Generation of Multi-Party Dialogues among Embodied

Conversational Agents to Promote Active Living and Healthy Diet for

Subjects Suffering from Type 2 Diabetes

Kuthethur Sneha Jagannath Das

1,2

, Tessa Beinema

, Harm op den Akker

and Hermie Hermens

Biomedical Signals and Systems, University of Twente, 7522NB, Enschede, The Netherlands

Telemedicine Group, Roessingh Research and Development,7522 AH, Enschede, The Netherlands

Keywords: Virtual Coaching, Embodied Conversational Agents, Behaviour Change, Type 2 Diabetes Mellitus.

Abstract: Diabetes Mellitus is a chronic condition that is highly prevalent in the geriatric population. Self-management

plays a key role in the management of this condition. Leading an active lifestyle and having a healthy meal

are cornerstones in managing this condition. Now with the advancement in Information and Communication

Technology (ICT), continuous and proactive care from the affected individual’s side is possible. A change in

the individual’s behaviour would be beneficial in pursuing physical activity and adopting a healthy diet. This

paper focuses on providing a dialogue based virtual health coaching through multiple Embodied

Conversational Agents (ECAs), using strategies such as Motivational Interviewing (MI), Theories of

Behaviour Change and Behaviour Change Techniques (BCTs). The dialogues are constructed to induce a

behaviour change and promote motivation in the affected subjects to work on improving their physical activity

and diet plan.

1 INTRODUCTION

With the advancement in medicine and technology,

the life span of individuals has increased (Maresca et

al., 2018). In many of the European Union countries,

the ageing index has increased significantly. Statistics

reveal that the average life expectancy would increase

from 72 to 81 years (Silva et al., 2018). In such a

situation, there is a need for explicit geriatric

rehabilitation because the longer life span of the

individual needs to be balanced with a good Quality

of Life (QoL) as well (Maresca et al., 2018; Backman

et al., 2018).

One of the new strategies that have been designed

for geriatric care is telemedicine technology (Maresca

et al., 2018). Telemedicine is based on ICT. The

common forms of telemedicine include

telemonitoring and teleconsultation. These forms of

telemedicine have proven to be beneficial for the

geriatric population (Maresca et al.,2018; Gokalp and

Clarke., 2013).

One of the main challenges faced by the elderly

population is the suffering of a chronic condition.

Some of the commonly prevalent chronic conditions

are diabetes, cardiovascular diseases and chronic

obstructive pulmonary disorder (COPD) (Silva et al,

2018). Using mobile health (m-health), for dealing

with chronic conditions have proven to be beneficial

(Silva et al., 2018). m-health has been useful in the

aspects of personalizing the care provided, engaging

the users and enabling them to manage their own

condition. Some key features of m-health are

nutrition management, planning of exercise,

scheduling an appointment and, tracking of

medicines (Silva et al., 2018; Karageorgos et al,

2018).

Diabetes Mellitus is a chronic condition related

to the deficiency of insulin or resistance to insulin and

high blood glucose levels (Inzucchi and Sherwin,

2011). Type 2 Diabetes Mellitus (T2DM), formerly

known as adult-onset diabetes is characterized by the

ineffective use of insulin (Inzucchi and Sherwin,

2011). People suffering from T2DM, are at risk of

developing cardiovascular conditions, stroke,

blindness, gangrene in the foot which could lead to

foot amputation and kidney failure to name a few

(Pardhan et al., 2018; Backholer et al., 2012).

Physical activity (PA) and a healthy diet have proven

to be beneficial in the management of T2DM

(Inzucchi and Sherwin,2011; Di Loreto et al., 2003).

Das, K., Beinema, T., Akker, H. and Hermens, H.

Generation of Multi-Party Dialogues among Embodied Conversational Agents to Promote Active Living and Healthy Diet for Subjects Suffering from Type 2 Diabetes.

DOI: 10.5220/0007750602970304

In Proceedings of the 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health (ICT4AWE 2019), pages 297-304

ISBN: 978-989-758-368-1

297

The treatment of diabetes can be enhanced through

self-management by the affected individual. This can

be achieved through monitoring and education (Sina

et al., 2018).

