In this paper we argue that in the multidisciplinary eld related to behavior change support systems using a body-centric approach where participants act rather than discuss and contemplate is a worthwhile technique and should also be investigated more often. Especially now that technology better allows to respond to bodily actions in an appropriate and experiential engaging setting. To this end we will introduce and re ect on two recent case studies we performed: 1) an interactive projection game to trigger better self-management of children with asthma and 2) a Virtual Reality (VR) environment to be integrated in therapy sessions on substance abuse for people with intellectual disabilities (i.e. IQ 50-85 and limited adaptive skills). This resulted in a realistic and controlled environment where individuals with substance use disorder and intellectual disabilities are confronted with substances related to alcohol or cannabis in order to trigger application of strategies for self-control. Both cases included interviews with experts and several user confrontations. These confrontations showed possibilities of including such technologies for persuasive purposes. Furthermore, the responses of the users and experts included where quite positive. This is why we suggest this might be a fruitful direction to look into for other use cases regarding behavior change support systems.
The experiential and body-centric role in learning by playing is a well known
topic of cognitive developmental theories. To paraphrase one of Vygotsky's
examples [
More traditional psychological oriented studies have already shown that
movements can impact attitudes or self-image and related this to a concept of
Embodied Persuasion, see [
In the next sections we will rst introduce the asthma case, the importance
to address asthma, a view on where the opportunities for improvement lay, the
resulting game, and the ndings of the exploratory user evaluation. In the
subsequent section we will introduce the SUD case for which we introduce a controlled
VR environment. In the nal section we will discuss how these cases might be
related to theories, and are clearly related and inspired by the PSD-model [
Asthma is the most common chronic diseases in childhood that frequently starts
at a young age. A recent estimate of the WHO is that worldwide there are 235
million people that su er from asthma 6. Cohort-based estimates of prevalence of
4 The results regarding an induced smiling movement were questioned later on due to
lack of support found in the replication studies https://en.wikipedia.org/wiki/
Facial_feedback_hypothesis, last visited 21-2-2019
5 Note that physically static experienced schema such as (dynamic) visuals,
related to previous acquired embodied movement counterparts (up/down,
avoidance/approach), might also be in uential on their own and are also called Embodied
Persuasion[
The ultimate goal of our envisioned intervention was to improve the physical condition of children with asthma. In this case study we included educational elements by over-simpli ed cause-e ect relations in an entertaining way to improve self-management skills while physically moving, as a step towards this overarching goal.
Previously Klaassen et al. showed a Behavior Change Support System for
children with asthma for which they proposed the combination of monitoring of
physical activity goals with giving rewards in an interactive playground located
in a waiting room of their local hospital [
In this study we again use our interactive playground [
An important conclusion regarding the overall therapy context of the interview
with the Dutch pediatrician specialized in addressing asthma, is that without
7 Although the also reported MAAS cohort from Manchester had a higher prevalence
of 23% this cohort seems to originate from a pre-selected high risk group[
The context of the visits and the waiting room with interactive playground is important to take into account into the design of the game. The average waiting time is about 15 to 30 minutes, there are about 5-10 children simultaneously in the waiting room, and in total about 50 children have an appointment in the morning. When the child is diagnosed they have at least 2 appointments over a period of 6 to 8 weeks. The frequency of visits after this period is very person depended, roughly a child with reasonably controlled asthma would visit the doctor about 3 times a year, whereas a child after exacerbation or even hospitalization would have recurrent appointments until it is under control. 2.3
Education is important in a way that it improves the ability of the patient to
control their asthma. A few of the most important topics of asthma control
following from the semi-structured interview and several scienti c studies are
explained below:
{ Having an active lifestyle, and in particular playing sports regularly.
