Proposed Use of a Conversational Agent for Patient Empowerment

Marco Alfano

1,4 a

, John Kellett

, Biagio Lenzitti

and Markus Helfert

Lero, Maynooth University, Maynooth, Co. Kildare, Ireland

Department of Emergency Medicine, Hospital of South West Jutland, Esbjerg, Denmark

Dipartimento di Matematica e Informatica, Università di Palermo, Palermo, Italy

Anghelos Centro Studi sulla Comunicazione, Palermo, Italy

Keywords: Digital Health, Patient Empowerment, Conversational Agent, Tailored Health Communication, Artificial

Intelligence, Big Data.

Abstract: Empowerment is a process through which people acquire the necessary knowledge and self-awareness to

understand their conditions and treatment options, make informed choices and self-manage their health

conditions in daily life, in collaboration with medical professionals. Conversational Agents in healthcare

could play an important role in the process of empowering a person but, so far, they have been seldom been

used for this purpose. This paper presents the basic principles and preliminary implementation of a

conversational health agent for patient empowerment. It dialogues with the user in a “natural” way, collects

health data from heterogeneous sources and provides the user with specific and relevant information. This

allows a person/patient to create his/her own opinion on health matters in the most complete and objective

way, and, therefore, it facilitates the empowerment process.

1 INTRODUCTION

Technological innovations are accelerating disruption

in consumer health and wellbeing but there is still a

disconnect between current healthcare, focused on

disease management, and the needs of empowered

people whose focus is on comprehension and

management of their health (Snowdon, 2020). An

empowered person/patient 1. has the necessary

knowledge and self-awareness to understand his/her

conditions and treatment options, 2. can make

informed choices (i.e. decide), and 3. can self-

manage his/her health conditions in daily life (i.e.

act), in collaboration with medical professionals

(European Health Parliament, 2017; WHO, 2016;

Alfano et al., 2019a; Alfano et al., 2019b). Few

applications exist for person/patient empowerment

and they often work as silos (Snowdon, 2020).

Artificial Intelligence (AI) in healthcare can play

an important role in the process of empowering a

person (Kondyalkis et al., 2013; Iatraki et al. 2018).

AI, however, often empowers machines rather than

https://orcid.org/0000-0001-7200-9547

https://orcid.org/0000-0002-4741-9242

https://orcid.org/0000-0003-2664-7788

https://orcid.org/0000-0001-6546-6408

people (e.g., self-diagnosis apps tend to be substitute

doctors and keep patients as passive recipients,

Davenport and Kalakota, 2019; Jiang et al., 2017; Fast

and Horvitz, 2017). Moreover, AI-driven healthcare

applications are used in separated contexts, use

different data, and work as silos (Herrero et al. 2016).

Finally, the existing virtual assistants are mainly used

for self-diagnosis (decide) and self-

monitoring/management (act). Although they

represent, in principle, the second and third step of the

empowerment process, the first step (understanding)

is unaddressed and, therefore, do not provide genuine

empowerment (Magyar et al., 2019; Herrero et al.,

2016).

How can AI be used to empower people and help

them to better comprehend health information, make

informed decisions and self-manage their health and

wellbeing in collaboration with their healthcare

professionals?

This paper presents the basic principles and

preliminary implementation details of a

conversational health agent for patient empowerment

Alfano, M., Kellett, J., Lenzitti, B. and Helfert, M.

Proposed Use of a Conversational Agent for Patient Empowerment.

DOI: 10.5220/0010414408170824

In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 5: HEALTHINF, pages 817-824

ISBN: 978-989-758-490-9

817

that facilitates the comprehension of health

information so that a person can create his/her own

opinion on health matters in the most complete and

objective way, using the most accurate and

appropriate information available. This is the first of

the three empowerment steps and the precondition for

the other two (i.e., making informed and conscious

health decisions, together with medical professionals,

and actively manage their health and well-being).