Health coaching provided to the affected

individuals could aid in the self-management (Gupta

et al., 2018). In this type of coaching, poor health

habits are targeted and worked upon by setting

adaptable and personalized goals, monitoring

progress, providing encouraging feedback and

suggestions for further improvement. Health

coaching aims to bring about a behaviour change in

the individual (Snaith et al., 2018). With the

advancements in the digital healthcare, the health

coaching can be provided virtually through the form

of virtual coaches. These virtual coaches could be the

ECAs. ECAs can be seen as a part of a system that

conveys messages to humans through facial

expressions, hand gestures and voice (Shaked, 2017).

In this paper, the focus lies on developing

dialogues for a panel of ECAs and a human user,

where the human user is suffering from T2DM. The

dialogues aim to provide education and awareness

about T2DM and, motivate the users to adopt PA and

a healthy diet. The dialogues are constructed using the

theories of behaviour change and behaviour change

techniques (BCT).

2 BACKGROUND

In the management of diabetes, Web of Things (WoT)

is being applied. Websites provide information,

problem-solving aids, social support and awareness.

Higher end versions in WoT enable tailored feedback,

self-monitoring of the condition and self-care in

terms of medication, exercise and diet (Al-Taee et al.,

2013). These developments are being made by using

the principle of behaviour change (Al-Taee et al.,

2013). Studies reveal that people suffering from

chronic conditions have to go through behaviour

changes in order to cope with the condition (Aujoulat

et al., 2007). WoT and ICT based systems are being

developed that take into account the feedback

provided to the affected individual. Reminders in the

form of text messages to take medications on time,

electronic diary entries for tracking the diet and

exercise pursued, education and tips to manage the

condition and goals set are sent to the users (Nieto-

Chaupis et al., 2017; Fioravanti et al., 2011).

Digital technology has enabled the creation of

context-aware technology. These technologies

contribute to virtual coaching from the confines of a

home setting (Albaina et al., 2009). Many virtual

coaching systems aid in the pursuit of PA. One such

system is Flowie which has been developed using the

strategies of goal setting, self-monitoring and

intrinsic motivation to make the elderly people walk

more (Albaina et al., 2009).

Virtual coaching using ECAs is being explored

for the purpose of telehealth and behaviour change

(Lisetti et al., 2011). MI is being used to create health

behaviour change interventions. MI is a client-

oriented counselling approach whose main goal is to

resolve discrepancies between the current behaviour

of the patients and their set goals (Lisetti et al., 2011).

Emp-ODVIHC has developed an ECA based

virtual coaching system based on MI. It includes

strategies such as feedback (Amini et al., 2013).

`Most of the times ECAs have a human-like

appearance with the ability to process natural

language and non-verbal communication. ECAs are

being tried to be integrated into a medical platform

(Ravenet et al., 2018). The ECAs being

anthropomorphic in nature has proven to be

beneficial in inducing motivation based on self-

efficacy, interest and attitude (Baylor, 2009).

Adoption of ECAs have been explored in the health

domain for children with autism and for adults with

cognitive impairments (Tanaka et al., 2017; Wargnier

et al., 2015).

Providing telemedicine-based services in terms of

self-monitoring and education has proven to be

beneficial in improving the individual’s diet intake

and PA (van Doorn-van Atten et al., 2018).

3 METHODS

The focus of this paper lies in creating dialogues

among the panel of ECAs and a subject with the

chronic condition of T2DM. The dialogues need to be

able to motivate the subject to adopt healthy

behaviour changes in terms of PA and diet since these

are the cornerstones in the self-management of

T2DM.

Since behaviour change is the key factor here, it

was decided that the dialogues would be written

based on the theories of behaviour change and BCTs.

Hence a thorough literature review as shown in

Figure 1, was implemented using keywords such as

motivation, behaviour change, intrinsic motivation,

goal setting in health care, virtual coaching, self-

management of T2DM, avatar-based coaching, health

coaching and PA monitoring to name a few.

At the end of the literature review, the theories of

behaviour change that could be used in the dialogue

creation and relevant BCTs were chosen. Along with

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

298

this, the types of exercise that could be performed and

the components of a healthy diet that would be

beneficial for people with T2DM were found out.

Based on these points the dialogues were created.

Figure 1: Method adopted in dialogue generation.

4 RESULTS

From the Methodology adopted, there are four main

results obtained and they are discussed in this section.

4.1 Theories of Behaviour Change

At the end of the literature review of psychological

theories related to behaviour change, the five theories

discussed below were chosen to write the dialogues

for the self-management of T2DM.