Unfortunately, physical activity can also be a trigger for an asthma attack,
thus it is not surprising that people with asthma often tend to quit sports
or at least reduce their physical activity drastically due to their asthma. It
could be helpful for the patients to be aware of the importance of physical
activity, despite their fear for a possible attack when combined with poor
asthma control, and especially what can be done to cope with asthma when
playing sports. A broad common consideration is to keep in mind the length
and intensity of the activity and proper medication intake to prevent
severe symptoms of asthma during physical exercise. Furthermore, Thomas
& Bruton [
The designed simple game consists of a central gure, called `Eldub', that needs to be protected from incoming elements, see Figure 1. Eldub shows its physical health both with a health bar and its representation at the centre, ranging from green and happy to red with a few tears. The player can physically walk around in the interactive space, where a circle is projected automatically around the player, see Figure 2. When the circle hits an incoming trigger it is dissolved preventing it from reaching Eldub. The actions are accompanied with simple sounds, such as a cough when a trigger impacts Eldub, spoken instruction texts made high pitched and ltered to give a `alien' impression. The game is intended for a single player but can be played with up to four players playing together.
The most frequent triggers are included, (cf 2.3), with exception of mold, food, and smog. In discussions with the research team, and also following from alternative game concepts, it was decided that appropriate nutrition would be harder to include, and that preventing contact with mold would be more targeted at the parents than children (primarily) playing the game. The games consists of multiple levels, reaching the next level when Eldub has enough health when the timer reaches zero, and each level incoming triggers speeding up throughout the level. Two levels were implemented. The rst level is a spring day environment (150 seconds) which includes various (seasonal) triggers: pollen, dust mites, cigarettes, and animals (dogs & cats). The second level is harder and lasts 200 seconds, it represents a cold environment, the cold itself a ects the health of Eldub negatively constantly. Additional triggers in this environment include smoke (using a camp re icon), strong odors, having a cold, and a temperature drop.
Two special features are included: sports (ball icon) and medicine intake (inhaler icon). The rst feature simulates a real life interaction of physical activity: doing sport is positive for the general health (a green circle around the icon) unless the health is already too weakened due to other triggers (then shown with a red circle). The medication intake is introduced with a small pause in the game with an overlay with written and spoken text that explains that taking the medication can make `Eldub' feel much better even when it does not seem to be really necessary. This object should not be touched by the player's circle, and when it reaches Eldbub it is protected from incoming triggers during 7 seconds which is accompanied with a short happy samba-like music clip.
In this way the game contains information about self-management including
watching out for and learning of certain triggers, the importance of medication,
and bene ts of physical activity. As some children might not be able or willing to
read well the texts are also presented in spoken instructions with a visualization.
We were inspired by the Persuasive System Design framework [
Multiple evaluations were performed of the game as explained above in various stages. The tests were following ethical guidelines of our department, written consent was given by the parents and information was anonymized, the tests were deemed standard non-medical research, and approval from the University's ethics committee was given. Di erent type of users were asked a few questions and were observed playing the game: a primary school teacher, students, e-health professionals, the involved health experts, and nally some of the children from the hospital. Observations included signs regarding the level of understanding (frown/pause/explanations), observable signs of emotions (joy/frustration/indi erence), usability issues, level of activity and interaction, and motivation. Questions di ered on whom was asked but included their opinion of the game, positive/negative feedback, the liking of the overall design, the understanding of the aim of the game and related to this the provided information, and for the health experts the correctness of the represented information.
During a university e-health external matchmaking event with about 60 participants from regional health care organization only 4 of those participants choose to go to the playground (it was in an accompanying room). On the other hand 21 students were interested in the game during two sessions at the university. The students were able to play the game, although this depended heavily on whether they understood enough Dutch to know the (spoken) instructional texts. Although the explanation was a bit hard due to the re ecting light, the used sound lters and sound system at the DesignLab (our university location of the playground). Remarkably, it was quite di cult to understand for the (student) players that the physical exercise feature could both function as a positive and negative trigger. The overall impression was positive, it was even suggested to start a spin-o surrounding the game. Most players found Eldub to be cute, only one student instead saw it as a monster that had to be killed.
The elementary teacher that played the game suggested to include more feedback. The commercial playground expert suggested that the projection itself could be improved and discussed the importance of a good mix between education and fun for such games.
The health experts mentioned that it was important that many of the triggers were included in the game and that the medical content was correct. They imagined it would be suitable for a wide range of age groups. The were concerned whether the children would understand and recognize the triggers and especially the way the physical activity feature worked.