The paper is organized as follows. Section 2

illustrates the background and motivation of the

research. Section 3 presents the principles of the

conversational health agent for patient empowerment.

Section 4 presents the details of the initial

implementation of the agent. Sections 5 presents

some conclusions and future work.

2 BACKGROUND AND

MOTIVATION

2.1 Review of the Literature

Conversational agents or chatbots are computer

programs that simulate conversations with users.

These AI systems, also known as relational agents,

inform the user by generating an easily

comprehended dialogue.

We have made a preliminary analysis of the

literature and we found only a limited number of

studies linking person/patient empowerment to

conversational agents: Issom et al. (2020) evaluated

the information provided by a chatbot designed to

help patients with Sickle-Cell Disease to prevent

vaso-occlusive pain. Denecke et al. (2018a; 2018b)

reported an AI-driven self-anamnesis mobile

application in music therapy; a conversational user

interface is chosen to simulate the patient-therapist

conversation but there is no direct empowering action

towards the patient.

We also searched for studies that connect AI or

machine learning with patient empowerment and

conversational agents used in healthcare. Ni et al.

(2020) propose a model that predicts human physical

activity status from sequential lifelogging data

collected from wearable sensors. It can be used as a

decision support tool to provide real-time monitoring,

statistical analysis, and personalized advice. Yadav et

al. (2020) present an AI-driven mobile application

used to predict cases of Anaemia and Thalassemia.

However, this application has been designed for use

by a doctor, nurse, or a health worker, and not by a

patient. The VASelfCare project (Magyar et al 2019)

aimed to develop a conversational agent to facilitate

the self-care of older adults with type 2 diabetes

mellitus (T2D) by improving medication adherence

and lifestyle changes (i.e., physical activity and diet).

The counselling step is tailored through diverse pre-

existing levels of knowledge, by means of a rule-

based decision system. The conversational agent

used reinforcement to learn appropriate behaviour

based on users’ preferences. Herrero et al. (2016)

developed a portable personalised decision support

system to empower individuals on insulin therapy to

self-manage their condition. The blood glucose levels

are collected by the sensors and are managed by a

Case-Based Reasoning (CBR) module to provide

personalised insulin recommendations, while a

second Model-Based Reasoning (MBR) module is

used to maximise users’ safety. You and Gui (2020)

conducted a review and interview study of eleven AI-

enabled chatbot-based symptom checker (CSC) apps.

They found that users perceive the current CSC apps

to lack support for a comprehensive medical history,

flexible symptom input, comprehensible questions,

and diverse diseases and user groups.

Although limited, the review of the current

literature shows:

● Conversational agents are mostly created for a

specific condition.

● Empowerment is almost never addressed directly

and, when it is, only some aspects are considered.

● Comprehension of the health

information/conditions is the least addressed step

of empowerment.

● Little user information is used to provide tailored

content to the user.

● The origin of the provided content is often

unknown and does not use all the information

that is available, for example, on-line (Alfano et

al. 2020b; Alfano et al. 2019c).

● User requirements, such as language level or

information quality (Alfano et al. 2020a; Alfano

et al. 2020b), are not considered.

Therefore, conversational agents currently focus on

specific conditions and mainly deal with self-

diagnosis or self-management of health/conditions.

Almost no agent deals with person empowerment by

considering the user characteristics and requirements,

and then providing him or her with up-to-date and

high-quality customized information.

2.2 Tailored Health Communication

Tailored health communication is the process of

adapting information to the specific characteristics of

an individual (Kamel Ghalibaf et al., 2019). Since it

is more personally relevant, it is more likely to be

read, understood, and acted on (Lustria et al., 2013;

Noar et al., 2011). The process of tailoring health

messages is the same as a tailor uses to make a

Scale-IT-up 2021 - Workshop on Scaling-Up Healthcare with Conversational Agents

818

custom-fit garment according to the customer’s

measurements and their preferred fabric,colour, and

style. Likewise, tailored health communication

considers the user’s needs, interests, and concerns, to

create appropriate “to fit” health information (Bol,

Smit, & Lustria, 2020; Kreuter et al., 2000).