4.1.1 Goal Setting Theory

This theory focuses on the transition from the set

goals to the actual performance. Two core factors are

involved in this theory: difficulty and specificity of

the set goals. Moderators on the effect of the goals are

self-efficacy, feedback and commitment (Locke and

Latham, 2002). Goals need to be realistically set and

followed up. In T2DM management, we focus on the

pursuit of PA and a healthy diet. Hence, dialogues are

needed that would focus on helping the user to set

achievable goals that would have a positive effect on

their T2DM control.

4.1.2 Social Cognitive Theory

This theory focuses on human behaviour that is

motivated and regulated through self-regulation.

According to the theory, self-regulation occurs

through three sub-processes: self-monitoring, self-

judgement and self-reaction (Bandura, 1991). The

concept of self-efficacy is introduced through this

theory. In keeping T2DM under control, self-

management plays a crucial role. Monitoring the

insulin and glucose levels and keeping track of the

food intake and the levels of PA is needed for T2DM

management. Hence, dialogues that focus on

improving the user’s self-regulation and self-efficacy

are needed.

4.1.3 Protection Motivation Theory

This health psychology theory was developed as a

framework to understand the impact of fear appeals

on an individual. This theory is governed by two

processes: 1. Threat appraisal 2. Coping appraisal.

The cognitive predictors of this framework include

severity, vulnerability, response efficacy which is

based on the recommendations given and self-

efficacy (Rogers, 1983). In the case of T2DM,

through the threat appraisal and coping appraisal, a

self- protection behaviour can be induced in the

subject. For instance, if the user comes to know about,

developing cardiovascular issues due to poor

management of T2DM, out of fear and the need to

protect himself from incurring the issue, he would be

willing to let go of the sedentary lifestyle and quit

eating junk food. The dialogues can be created based

on this theory.

4.1.4 Information-Motivation Behavioural

Skills Model

This model is developed based on three factors. They

are the availability of information and knowledge in

the particular health topic, motivation to change

behaviour and the availability of behavioural skills to

do the specific acts required to tackle the health

condition (Fisher and Fisher, 1992). According to this

model, information and motivation influence each

other. In the case of T2DM, if the subject has the

knowledge that he needs to adopt more PA to manage

the condition, he would be motivated to take the stairs

instead of an elevator. Similarly, if he is motivated to

Generation of Multi-Party Dialogues among Embodied Conversational Agents to Promote Active Living and Healthy Diet for Subjects

Suffering from Type 2 Diabetes

299

manage T2DM, he would take the effort to gather

information about the same. For instance, he would

consult a nutritionist about a healthy diet plan.

4.1.5 Health Belief Model

This model tries to predict and explain when an

individual would adopt actions pertaining to his

health. According to this model, an individual would

be ready to adopt healthy behaviours if he

understands the susceptibility and severity of the

health condition and would adopt the changes when

the cues to action are provided (Janz and Becker,

1984). This model can be incorporated in the

dialogues. For instance, if the subject is made to

understand that eating too many sweets can worsen

T2DM and hence the person is more likely to develop

gangrene in the foot, he would understand that he is

susceptible to it and the severity of the condition as

well. Cues such as reminders by the virtual coaches

that sweets shouldn’t be eaten often can be provided.

4.2 Behaviour Change Techniques

Based on the theories selected in the previous section

and taking into account that the subjects need to adopt

a change in behaviour in terms of an active lifestyle

and diet, from the latest BCT taxonomy, the relevant

BCTs were chosen (Hankonen et al., 2014; Michie et

al., 2013).

The BCTs chosen belong to the cluster of goals

and planning, feedback and monitoring, social

support, shaping knowledge, natural consequences,

repetition and substitution and self-belief (Michie et

al., 2013).

For instance, BCTs belonging to the cluster of

goals and planning can be used in the improvement of

PA. This could be in terms of achieving a required

step count. The feedback and monitoring cluster

would help the individual to follow up his set goals

and his physiological parameters such as the glucose

level. Similarly, through the BCT clusters of shaping

knowledge and natural consequence, awareness

would be created in the subject about T2DM

management and keep the subject informed about the

negative conditions that could occur if T2DM is not

under control. BCTs of repetition and substitution

cluster is useful in the formation of healthy habits

such as cycling to the grocery store instead of using

the car. These kinds of suggestions can be given via

the virtual coaches through the dialogues. By using

the BCTs from self-belief, the subject can be given

the confidence that he could achieve the set goals. For

instance, by reflecting on the subject’s past success of

achieving good PA levels, the subject can be inspired

to achieve higher levels of PA.