User Test with Children at the Hospital Although the original target group was 7 to 10 years old, for practical reasons in the user test the six participating children ranged from 3 to 9 years old (three were 6 years or older) and their parents often participated as well. Children were asked to play but parents were not discouraged from joining and could tag along. Especially for the younger children this made for a more representative setting of actual use, and would also t well with the importance of including parents in the process towards acceptance of the disease. Sessions were done on two consecutive days in a time span of about 3 hours in total, including the semi-structured interviews each individual session took about 10 to 15 minutes.Small breaks were included directly after the session to write down important observations, feedback, suggestions, problems, and ideas.
The children responded very positively, the most common answer when asked what they liked was: `Everything!'. Two of the six children were unaware of all the triggers and personalized the game by responding that they were allergic to ... as well. Three children stated they loved the physical activity performed during the game. Four of the children played together with their parent, most because they were a bit shy (possibly also due to having an audience), and one child played with his parent because he wanted to win the game and could not manage to do that alone. The number of children playing with their parents was unexpected when designing the game but as explained this might be positive in itself. Many of the children were too shy to truly answer all the indicated questions, the observations provided a better representation. Two of the children did indicate they were introduced to new triggers to consider, some other educational aspects did not come across as planned for other children.
Even the most shy children started to play enthusiastically after about a minute. The di culty seemed to be alright and did result in some children some acceptable level and mix of frustration and fun while playing the game. One of the children managed to survive the rst level on its own. The children really cared about their scores and were still very excited, even if they lost, when they noticed the high number shown on the oor (e.g. 600 points). The younger children were more guided by their parents to play the game together. Some children would rst hold hands. The children above 6 were more excited and less shy when playing the game. One eight-year-old child even nearly begged the parent to continue playing until they won the game, which they unfortunately did not manage to do. The children above 6 also even understood the game without additional explanation, including the distinction between the orange border of triggers, and a green border of the medication intake. However, the game was not clear enough from the beginning, and the physical exercise feature was only interpreted as a positive item throughout the game.
It takes quite some time until the rst triggers enter the screen, especially in the second level, up to 15 seconds, which seemed to be way too long, and this required to mention that the children had to wait for a bit during the user tests. The symbols used for the triggers were not clear for at least two of the children and they indicated they had a problem with this. For instance, one child recognized a sock in both the inhaler and the cigarette butt. As some of the children were fairly small, the tracking system was not yet set to recognize them accurately enough which frustrated the children. Another technical issue to be xed is that the game becomes unresponsive after around 15 minutes of playing. 2.7
To circumvent a bug in which the color of the physical activity feature would not match its e ect (as Eldub's health might have changed), in the most recent implementation the color and its accompanying e ect was decided upon when it spawned.
Several people, including the primary school teacher, noticed that there was a lack of feedback in the sessions, therefore thought bubbles and short spoken texts were added to indicate the current health of Eldub every 50 seconds (e.g. `Wow! You are protecting me very well! ), or when it dropped below a certain threshold for the rst time: `I don't feel so well'. Furthermore, some adjustments were made in the way the triggers were spawned at the start of the second level as this took a bit too long to notice.
Based on feedback of the health experts, a short tutorial of 10 to 15 seconds was included in which all the triggers were explained, which was implemented before the test with the children. In an improved version the triggers should also generalize more, they should not just be about cats and dogs, but about all pets. The game should put more emphasis on the hay-fever season, as depicted in the environment, over a pollen trigger. Furthermore, the confusion and limited recognition of some of the triggers needs additional work, and perhaps going beyond the limited availability of the royalty free on-line resources but simply drawing or buying better depictions. Before the user tests it was suggested to stimulate more to move more, as children could also play with minimum e ort standing near the centre. Instead, the game could also give scores if negative triggers would be dissolved quickly, enticing players to move more. However, the user tests showed that children would run around and none of the children stood still on top of Eldub to catch the triggers more easily.