Interpersonal communication has the potential to

be the most customized “tailored” type of

communication, provided all participants understand,

know, and listen to each other. Patients often complain

that discussions with their doctor are unsatisfactory

because they are frequently interrupted and not

listened to (Snyder, 2008; Schouten & Meeuwesen,

2006). Ideally, for any health-related communication

to be empowering it should support an attentive

conversational dialogue with the user to assess his or

her health needs and ensure any information provided

is appropriate and comprehensible (Cheung et al.,

2019); this applies both to traditional inter-personal

and conversational agents.

2.3 Rationale for Tailored Health

Information and Patient

Empowerment

Although provision of tailored health information is

often aimed at a change in behaviour (e.g., smoking

cessation, dietary change, and physical activity),

person empowerment already represents, by itself, a

behavioural change (when it is seen as an outcome

other than a process) because it provides a person

with skills and “power” to make informed decisions,

self-manage health and carry out further lifestyle

changes as needed.

Petty and Cacioppo’s Elaboration Likelihood

Model (Petty and Cacioppo, 1981) provides a

rationale for this approach (Kreuter et al., 1999):

1. by tailoring materials, superfluous

information is eliminated

2. the information that remains is more

personally relevant to the message recipient

3. the message recipient will pay more

attention to information he or she perceives

to be personally relevant

4. information that is attended to is more likely

to have an effect than that which is not

5. when attended to, information that

addresses the unique needs of a person will

be useful in helping him or her decide and

act upon the decision.

In addition, we believe the information provided

to the user must be objective and factual and must not

contain any kind of bias and opinion, unless explicitly

required. In this way, a user will be able to create his

or her own opinion without being influenced (even

indirectly) by biased information.

3 CONVERSATIONAL HEALTH

AGENTS FOR PATIENT

EMPOWERMENT

The requirements of a conversational health agent

that empowers users to understand health

information, make informed decisions, self-manage

their health and wellbeing, and interact better with

healthcare professionals are:

 It dialogues with the user in the most “natural”

way.

 It collects health data from heterogeneous

sources (health information, health status,

empowerment level, needs, etc.) and can

understand, process, and combine them.

 It provides tailored information that is specific

and relevant to patients.

 It provides a person/patient with up-to-date

health information that is easy to understand and

directly actionable.

 The information provided to the user is objective

and factual and does not contain any kind of bias

and opinion, unless explicitly required.

 It explains the principles on how information is

selected in terms that a person/patient can

understand, thus increasing his/her trust and

acceptance.

On the input side, we assume that a user makes a

query (clearly defined or undefined) about complaints

or diseases. The system understands/establishes the

user query and related subqueries in terms of:

 Complaints - re. definition, causes, related

diseases, remedies.

 Diseases - re. presentation, related complaints,

causes, treatment, prognosis, course of disease/

range of severity.

Moreover, the user provides further information

(either directly or indirectly) that is going to be used

as explicit tailoring criteria for the answer:

 Current health status (e.g., symptoms and/or

conditions).

 Background health status (e.g., sex, age, gait,

BMI, comorbid illness).

 Lifestyle information (e.g., sleep, drugs, meal

composition, alcohol consumption, hormonal

cycles).

 Dynamic health indicators (e.g., vital signs

monitoring, physical activity monitoring, stress

level).

 Empowerment level (health literacy,

motivation, gaining control).

Proposed Use of a Conversational Agent for Patient Empowerment

819

 Health and wellbeing needs (e.g., urgent health

improvement, elective quality of life).