4.3 Physical Activity and Diet in T2DM

Literature review on the role of PA and diet in T2DM

management revealed the type of exercises and

components of a healthy diet that could be adopted by

the subject. With regards to the PA, four groups of

exercises were considered to be optimal for T2DM

management. The groups are endurance, passive,

resistance and aerobic types of exercise. These

exercises can alter the body metabolism associated

with T2DM by improving insulin sensitivity, blood

glucose level and glycemic control. This is especially

the case with resistance and aerobic type of exercises.

Hence activities such as walking, running,

swimming, cycling, rowing, skipping and exercises

performed against resistance such as weight lifting

are recommended for subjects with T2DM (Thent et

al., 2013).

With regards to the diet, it was found that the diet

adopted has an explicit effect on obesity and glucose

level regulation. Hence it is highly imperative that a

person with T2DM avoids food with excessive sugar

content, oil and the intake of alcohol as well. Meals

with low-calorie content are recommended for people

with T2DM since it can prevent further progression

of the condition because of insulin sensitivity

improvement and reduction in weight. The ideal meal

should have a balance between the contents of

carbohydrates and high fibre (Lagerros and Rössner,

2013).

4.4 The Dialogues

The results of section 4.1, 4.2 and 4.3 are combined

to generate the dialogues. The below dialogue set is a

result of this combination. Three ECAs have been

chosen for the virtual coaching panel. These coaches

focus on motivating the user to adopt PA and a

healthy diet. A brief persona description of the

coaches and the user is discussed in the following

paragraph (Le Rouge et al, 2013).

Bob: Exercise trainer (Male- 38 years)

Kim: Nutritionist (Female- 30 years)

Emma: Volunteer (Female- 57 years)

Dora: Subject (Female-55 years)

Six months ago, the subject Dora was diagnosed

with T2DM and she is not happy with the way her

condition is progressing. Hence, she has scheduled

her first meeting to meet this virtual team of coaches.

The coaches are Bob whose expertise lies in fitness

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

300

training through exercises, for elderly people since

the last 10 years and Kim who has a Master’s degree

in nutrition sciences and specializes in coming up

with a healthy diet plan for people with T2DM and

cardiovascular conditions. In the panel, Emma a

volunteer who has been suffering from T2DM but is

successfully coping with it for the last 3 years is also

present.

Bob and Kim are amiable kinds of personalities.

Bob says his suggestions right away and is more

authoritative whereas Kim likes to understand her

subject and then give the required suggestions in a

milder manner. Emma is a part of the panel since she

feels that she would be able to relate to the subject’s

worries because she herself has gone through it in the

past and by sharing her experiences she wants to

boost the confidence of the subject. Dora has been

recently diagnosed with T2DM and she is worried

about her condition and wants to take the necessary

steps to control it but she is confused about how to

proceed with it. Hence, she has come to meet the

coaches.

Below is a brief illustration of the dialogues that

could take place between the panel and the user.

Dora: Hello. For the last 6 months, I have been

suffering from Diabetes and with the medicines

prescribed, I tried to take care of the condition. But I

guess it is not working.

Bob: Hello! Don’t worry. A lot of people who have

come to us have been in the same state as you and are

now doing much better. Isn’t it so, Kim?

Kim: Yes. Of course Bob!

Emma: Yes Dora. Even, I faced a similar situation

around 3 years back and with a good exercise and diet

plan, I am able to manage the condition really well.

(BCT: Social Support(unspecified), Information

about Health Consequences)

Dora: Okay. This sounds promising. From where

should I start?

Bob: Dora, to keep Diabetes under check, you need

to follow both a good exercise plan and diet. Only

then would your insulin and glucose levels be under

control. (BCT: Goal Setting(outcome), Goal

Setting (Behaviour), Problem Solving,

Information about Health Consequences)

Emma: Bob is right. Could you tell us a little bit

about your comfort with performing exercises?

Dora: I am not a very athletic person nor do I go to

the gym. Once a while I go for walks but most of the

time, I drive around.

Bob: Hmm. Okay. Maybe you could start with a

45minutes to 1-hour walk every morning. You can

follow this time period for 10 days and then gradually

increase the duration by about half an hour or so.

