to the

younger ones between 20 and 30 years of age.

This results was also supported in the discussion

with the patients, and the conclusion was that a good

artificial case manager should not be too young

(over 30 years of age), which would guarantee

experience and knowledge, but look as up-to-date as

the younger category (women between 20 and 30

years of age), which would guarantee that you would

like to talk to her about anything. The best-liked

women from the case manager categories are

presented in figure 2.

Older than 30 years

20-30 years of age

Figure 2: Best liked case manager characters.

The comments regarding the category of older

men (category 3) were that they seemed friendly.

They were ascribed traits and discussed in terms of

being nice, wise and as interesting storytellers. The

most popular man is presented in figure 3. The

group of elder men seemed about equally popular as

potential talking partners (m=1.8) as the group of

women over 30 years of age.

The animal-like characters (category 4) were the

most popular category in the study (m=7.2). To meet

the design goal of providing distraction, it was

important to create a character that the patients

fancied and enjoyed to interact with, otherwise it

would not provide the intended distraction. The most

popular individual of all categories was the sheep

(see figure 3).

The character for motivation

The pet for distraction

Figure 3: The characters for motivation and distraction.

3.3 Developing the Dialogue

for the Artificial Case Manager

using WoZ

To meet the design goal of providing an artificial

character that could provide trustworthy answers to

the patients questions the Wizard of Oz (WoZ)

method was used. It is a method used to test device

design and techniques and functionality before it is

implemented in a system (e.g. Dahlbäck, Jönsson &

Ahrenberg, 1993). The method is based on using a

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

human to simulate the response of a system. The

"wizard" sits in a back room and simulates the

system’s responses to the user.

Based on the WoZ method we wanted to collect

material to a database in terms of frequently asked

questions as well as examples of how patients would

be expressing themselves when “chatting” to an

artificial case manager. Based on ethical reasons we

informed the patients prior to the WoZ session that

the replies were not automatically delivered by the

computer but written by a real case manager sitting

in another room. However, neither patient nor case

manager were informed about the identity of the

other person.

Sixteen patients were participating, 2 males and

14 females. The patients were told to ask anything

they wanted by writing freely phrased questions in a

“chat window”. They were also asked to imagine

that they were talking to the artificial case manager

seen on the computer screen. Further, they were

informed that the questions would be saved in a log-

file to be used in the design of a dialogue system for

clinical use. No time limits were set for the session.

After the session the patient was interviewed. In

the interviews, the patients expressed their

appreciation of being able to ask questions

anonymously. They felt that this fact made it easier

for them to ask about things they had considered

being embarrassing and private.

The case managers acting like wizards were also

asked about their opinions. All had enjoyed the

method of communicating with the patients, by

“chatting”. Several thought the patients were more

attentive and interested than during traditional face-

to-face sessions. They also thought it was an

excellent opportunity for the patients to get a second

opinion from someone else than their regular caser

manager and thus be convinced of the accuracy in

the replies, e.g. that it is possible for them to eat

without getting overweight.

4 THE FINAL VERSION

OF THE PROTOTYPE

The results from the Repertory Grid study showed

that the patients wanted the artificial case manager

to be a woman older than 30 but with a modern look.

Based on this, the patients were asked to select

between a new selection of sketches of case

managers. The preferred character of this selection

shows in figure 4.

Based on this version, a 3D-model was

developed. However, due to limitations, both of the

Figure 4: The most preferred character.

software and to resources, some simplifications were

made. The 3D-model was also discussed several

times with patients at the clinic and the appearance

was changed according to their suggestions. The

final version of the artificial 3D case manager is

presented in figure 5A. The character with the

function of a case manager was placed in the living

room of an apartment.

To fill the design goal of making an appealing

environment and the patients’ reluctance against

hospital-like environments a 3D-environement

looking like an apartment was developed, figure 5B.

The apartment consisted of two rooms, a hall and a

balcony.