Notice that, beside the explicit tailoring criteria, we

also consider some implicit tailoring criteria such as

the language level, to provide the user with

information he/she can easily understand (Alfano et

al., 2020a; Alfano et al., 2020c; Alfano et al., 2018),

and the information quality, in terms of selection of

authoritative sources and factual (unbiased)

information (Alfano et al., 2020b; Alfano et al.,

2019a).

Possible external sources of health information

are:

 Semantic Web (e.g., health-lifesci.schema.org).

 UMLS Metathesaurus.

 Specialized Websites.

 World Wide Web (selected sources).

 Other health information repositories.

We also consider internal information that comes

from the user’s previous data and from data related to

other users (appropriately anonymized).

The system provides a primary output in terms of:

 Tailored health information on complaints

(Definition, Causes, Related diseases,

Remedies).

 Tailored health information on diseases

(Presentation, Treatment, Prognosis, Course of

disease/ Range of severity).

It also provides a secondary output in terms of:

 Suggested available options (based on a

threshold mechanism) on talking to a doctor or

going immediately to a hospital (Wasingya-

Kasereka, 2020).

 Connection with relevant healthcare services

and professionals.

The overall process of the conversational agent is

summarised in Fig. 1 and works as follows:

1. The conversational agent receives the user

input (query and tailoring data) through a

dialogue interface.

2. The conversational agent processes the data,

applies a decision rule algorithm, using the

tailoring data, and retrieves the proper

content from external repositories/databases.

3. The conversational agent

provides the user

with a primary and secondary output.

4. The user provides one of the following

responses to the conversational agent:

o Not OK, I did not get the required

info.

o Partially OK, Explain better.

o OK, Explore further.

o OK, Stop.

For the first three responses, the user can

provide additional information to better

specify his/her further request.

The system will then process the initial and

new data (if present) and the user’s previous

responses (or responses of other users with a

similar profile) and use a reinforcement

learning algorithm to provide the new output.

5. The conversational agent provides the user

with a new primary and secondary output and

goes to step 4.

Figure 1: Summary of the process of the conversational health agent for person/patient empowerment.

Scale-IT-up 2021 - Workshop on Scaling-Up Healthcare with Conversational Agents

820

4 INITIAL IMPLEMENTATION

The implementation of a conversational health agent

for patient empowerment is currently being piloted.

4.1 Complaints and Diseases

There are only a finite number of symptoms and signs

of a disease. For our implementation we have used the

common complaints listed in a classic text

(MacBryde CM, Blacklow RS, 1970):

 Pain

 Gastro-intestinal complaints

 Feverishness

 Cough

 Breathlessness

 Faints or fits

 Insomnia

 Anxiety

 Dizziness

 Palpitations

 Weakness

 Loss of vision

 Impaired hearing

 Bleeding

The number of diseases considered in our first “proof

of concept” iteration has been confined to:

A. the ten conditions most associated with in-

hospital death (Kellett and Deane, 2007):

o Infection

o Heart failure

o Myocardial infarction

o Chronic obstructive lung disease

o Cancer

o Diabetes

o Kidney disease

o Stroke

o Dementia

B. The commonest diagnoses in primary care

(Finley CR et al, 2018):

o Upper respiratory tract infection

o Asthma

o Otitis media

o Tonsillitis

o Urinary tract infection

o Hypertension

o Arthritis

Dyspepsia

o Depression

https://schema.org/

https://health-lifesci.schema.org/

http://webdatacommons.org/

http://commoncrawl.org

o Dermatitis

o Tuberculosis

4.2 Creation of Repository of

Schema.org and Health-lifesci

Structured Data

We have started the creation of a repository of health

information by exploiting the semantic information

available in the World Wide Web and, in particular,

that provided by schema.org

, an initiative funded by

some major Web players, that aims to create,

maintain, and promote schemas for structured data on

the Internet. For the present work, we consider the

health-lifesci extension

that contains 80 types, 162

properties and 125 enumeration values related to the

health/medical field.