(BCT- Action planning, Graded Tasks, Habit

Formation, Instruction on how to perform a

Behaviour, Goal Setting (Behaviour))

Emma: You need to focus on your PA levels. If it is

not up to the mark, then you would land up in a

problem. I have been through that phase! (BCT:

Information about Health Consequences,

Information about Antecedents)

Bob: Regarding your diet, the ﬁrst step is to

completely cut down on junk food. This means that

you should cut down on sugary and oily foods. Work

on this for the next three weeks starting from today.

(BCT: Goal setting (Behaviour), Habit Reversal)

Kim: Yes. We can see how this goes.

Dora: okay. I am not very conﬁdent about this but I

will try my best to stick to this plan.

Kim: I think you can do it! Find a friend who would

be your teammate in this for a few days. (BCT: Social

Support (unspecified)

Emma: Yes! I think you will be ﬁne as well. I would

also suggest that you use a diary to write down about

all this on a daily basis. It has helped me to stay

motivated and get things back on track. (BCT-Self-

monitoring (behaviour), Self -monitoring

(outcome), Information about Antecedents)

Dora: Okay. Would you also suggest that I weigh my

weight every week and check for the sugar level?

Kim: Yes. That’s a good idea and it will help in

getting a thorough feedback. (BCT-Bio-feedback,

focus on past success, self-monitoring, feedback

from others, Information about Antecedents)

Dora: Okay. See you!

Kim: Good luck!

Dialogues based on the BCTs such as Goal

Setting(outcome) and Goal Setting (behaviour)

follow from the Goal Setting Theory (Locke and

Latham, 2002). Dialogues that use the BCTs

belonging to the self-monitoring cluster are derived

from the Social Cognitive theory which focuses on

self-efficacy (Bandura, 1991). Overall, the

information provided to the user in the dialogue is

based on the Protection Motivation Theory,

Information- Motivation Behavioural Skills Model

and Health Belief Model. The dialogues provide

information about how the PA should be pursued and

the nature of diet and also focus on what could go

wrong with the subject’s health if the behaviour is not

followed (Rogers,1983; Fisher and Fisher, 1992; Janz

and Becker, 1984). The coaching strategies

incorporated through these dialogues are MI,

Motivational Enhancement Therapy (ME) and

diabetes specific coaching (Lisetti et al., 2011; Miller,

1995; Sherifali, 2017).

Generation of Multi-Party Dialogues among Embodied Conversational Agents to Promote Active Living and Healthy Diet for Subjects

Suffering from Type 2 Diabetes

301

5 DISCUSSION

A dialogue set between a panel of ECAs and a subject

with T2DM has been created for the purpose of

virtual coaching. The dialogues are created using the

psychological theories of behaviour change and

BCTs. Predominantly in the dialogue set, the

coaching strategies of MI and ME along with diabetes

specific management is adopted.

MI and ME are useful in trying to induce an

intrinsic motivation in the subject (Lisetti et al., 2011;

Miller, 1995). Theories and BCTs on which the

dialogues are constructed have been adopted in

coaching systems related to the pursuit of PA and

healthy diet (Hankonen et al., 2014). Goal Setting

Theory, Social Cognitive Theory and Information

Motivation Behavioural Skills model are used

commonly in health psychology and are known for

inducing intrinsic motivation (Van Doorn-van Atten

et al., 2018; Locke and Latham, 2002; Fisher and

Fisher, 1992; Miller, 1995). Goal setting is a very

integral part of health coaching in order to induce a

behaviour change (Snaith et al, 2018). The goals set

need to be understood and agreed upon by the subject

and through the dialogues, the concept of

commitment in pursuing the set goal is being aimed

for (Snaith et al., 2018).

The BCTs chosen focus on inducing an intrinsic

motivation since this type of motivation would help

the subject on a long- term basis (Seifert et al., 2012).

This is the reason why the BCTs are chosen from the

clusters mention in Section 4.2 and BCTs belonging

to the cluster such as rewards have not been included

(Michie et al., 2013). BCTs such as satiation or

exposure is not used in this dialogue set because of

the possibility that it would lead to poor management

of T2DM. For instance, in satiation, the subject could

be asked to eat a lot of chocolates in order to let go of

the habit of eating sugary stuff but it might not go well

with T2DM management (Michie et al., 2013). The

BCTs focus on creating awareness about T2DM and

its self-management. It tries to motivate the subject

by giving her the confidence that she can do it.