The diversity of the design goals, in terms of

providing answers and convey trust, enhance

motivation to get well and to provide distraction

away from thoughts related to the disease made it, as

mentioned, necessary to include several characters

within the system, each of them with functionalities

meeting some of the design goals, but not all of

them.

Beside the case manager a character for

providing motivation was realised by a character

looking like an old wise man. The man, who was

placed on the balcony of the apartment, delivered

encouraging quotations when clicking on him, figure

5C. The aim was that the character would say

meaningful and encouraging things about life. The

quotations were selected by a group of users that

ranked 40 aphorisms on a scale. The 23 most liked

aphorisms were chosen as material for the character.

The design goal set up to enhance the patients’

distraction also was included in the system. It was a

pet that had been designed based on the users

suggestions, and it had the appearance of a sheep. It

was kept in an enclosure of glass looking like an

aquarium to make the users curious, see figure 5D.

The pet had the function of a small game where the

user could take care of and play with the sheep.

During the work with this part of the system, the

patients were very engaged and came with many

suggestions. Some of these were implemented in a

sub-application.

USER INVOLVEMENT OF PATIENTS WITH EATING DISORDER - The Design Process from User Needs to Prototype

A) The character for

answering questions

B) The environment

C) The old wise man for

motivation

D) The game for distraction

Figure 5: The final version of the prototype.

5 USER FEEDBACK

A user feedback study was conducted at AB Mando

during three weeks of time. During this phase the

patients had free access to computers located at the

clinic. The only criterion for inclusion was that the

patient should visit the clinic frequently (everyday or

at least almost everyday) during the weeks for the

study.

Six patients participated in the study, they were

all young women and their average age was 17

years, Before the study started the patients were

given instructions to use the prototype at least twice

a week during the test phase, and also to write down

comments about the prototype. After the test phase

semi-structured interviews were conducted with the

patients, and the data from the log-files were

analysed.

5.1 Usage and Overall Opinions

The overall attitudes during the interviews were

positive. All patients in the evaluation thought that

the prototype was easy to use and that it worked well

without technical problems.

All patients in the evaluation, except for one,

said that they would like to have access to the

system on a permanent basis. They also wanted to

have access to the system both at the clinic and at

home. However, the purpose with the usage would

be different depending on location. At the clinic the

main purpose would be for distraction, but at home

the system would be used for asking questions to the

artificial case manager.

5.2 The Artificial Case Manager

and the Dialogue System

None of the patients in the evaluation made any

judgements regarding the looks of the artificial case

manager since they had focused on other things in

the system. However, during the interviews, it was

pointed out that the visualisation of the artificial case

manager was important since it would have been

boring to only get answers delivered in a textbox.

With respect to the dialogue system and its

capacity to provide correct answers to the patients’

questions, there was a hit rate of .35. Of the .65

answers that were wrong .19 was explained by the

fact that the question did not exist in the database

and .46 was explained by failure to interpret the

question.

The patients thought the length and content of

the replies of the questions were appropriate and that

the content was highly relevant. However,

sometimes the answers were too medical or too

general. The patients thought that the dialogue

lacked the possibility to talk any deeper of feelings

and emotions, and that the system/dialogue should

provide more comfort when feeling sad.

5.3 Motivation and Distraction

The sheep and the old wise man were included in the

system to provide distraction and motivation. The

patients in the evaluation liked the appearance of

both.

The old wise man delivering motivating

quotations was, for some of the patients, difficult to

find in the environment since he was located at the

balcony. However, the things he delivered were very

much liked by the patients, even though the number

of different quotations was found to be too low. This

was not surprising as the aim of the work was not to

deliver large amounts of content to the system.

When trying to provide distraction it is important

that the content of the application is well suited for

the particular target group. This sub-application and

the appearance of the character were developed in

close cooperation with the patients. As a result of

this, the patients also liked the character and found

the sub-application enjoyable to use. The interaction

with the sheep provided distraction as far as the

limited amount of content could do.