We have performed an analysis of the health-

lifesci elements using the data made available by the

Web Data Commons initiative

(Meusel, 2014. The

Web Data Commons contain all Microformat,

Microdata and RDFa (Resource Description

Framework in Attributes) data extracted from the

open repository obtained by the Web Common

Crawl

. The whole dataset contains about 2.5 billion

pages and almost 1 billion pages contain structured

data. The dataset consists of 44 billion RDF n-quads

These are sequences of RDF terms in the form {s, p,

o, u}, where {s, p, o} represents a statement about

semantic data consisting of subject, predicate, object,

and {u} represents the Uniform Resource Identifier

(URI) of the document from which the statement has

been extracted.

From the whole dataset, we have extracted the

subset containing health-lifesci.schema.org elements

and the schema.org elements associated with each

URI that contains health-lifesci.schema.org elements.

Since the queries (direct or indirect) from the

users are about complaints and diseases, we have

initially identified the corresponding health-

lifesci.schema.org types (with the corresponding

explanation). For what concerns the complaints, we

have selected the MedicalSignOrSymptom

type. For

what concerns the diseases, we have selected the

MedicalCondition

type. Moreover, according to the

user query and tailoring data, defined in Section 3, we

have selected the following properties and types

(explanations are taken from health-

lifesci.schema.org):

 name, i.e., the name of the item.

https://www.w3.org/TR/n-quads/

https://schema.org/MedicalSignOrSymptom

https://schema.org/MedicalCondition

Proposed Use of a Conversational Agent for Patient Empowerment

821

 description, i.e., a description of the sign,

symptom, or condition.

 signOrSymptom, i.e., a sign or symptom of the

condition.

 MedicalCause, i.e., the causative agent(s) that

are responsible for the pathophysiologic process

that eventually results in a medical sign,

symptom, or condition.

 possibleTreatment, i.e., a possible treatment.

 drug, i.e., a drug or medicine used in the

treatment.

 expectedPrognosis, i.e., the likely outcome in

either the short term or long term of the medical

condition.

 stage, i.e., the stage of the condition, if

applicable. It is used in the course of the disease.

 epidemiology, i.e., the characteristics of

associated patients, such as age, gender, race, etc.

They are used for tailoring the information.

 riskFactor, i.e., a modifiable or non-modifiable

factor that increases the risk of a patient

contracting this condition, e.g., age, coexisting

condition. It is used for tailoring the information.

 code, i.e., a medical code for the entity, taken

from a controlled vocabulary or ontology such as

ICD-9, DiseasesDB, MeSH, SNOMED-CT, etc.

It is used for connecting the schema.org data to

other vocabularies such as the ones related to the

Unified Medical Language System (UMLS)

The health-lifesci.schema.org types and properties

allow us to provide users with complete information

about both complaints and diseases. In terms of

tailoring, the system is presently limited to the current

and background health status. We are in the process

of adding further tailoring data to our repository and

in particular:

 Lifestyle information

 Dynamic health indicators

 Empowerment level

 Health and wellbeing needs

4.3 User Interfaces

Pilot interfaces have been developed that allow the

user to insert his/her health information in a

conversational way so that the system can create a

profile to tailor the user’s query (Fig. 2 and 3).

Once the user has inserted all the information, the

conversational agent processes the data, applies a

decision rule algorithm, using the tailoring data, and

retrieves the proper content from the schema.org

database (Fig. 4).

https://www.nlm.nih.gov/research/umls/index.html

Figure 2: User interface for collecting current health status

(e.g., complaint).

Figure 3: User interface for collecting background health

status (e.g., gait information).

Figure 4: Preliminary output of the conversational agent for

the sepsis disease.

Given that the present repository, as seen in Section

4.2, does not contain all the data that allow a complete

tailoring of information, the conversational agent

only uses the subset of information that allows such

tailoring.