In the coaching panel, three ECAs are present who

have expertise in T2DM management. According to

the literature, a positive effect in health coaching is

possible through a multi-virtual coach system

(Kantharaju et al., 2018; op den Akker et al., 2018).

Persuasiveness from the coach’s side also adds more

value to the health coaching. Studies indicate that a

coach with an authoritative personality can be more

successful in inducing a behaviour change (Prochaska

et al., 1994). Hence, Bob’s character is shown to be

authoritative. Cross-gender interaction has also been

recommended in this type of coaching (Kantharaju et

al., 2018).

As a part of the future work, a user evaluation

needs to be conducted to test the competency of the

dialogues delivered by this multi-party ECAs. The

persuasiveness of the dialogues and whether the users

would be intrinsically motivated to adopt an active

living and a healthy diet needs to be evaluated.

Parameters such as how quickly the changes can be

seen, that is whether and when they would move to

the action phase in the five stages of change model

needs to be looked into (Prochaska et al., 1994). For

the virtual coaching to have maximal efficiency, the

ECAs which comprise the heart of the coaching

system need to be fully aware and updated with the

overall health and behaviour of the subject (op den

Akker et al., 2018). For instance, if Dora is attending

the second session with the panel and if there is an

improvement in PA, then the coach needs to be

automatically able to say it during the session. This

could be implemented for instance through a Heart

Rate (HR) sensor worn by Dora, where the HR values

are indicated in prior to the system of coaches. This

way, the system can be tailored and adapted in

specific to each user. Similarly, to improve the

adaptability and smartness of the system, the coaches

need to be able to predict the stage of change of the

user, this could be done through a short questionnaire

that could be filled by the user before the session

starts and based on the questionnaire results, the

ECAs can provide a tailored coaching to the user.

In conclusion, a dialogue set was created between

a set of virtual coaches and a geriatric subject

suffering from T2DM using the theories of behaviour

change, BCTs and coaching strategies such as MI and

ME in order to make the subject adopt better lifestyle

changes in terms of PA and diet.

REFERENCES

Al-Taee, M. A., Sungoor, A. H., Abood, S. N. and Philip,

N. Y., 2013, December. Web-of-Things inspired e-

Health platform for integrated diabetes care

management. In Applied Electrical Engineering and

Computing Technologies (AEECT), 2013 IEEE Jordan

Conference on (pp. 1-6). IEEE.

Albaina, I. M., Visser, T., Van Der Mast, C. A. and

Vastenburg, M. H., 2009, April. Flowie: A persuasive

virtual coach to motivate elderly individuals to walk. In

Pervasive Computing Technologies for Healthcare,

2009. PervasiveHealth 2009. 3rd International

Conference on (pp. 1-7). IEEE.

Amini, R., Lisetti, C., Yasavur, U. and Rishe, N., 2013,

September. On-demand virtual health counselor for

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

302

delivering behavior-change health interventions.

In Healthcare Informatics (ICHI), 2013 IEEE

International Conference on (pp. 46-55). IEEE.

Aujoulat, I., d’Hoore, W. and Deccache, A., 2007. Patient

empowerment in theory and practice: polysemy or

cacophony?. Patient education and counseling, 66(1),

pp.13-20.

Backholer, K., Chen, L. and Shaw, J., 2012. Screening for

diabetes. Pathology, 44(2), pp.110-114.

Backman, C., Harley, A., Peyton, L., Kuziemsky, C.,

Mercer, J., Monahan, M. A., Schmidt, S., Singh, H. and

Gravelle, D., 2018. Development of a Path to Home

Mobile App for the Geriatric Rehabilitation Program at

Bruyère Continuing Care: Protocol for User-Centered

Design and Feasibility Testing Studies. JMIR research

protocols, 7(9).

Bandura, A., 1991. Social cognitive theory of self-regulation.

Organizational behavior and human decision processes,

50(2), pp.248-287.

Baylor, A. L., 2009. Promoting motivation with virtual

agents and avatars: role of visual presence and

appearance. Philosophical Transactions of the Royal

Society B: Biological Sciences, 364(1535), pp.3559-

3565.

Di Loreto, C., Fanelli, C., Lucidi, P., Murdolo, G., De

Cicco, A., Parlanti, N., Santeusanio, F., Brunetti, P. and

De Feo, P., 2003. Validation of a counseling strategy to

promote the adoption and the maintenance of physical

activity by type 2 diabetic subjects. Diabetes care,

26(2), pp.404-408.