The 3D-environmnet developed for the final

version of the prototype had been redesigned

extensively to convey a feeling of a home

environment. All patients in the evaluation found the

environment attractive, however they thought that it

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT

could be more colourful and that it should have more

soft decorations such as pillows and blankets.

6 DISCUSSION AND GENERAL

CONCLUSIONS

6.1 Achievement of the Design Goals

Providing an Artificial Case Manager: The most

important aspect of the system was that it should

contain an artificial case manger able to answer

questions in a trustworthy way. With respect to this

design goal we reached quite far, even if the systems

dialogue section needs to be adjusted and improved.

The patients trusted the answers even though they

thought they, to some extent, were too medical and

lacked emotions. On the other hand, the medical

approach in the answers serves the purpose of

disrupting ruminating questions and thinking with

respect to worries about weight and what to eat.

Enhancing Motivation: The design goal of

enhancing the patients’ motivation could both be

long-term and short-term. The only aspect that was

possible to investigate, at this stage, was short-term

motivation and whether the character of an old wise

man could make the patients look at life in a little

more positive way for a short while when talking to

him. However, the patients actually liked the things

this character said and they felt encouraged and

strengthened by his words of wisdom.

Providing Distraction The design goal related to

this aspect was to find interesting functionalities to

include in the system. The functionalities had to be

interesting enough to actually serve the purpose of

distraction. For this design goal it was extremely

important to take the profile of the user group into

consideration. The sheep-game and the appearance

of the character were developed in close cooperation

with the patients. As a result of this the patients also

liked the character and found the sub-application

enjoyable to use, even though it was not possible to

provide a large amount of content within the

framework of this project.

Providing an Appealing Environment: Within

this work, much effort had been placed on actually

finding functionalities and traits that are appreciated

and liked by this user group. The work with the

patients also provided much input about further

functionalities to include in the system. The design

goal of providing an appealing environment must be

considered to be met to a quite great extent. The

patients overall positive judgements about the

system and its functionalities and the patients

comments about wanting to have access to a system

like this in their every-day life support this

conclusion.

6.2 The use of Different Methods

The first method we applied was based on the use of

“Cultural Probes” (Gaver et al., 1999). This method

was selected to capture input to the design goals

related to finding functionalities that would enhance

the patients’ motivation to fight the disease and to

find distraction from thoughts related to the disease.

By asking the patients to describe good tings in life

and things they long for, instead of observing their

present life or behaviour, the material from this

study could be used as positive input to the design

process. Within this study the patients found the

study tasks themselves also entertaining and

therefore they also provided distraction away from

the disease. One explanation to why this study

worked so well even though that this group of

patients normally lack the ability to engage in tasks

not related to food intake might be that very clear

instructions were given to the patients. These

instructions made the tasks creative but they also

provided enough information for the patients so that

they could conduct the tasks without being affected

by the disease related difficulties in engaging in new

activities that demands a large amount of initiative

from the individual. One problem with using this

method though, at least in this way, could be that it

only attracts a certain kind of people. The

involvement place high demands on the participants’

engagement, and it might be the case that only

people fond of scrapbooking activities will

participate.

The study that investigated the appearance of the

characters was inspired by the “Repertory Grid

Technique” (Fallman & Waterworth, 2005). This

method was chosen with the aim of letting the

patients describe their own subjective way of

perceiving the appearances and the different traits of

the characters. Since the things to be described were

humans or characters of different kinds a variety of

interpretations could be made, therefore an open

method seemed to be a relevant choice. However, in

our case it also turned out that this complexity

contributed to difficulties in using the method. When

asking people to group something complex as

humans a more detailed instruction about what the

grouping should be based on is needed, otherwise

there is risk that the characters will be grouped or

USER INVOLVEMENT OF PATIENTS WITH EATING DISORDER - The Design Process from User Needs to Prototype

described upon very simple traits related to looks

such as hair colour or hair style.