5 CONCLUSIONS

Our overall objective is that anyone anywhere,

regardless of educational level or health literacy, will

have instant access to health information they

Scale-IT-up 2021 - Workshop on Scaling-Up Healthcare with Conversational Agents

822

understand, which will empower them to decide the

wisest interventions, if any, for their immediate and

future wellbeing. In this paper, we presented the

principles and preliminary implementation of a

conversational agent for patient empowerment that

allows the user to specify his or her requests (either

explicitly or implicitly) in terms of complaints and

diseases and receives tailored health results for his or

her understanding and empowerment.

To our best knowledge, this is the first attempt to

create a conversational agent for patient

empowerment (with a specific focus on

comprehension) for general complaints and diseases.

Moreover, the system is making a novel effort to mix

data that come from different fields and are usually

used separately. We are at the initial stage of the

implementation phase and in the process of

completing the health information repository by using

other sources (such as the UMLS) and adding further

tailoring data. We are also implementing the machine

learning and selection mechanism that will provide

the user with high-quality tailored information in a

language that the user can easily understand. We are

also implementing the quality/accuracy mechanism

for the provided information (Alfano et al. 2020b).

We plan to simulate different user typologies to test

the system in all its different aspects and then run

some tests with real users to evaluate its efficacy and

fine-tune it.

ACKNOWLEDGEMENTS

This work was supported, in part, by Science

Foundation Ireland grant 13/RC/2094 and by the

European Union’s Horizon 2020 research and

innovation programme under the Marie Skłodowska-

Curie grant agreement No 754489.

REFERENCES

Alfano, M., Lenzitti, B., Taibi, D., Helfert, M. (2020a)

Language Complexity in On-line Health Information

Retrieval.. In: Ziefle M., Maciaszek L. (eds) Information

and Communication Technologies for Ageing Well and

e-Health. ICT4AWE 2019. Communications in

Computer and Information Science, vol 1219. Springer,

Cham

Alfano, M., Lenzitti, B., Taibi, D., Helfert, M. (2020b).

Tailored Retrieval of Health Information from the Web

for Facilitating Communication and Empowerment of

Elderly People. In Proceedings of the 6th International

Conference on Information and Communication

Technologies for Ageing Well and e-Health - Volume 1:

ICT4AWE, ISBN 978-989-758-420-6, pages 205-216.

DOI: 10.5220/0009576202050216.

Alfano, M., Lenzitti, B., Lo Bosco, G., Muriana, C., Piazza,

T., Vizzini, G. (2020c). Design, Development and

Validation of a System for Automatic Help to Medical

Text Understanding. International Journal of Medical

Informatics, Elsevier.

DOI:https://doi.org/10.1016/j.ijmedinf.2020.104109

Alfano, M., Lenzitti, B., Taibi, D., Helfert, M.. (2019a).

Provision of tailored health information for patient

empowerment: An initial study. In CompSysTech '19

Proceedings of the 20th International Conference on

Computer Systems and Technologies (pp. 213-220).

Alfano M., Lenzitti B., Taibi D., Helfert M. (2019b) ULearn:

Personalized Medical Learning on the Web for Patient

Empowerment. In: Herzog M., Kubincová Z., Han P.,

Temperini M. (eds) Advances in Web-Based Learning –

ICWL 2019. ICWL 2019. Lecture Notes in Computer

Science, vol 11841. Springer, Cham.

Alfano, M., Lenzitti, B., Taibi, D., Helfert M., (2019c).

Facilitating access to health Web pages with different

language complexity levels. Proc. of the 5th Inter.

Conference on Information and Communication

Technologies for Ageing Well and e-Health (ICT4AWE

2019), 2-4 May 2019, Heraklion-Crete.

Alfano, M., Lenzitti, B., Lo Bosco, G., and Taibi, D., (2018).

Development and Practical Use of a Medical

Vocabulary-Thesaurus-Dictionary for Patient

Empowerment. Proc. of ACM International Conference

on Computer Systems and Technologies

(CompSysTech’18), 13-14 September 2018, Ruse.