Fioravanti, A., Fico, G., Arredondo, M. T. and Leuteritz, J.

P., 2011, August. A mobile feedback system for

integrated E-health platforms to improve self-care and

compliance of diabetes mellitus patients. In 2011

Annual International Conference of the IEEE

Engineering in Medicine and Biology Society (pp.

3550-3553). IEEE.

Fisher, J. D. and Fisher, W. A., 1992. Changing AIDS-risk

behavior. Psychological bulletin, 111(3), p.455.

Gokalp, H. and Clarke, M., 2013. Monitoring activities of

daily living of the elderly and the potential for its use in

telecare and telehealth: a review. TELEMEDICINE and

e-HEALTH, 19(12), pp.910-923.

Gupta, I., Di Eugenio, B., Ziebart, B., Liu, B., Gerber, B.,

Sharp, L., Davis, R. and Baiju, A., 2018, June. Towards

Building a Virtual Assistant Health Coach. In 2018

IEEE International Conference on Healthcare

Informatics (ICHI) (pp. 419-421). IEEE.

Hankonen, N., Sutton, S., Prevost, A. T., Simmons, R. K.,

Griffin, S. J., Kinmonth, A. L. and Hardeman, W.,

2014. Which behavior change techniques are associated

with changes in physical activity, diet and body mass

index in people with recently diagnosed diabetes?.

Annals of Behavioral Medicine, 49(1), pp.7-17.

Inzucchi, S. E. and Sherwin, R. S., 2011. Type 2 diabetes

mellitus. Cecil Medicine. 24th ed. Philadelphia, Pa:

Saunders Elsevier.

Janz, N. K. and Becker, M. H., 1984. The health belief model:

A decade later. Health education quarterly, 11(1), pp.1-

47.

Kantharaju, R. B., De Franco, D., Pease, A. and Pelachaud,

C., 2018, November. Is Two Better than One?: Effects

of Multiple Agents on User Persuasion. In Proceedings

of the 18th International Conference on Intelligent

Virtual Agents (pp. 255-262). ACM.

Karageorgos, G., Andreadis, I., Psychas, K., Mourkousis,

G., Kiourti, A., Lazzi, G. and Nikita, K. S., 2018. The

Promise of Mobile Technologies for the Health Care

System in the Developing World: A Systematic

Review. IEEE reviews in biomedical engineering.

Lagerros, Y. T. and Rössner, S., 2013. Obesity management:

what brings success?. Therapeutic advances in

gastroenterology, 6(1), pp.77-88.

LeRouge, C., Ma, J., Sneha, S. and Tolle, K., 2013. User

profiles and personas in the design and development of

consumer health technologies. International journal of

medical informatics, 82(11), pp.e251-e268.

Lisetti, C.L., Yasavur, U., Visser, U. and Rishe, N., 2011,

May. Toward conducting motivational interviewing

with an on-demand clinician avatar for tailored health

behavior change interventions. In Pervasive Computing

Technologies for Healthcare (PervasiveHealth), 2011

5th International Conference on (pp. 246-249). IEEE.

Locke, E. A. and Latham, G. P., 2002. Building a practically

useful theory of goal setting and task motivation: A 35-

year odyssey. American psychologist, 57(9), p.705.

Maresca, G., De Cola, M. C., Caliri, S., De Luca, R.,

Manuli, A., Scarcella, I., Silvestri, G., Bramanti, P.,

Torrisi, M., Calabrò, R. S. and Bramanti, A., 2018.

Moving towards novel multidisciplinary approaches for

improving elderly quality of life: The emerging role of

telemedicine in Sicily. Journal of telemedicine and

telecare, p.1357633X17753057.

Michie, S., Richardson, M., Johnston, M., Abraham, C.,

Francis, J., Hardeman, W., Eccles, M. P., Cane, J. and

Wood, C. E., 2013. The behavior change technique

taxonomy (v1) of 93 hierarchically clustered

techniques: building an international consensus for the

reporting of behavior change interventions. Annals of

behavioral medicine, 46(1), pp.81-95.

Miller, W. R., 1995. Motivational enhancement therapy

manual: A clinical research guide for therapists

treating individuals with alcohol abuse and

dependence. DIANE Publishing.