Fitting the Persuasive Systems Design framework [
Interestingly, there are also two more contextual related suggestions. The rst is to make the playground more recognizable with a white oor and physical sign indicating the playground area. The second is to make the game start automatically including explanation when a player enters after several minutes of inactivity, although sound levels and the repetitive sounds accompanying the explanation might annoy the people around which also need further investigation and work. 3
Individuals with MBID are a risk group for SUD [
Because of these limitations, mainstream treatment protocols need to be
adapted to better suit the needs of individuals with MBID (e.g. personalization
in the form of use of language and content). Over the last few years, several
interventions have been developed targeting individuals with SUD and MBID
[
VR o ers an attractive opportunity to treat SUD in a safe and personalized environment. It can be used to induce craving by introducing the participants to virtual risk situations. Also, participants learn to reduce craving, by applying self-control techniques, and learn which, factors induce and reduce their cravings. 3.1
Qualitative interviews with treatment providers and a review of existing treatment protocols were used to de ne user and functional requirements for the VR environments. Based on these requirements, two prototypes of a virtual bar were designed and developed in an iterative process. Recurring themes within the interventions developed for this target group are self-control techniques. These techniques can be used by the clients when they nd themselves in risky situation with high changes of SUD. The self-control techniques are called the six D's (Deals, Distance, Distraction, Declaring, Di erent thinking di erent acting, Doing great). The self-control techniques are implemented in two di erent VR environments where clients could interact with the environment and apply the learned self-control techniques. The prototypes are developed in Unity 3D and the HTC VIVE is used as VR headset. The rst prototype was a bar environment designed for users with a substance use disorder related to alcohol. Based on the feedback on this environment a second prototype was developed for users with a SUD related to cannabis use. Below we will introduces both environments. The bar environment In the bar environment users are confronted with substances (all related to alcohol) in order to apply the 6Ds, see Figure 3. User can be confronted with situations where someone will o er them a beer or where they have to order a drink at the bar. The 6D's were implemented as follows: { Distance - The user can leave the bar by going out the door and entering an outside area. { Deals - This is a response consequence and is not implemented. { Distraction - The user can play a game of darts in the bar. { Declaring - The user can pick up a mobile phone to make a phone call, see
Figure 3. { Di erent thinking di erent acting - The user can choose other drinks such as co ee and soda instead of alcoholic drinks. { Doing great - Making it a pleasant experience in the virtual environment when the user makes the right choices.
When users enter the virtual environment all elements (other virtual characters, the 6D's and the full interior of the bar) are available. Users have to stand while being in this virtual environment. User can move through the environment by walking around and by using the controllers of the HTC VIVE. The Co eeshop environment In the co eeshop9 environment users are confronted with substances (all related to cannabis) in order to apply the 6Ds. User can be confronted with situations where someone will o er them a joint or where they can order all kinds of cannabis at a counter. The design of the co eeshop 9 A Co eeshop is Dutch for a bar-like environment dedicated to legally tolerated cannabis use. environment also included o ering the user a joint at the table in the co eeshop (exposure), and the 6D's were implemented in a similar fashion as the bar, with some changes: Distraction - The user can play a game of checkers in the coffeeshop. Di erent thinking di erent acting - The user can choose not to order or smoke a joint.
When users enter the virtual environment they enter an empty co eeshop
where only the wall, tables, chairs and the counter are visible. The virtual
environment can be build up by a therapist who has control over all available
elements via the therapist user interface (other virtual characters, the 6D's and
the full interior of the bar), see Figure 3. This control by the therapist provides
an important distinction to make, turning the application into a tool for the
professional, rather than an unrealistic replacement approach, similar to what
Schell addressed for what he considers to be Transformational Games [
The two di erent prototype discussed above are designed and developed with seven patients at a clinic for individuals with SUD and ID. The prototypes are evaluated using the think aloud method, observation, and qualitative interviews. The rst pilot revealed that individuals with SUD and ID were well capable to work with the VR equipment and interact with the virtual environment. However, the target group proved to be easily overwhelmed and distracted and therefore requires a VR product that can gradually increase in complexity. In addition, personalization of the VR environment (i.e. o ering VR stimuli that are highly relevant for the patient, such as using their favorite beer brand), was suggested to be crucial. Therefore, the second prototype (the co eeshop) included the option to start with a simple environment, with the option to increase complexity in both the realism of the virtual environment as well as the interaction with this environment. In this prototype patients could not freely navigate in the virtual space. Therapist had the control over the navigation and the complexity of the environment. The option to personalize stimuli (i.e. choosing between a bar and a co eeshop environment', using the patients favorite substances etc.) was also added in the second prototype. Patients rated these changes positively, and con rmed sta observations that cue exposure within the VR environment triggered craving, and that activities in the VR could be of assistance to reduce these cravings and train self-control techniques. 4
We showed two di erent ways, both regarding strategies and technical implementation, in which an embodied persuasion-while-doing approach was implemented. The last use case was a VR environment in which individuals with SUD and ID could be confronted with substance they crave in a realistic but safe environment. The user confrontations showed that VR o ers opportunities in the treatment of individuals with ID and SUD. It has the potential to become an extension of current treatments as it provides more practical learning opportunities. It also o ers clients to repeat exercises with the push of a button, in their own pace and time, which could improve the treatment satisfaction and adherence of patients. The rst use case resulted in an embodied learning environment that included several elements important for improving self-management of asthma. Children liked to play in it and showed interest in the material to be learned while being entertained and physically moving. The user confrontations of both studies show opportunities to start changing behavior by using interactive embodied technologies, although both were done in a fairly exploratory way and do not provide conclusive evidence.