Bol, N., Smit, E. S., & Lustria, M. L. A. (2020). Tailored

health communication: Opportunities and challenges in

the digital era. Digital Health, 6, 1–3.

https://doi.org/10.1177/2055207620958913

Cheung, K. L., Durusu, D., Sui, X., & de Vries, H. (2019).

How recommender systems could support and enhance

computer-tailored digital health programs: A scoping

review. Digital Health, 5, 1–19.

https://doi.org/10.1177/2055207618824727

Davenport T, Kalakota R. The potential for artificial

intelligence in healthcare. Future Healthc J. (2019).

6(2):94-98. doi: 10.7861/futurehosp.6-2-94.

Denecke, K., Pöpel, A., Hochreutener, S. L., & May, R.

(2018a). Talking to ana: A mobile self-Anamnesis

application with conversational user interface. ACM

International Conference Proceeding Series, 2018-

April, 85–89.

https://doi.org/10.1145/3194658.3194670

Denecke K, Hochreutener SL, Pöpel A, May R. Self-

Anamnesis with a Conversational User Interface:

Concept and Usability Study. Methods Inf Med.

(2018b).57(5-06):243-252. doi: 10.1055/s-0038-

1675822.

European Health Parliament. (2017). Patient empowerment

and centredness.

Fast, E., & Horvitz, E. (2017). Long-term trends in the

public perception of artificial intelligence. 31st AAAI

Conference on Artificial Intelligence, AAAI 2017,

January 1986, 963–969.

Proposed Use of a Conversational Agent for Patient Empowerment

823

Finley CR, Chan DS, Garrison S, Korownyk C, Kolber MR,

Campbell S, Eurich DT, Lindblad AJ, Vandermeer B,

Allan GM. What are the most common conditions in

primary care? Systematic review. Canadian Family

Physician 2018;64:832-840

Herrero, P., López, B., & Martin, C. (2016). PEPPER:

Patient Empowerment Through Predictive Personalised

Decision Support. Artificial Intelligence for Diabetes.

1st ECAI Workshop on Artificial Intelligence for

Diabetes at the 22nd European Conference on Artificial

Intelligence (ECAI 2016), 8–10.

Iatraki, G., Kondylakis, H., Koumakis, L., Chatzimina, M.,

Kazantzaki, E., Marias, K., & Tsiknakis, M. (2018).

Personal Health Information Recommender:

implementing a tool for the empowerment of cancer

patients. Ecancermedicalscience, 12, 1–11.

https://doi.org/10.3332/ecancer.2018.851.

Issom, D. Z., Rochat, J., Hartvigsen, G., & Lovis, C. (2020).

Preliminary evaluation of a mHealth coaching

conversational artificial intelligence for the self-care

management of people with sickle-cell disease. Studies

in Health Technology and Informatics, 270, 1361–1362.

https://doi.org/10.3233/SHTI200442

Jiang F, Jiang Y, Zhi H, et al. (2017). Artificial intelligence

in healthcare: past, present and future. Stroke and

Vascular Neurology 2017;2: e000101. doi:10.1136/svn-

2017-000101.

Kamel Ghalibaf A, Nazari E, Gholian-Aval M, et al. (2019)

Comprehensive overview of computer- based health

information tailoring: a systematic scoping review. BMJ

Open; 9:e021022. doi:10.1136/ bmjopen-2017-021022

Kellett, J., & Deane, B. (2007). The diagnoses and co-

morbidity encountered in the hospital practice of acute

internal medicine. European Journal of Internal

Medicine, 18(6), 467–473.

https://doi.org/10.1016/j.ejim.2007.02.019

Kondyalkis, H., Koumakis, L., & Tsiknakis, M. (2013).

Smart Recommendation Services in Support of Patient

Empowerment and Personalized Medicine. In

Multimedia Services in Intelligent Environments. Smart

Innovation, Systems and Technologies (Vol. 24).