Nieto-Chaupis, H., Matta-Sólis, H., Asencios, B. B., Perez-

Siguas, R., Talledo, R. A., Moreno, T. E. and Sifuentes,

V. N., 2017, October. Proposal of a telecare system for

monitoring glucose anomalous behavior in type-2

diabetes patients. In Electrical, Electronics

Engineering, Information and Communication

Technologies (CHILECON), 2017 CHILEAN

Conference on (pp. 1-6). IEEE.

op den Akker, H., op den Akker, R., Beinema, T., Baños,

O., Heylen, D., Bedsted, B., Pease, A., Pelachaud, C.,

Salcedo, V. T., Kyriazakos, S. A. and Hermens, H.,

2018, March. Council of Coaches-A Novel Holistic

Behavior Change Coaching Approach. In ICT4AWE

(pp. 219-226).

Pardhan, S., Nakafero, G., Raman, R. and Sapkota, R.,

2018. Barriers to diabetes awareness and self-help are

Generation of Multi-Party Dialogues among Embodied Conversational Agents to Promote Active Living and Healthy Diet for Subjects

Suffering from Type 2 Diabetes

303

influenced by people's demographics: perspectives of

South Asians with type 2 diabetes. Ethnicity & health,

pp.1-19.

Prochaska, J. O., Velicer, W. F., Rossi, J. S., Goldstein, M.

G., Marcus, B. H., Rakowski, W., Fiore, C., Harlow, L.

L., Redding, C. A., Rosenbloom, D. and Rossi, S. R.,

1994. Stages of change and decisional balance for 12

problem behaviors. Health psychology, 13(1), p.39.

Ravenet, B., Pelachaud, C., Clavel, C. and Marsella, S.,

2018. Automating the production of communicative

gestures in embodied characters. Frontiers in

psychology, 9.

Rogers, R. W., 1983. Cognitive and psychological processes

in fear appeals and attitude change: A revised theory of

protection motivation. Social psychophysiology: A

sourcebook, pp.153-176.

Seifert, C. M., Chapman, L. S., Hart, J. K. and Perez, P.,

2012. Enhancing intrinsic motivation in health

promotion and wellness. American Journal of Health

Promotion, 26(3), pp.1-12.

Shaked, N. A., 2017. Avatars and virtual agents–relationship

interfaces for the elderly. Healthcare technology letters,

4(3), pp.83-87.

Sherifali, D. (2017). Diabetes coaching for individuals with

type 2 diabetes: A state-of-the-science review and

rationale for a coaching model. Journal of Diabetes,

9(6), 547-554.

Sina, M., Graffy, J. and Simmons, D., 2018. Associations

between barriers to self-care and diabetes complications

among patients with type 2 diabetes. Diabetes research

and clinical practice, 141, pp.126-131.

Silva, S., Felgueiras, R. and Oliveira, I. C., 2018. Geriatric

Helper: An mHealth Application to Support

Comprehensive Geriatric Assessment. Sensors (Basel,

Switzerland), 18(4).

Snaith, M., De Franco, D., Beinema, T., Op Den Akker, H.

And Pease, A., 2018. A dialogue game for multi-party

goal-setting in health coaching. In Proceedings of the

Seventh International Conference on Computational

Models of Argument (COMMA 2018), IOS Press,

accepted for publication.

Tanaka, H., Negoro, H., Iwasaka, H. and Nakamura, S.,

2017. Embodied conversational agents for multimodal

automated social skills training in people with autism

spectrum disorders. PloS one, 12(8), p.e0182151.

Thent, Z. C., Das, S. and Henry, L. J., 2013. Role of

exercise in the management of diabetes mellitus: the

global scenario. PloS one, 8(11), p.e80436.

van Doorn-van Atten, M., de Groot, L., de Vries, J. and

Haveman-Nies, A., 2018. Determinants of Behaviour

Change in a Multi-Component Telemonitoring

Intervention for Community-Dwelling Older Adults.

Nutrients, 10(8), p.1062.

Wargnier, P., Malaisé, A., Jacquemot, J., Benveniste, S.,

Jouvelot, P., Pino, M. and Rigaud, A. S., 2015, March.

Towards attention monitoring of older adults with

cognitive impairment during interaction with an

embodied conversational agent. In Virtual and

Augmented Assistive Technology (VAAT), 2015 3rd

IEEE VR International Workshop on (pp. 23-28). IEEE.

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

304