The WoZ method was appropriate to use for the

purpose of gather dialogue. Besides providing

material for the database, this study also showed the

need among the patients for being able to ask

questions anonymously. Despite the ethical

considerations that made us reveal to the patients

that there was a human on the other side, the study

worked well and the patients forgot that they did not

interact with a system or, at least, did not bother by

this fact. However, when using this method for

developing systems for treatments, it is important to

be aware of that patients can reveal issues that could

be a hazard to the patient’s health if they are kept

anonymous.

6.3 Concluding Remarks

It is important to state that the aim with this work

only was to develop a prototype and to investigate

different methods for user involvement. The

knowledge gathered during this work was intended

to serve as a starting point for the next phase of

developing a system for patients with eating

disorders.

One of the most important results regarding

design of environments like this was the patients’

opinion about the 3D-design/interface. The patients

pointed out the importance of colours and homelike

environments since several of them had spent time at

hospitals against their own will. Further, the studies

showed that it is important that when using an

artificial character to answer questions as a

complement to the treatment for this group of

patients the character has to have an appearance and

an age that conveys trust but also that it is young and

modern enough to be someone that the patients feel

that they would like to talk to. Finally, the patients’

enthusiasm for the Cultural Probes study supports

the idea of using the design process as a direct

means to an end. In this case, using the design

process as a part of the treatment, an interesting

aspect well worth investigating further.

ACKNOWLEDGEMENTS

The authors would like to thank AB Mando and their

staff. The work is based on their method for treating

eating diseases, and all the work has been conducted

at the clinic in cooperation with their staff. We will

also thank all the patients at the clinic that have

participated in the studies. The funding for this

project was given by VINNOVA (Swedish

Governmental Agency for Innovation Systems),

along with AB Mando and SICS.

REFERENCES

Bergh, C., Brodin, U., Lindberg, G., & Södersten, P.

(2002). Randomised controlled trial of a treatment for

anorexia and bulimia nervosa. Proc Natl Acad Sci USA

2002; 99: 9486.

Berry, D., Butler, L., de Rosis, F., Laaksolahti, J.,

Pelachaud, C., & Steedman, M. (2004). Final

evaluation report, the Magister project. January 2004.

Carlbring P. (2004). Panic! Its Prevalence, Diagnosis and

Treatment via the Internet. Doctoral thesis,

Department of Psychology, Uppsala University, May

2004.

Coyle, D., & Matthews, M. (2004). Personal Investigator:

a Therapeutic 3D Game for Teenagers, in workshop

proceedings 'Social Learning through Gaming, CHI

conference Wienna, April 2004.

Dahlbäck, Jönsson & Ahrenberg, 1993). Wizard of Oz

Studies -- Why and How. Knowledge-Based Systems,

Vol. 6, No. 4, pp. 258-266, December, 1993. Previous

version published in Proceedings from the 1993

International Workshop on Intelligent User Interfaces,

Orlando, Florida, pp. 193-200. 1993.

de Rosis, F., Cavalluzi, A., Mazotta, I. & Novielli, N.

(2005). Can embodied conversational agents induce

empathy in users? AISB'05 "Virtual Characters

Symposium".

Fallman, D., & Waterworth, J. (2005). Dealing with User

Experience and Affective Evaluation in HCI Design: a

Repertory Grid Approach. Presented at the Workshop

Innovative Approaches to Evaluating Affective

Interfaces, CHI April 2-7 2005, Portland, Oregon,

USA.

Gaver, B., Dunne, T., & Pacenti, E. (1999). Cultural

Probes. Interactions jan-feb, 21-29.

Klingberg T., Forssberg H., & Westerberg H. (2002).

Training of working memory in children with ADHD.

In J Clin Exp Neuropsychol. 2002 Sep; 24(6):781-91.

Södersten, P., Bergh, C., & Zandian, Z. (2006).

Understanding eating disorders. Horm Behav 2006;

50:572.

INNOV 2010 - International Multi-Conference on Innovative Developments in ICT