It is good to notice the success of application of Virtual Reality in the context
of Exposure Therapy (VRET) for both PTSD and various anxiety disorders
[
Unlike these disorders, for that should play a central role, for obvious reasons related to negative e ects on craving. The therapy is focused on training coping mechanisms and therefore care should be taken to nd the right balance. There might also be elements where research ndings can build on shared knowledge generation, for instance considering personalization and integration in a user friendly therapy setting. In the coming years we will continue to work on this interesting topic.
Although we started with some applied knowledge originating from the
theoretical framework of embodied cognition in our mind, we also started from
a technology-driven approach, where we knew the platform before
implementation. As Malinverni, Ackerman, & Pares warned for, we primarily looked at
online (constructing of knowledge while doing) and not yet speci cally to the
o ine transformation of these experiences, nor the integration of the two. We
also anticipated certain pre-conceived and planned outcomes, which shows a
rather reductionist approach, and calls for future research looking into relations
of knowledge construction [
Another important aspect to address is to further accommodate di erent
ways in how participants interact and make sense (cf. [
Furthermore, building on the argumentation of Brin~ol & Petty related to
the Elaboration Likelihood Model, it might be that certain user groups that
are a-priori less motivated or capable to think will experience a bigger e ect
of (a ective) bodily responses [
Beyond the link to related theories our cases also showed the importance of unanticipated outcomes and interactions. We reported on several exploratory evaluations that focused on the interaction, and nding the meaning beyond actions and responses in order to improve the systems. This seems important especially for these speci c user groups. For instance we saw the importance of details (e.g. the brand of the beer o ered) and the subsequent possibilities, as well as the impact of testing in a realistic context (e.g. nding that the younger children actually start of playing with their parent). With these cases we show some of the possibilities of investigating a persuasion-while-doing approach.
Both cases also show a clear relation to implementation of the PSD framework (both include simulation, tunneling, and reminders to target behavior), one might recognize elements of self-monitoring, praise, suggestions, and see opportunities to future extensions of implementations (e.g. additional reward systems in the VR environment and increased tailoring regarding age or the social role of the parent in the Eldub game). Noteworthy is that the system credibility support of the system is not dealt with extensively, especially not in the implementation of the interactive system, but it is more linked to the entire context of use, perhaps while implementing the systems the developers were (unconsciously) expecting these features to be less e ective for these target groups. 5
In this paper we suggested to apply a persuasion-by-doing approach for behavior change support systems. We showed two possible implementations, one focusing on improving self-management of asthma using an interactive playground, and the second focusing on therapy following substance abuse for people with with mild to borderline intellectual disabilities using a virtual reality environment that can be controlled by a therapist. Both exploratory user confrontations showed positive responses that suggest this is a worthwhile direction to look into for the multidisciplinary eld of behavior change support systems.
We like to thank all the sta , children, parents, and other participants for joining our studies, and especially the students Denise van Ingen and Joost van Aggelen who made the implementations in their bachelor and master thesis assignment.