Springer Cham. https://doi.org/10.1007/978-3-319-

00372-6.

Kreuter, M., Farrell, D., Olevitch, L., & Brennan, L. (2000)

Tailoring health messages: customizing communication

with computer technology. Mahwah New Jersey:

Lawrence Erlbaum Associates. Available at:

https://www.popline.org/node/174671.

Kreuter MW, Farrell D, Olevitch L, Brennan L. Tailoring

Health Messages: Customizing Com- munication Using

Computer Technology. Mahwah, NJ: Lawrence

Erlbaum, 1999.

Lustria, M. L. A., Noar, S. M., Cortese, J., Van Stee, S. K.,

Glueckauf, R. L., & Lee, J. (2013). A meta-analysis of

web-delivered tailored health behavior change

interventions. Journal of Health Communication, 18(9),

1039–1069.

https://doi.org/10.1080/10810730.2013.768727

Magyar, G., Balsa, J., Cláudio, A. P., Carmo, M. B., Neves,

P., Alves, P., Félix, I. B., Pimenta, N., & Guerreiro, M.

P. (2019). Anthropomorphic virtual assistant to support

self-care of type 2 diabetes in older people: A

perspective on the role of artificial intelligence.

VISIGRAPP 2019 - Proceedings of the 14th

International Joint Conference on Computer Vision,

Imaging and Computer Graphics Theory and

Applications, 1(Visigrapp), 323–331.

https://doi.org/10.5220/0007572403230331.

MacBryde CM, Blacklow RS, eds. Signs and Symptoms:

Applied Pathologic Physiology and Clinical

Interpretation. 5th edn. Philadelphia: Lippincott, 1970.

Meusel, R., Petrovski, P., and Bizer, C. 2014. The

WebDataCommons Microdata, RDFa and Microformat

Dataset Series. Proc. of the 13th International Semantic

Web Conference (ISWC14), Springer-Verlag New York,

USA, 277-292.

Ni, J., Chen, B., Allinson, N. M., & Ye, X. (2020). A hybrid

model for predicting human physical activity status from

lifelogging data. European Journal of Operational

Research, 281(3), 532–542.

https://doi.org/10.1016/j.ejor.2019.05.035

Noar, S. M., Grant Harrington, N., Van Stee, S. K., &

Shemanski Aldrich, R. (2011). Tailored Health

Communication to Change Lifestyle Behaviors.

American Journal of Lifestyle Medicine, 5(2), 112–122.

https://doi.org/10.1177/1559827610387255.

Petty RT and Cacioppo JT. (1981). Attitudes and

Persuasion: Classic and Contemporary Approaches.

Dubuque, IA:Wm C. Brown, 1981.

Schouten BC, Meeuwesen L. (2006). Cultural differences in

medical communication: A review of the literature.

Patient Education and Counseling 64, 21–34

Snowdon, A. (2020). Digital health: A Framework for

Healthcare Transformation. In HIMSS.

https://doi.org/10.1038/d41586-019-02869-x.

Snyder U. (2008). The doctor-patient relationship II: not

listening. Medscape J Med; 10: 294.

Yadav S. et al. (2019) Suśruta: Artificial Intelligence and

Bayesian Knowledge Network in Health Care –

Smartphone Apps for Diagnosis and Differentiation of

Anemias with Higher Accuracy at Resource

Constrained Point-of-Care Settings. In: Madria S.,

Fournier-Viger P., Chaudhary S., Reddy P. (eds) Big

Data Analytics. BDA 2019. Lecture Notes in Computer

Science, vol 11932. Springer, Cham.

https://doi.org/10.1007/978-3-030-37188-3_10.

You, Y. & Gui, X. (2020). Self-Diagnosis through AI-

enabled Chatbot-based Symptom Checkers: User

Experiences and Design Considerations.

Wasingya-Kasereka, L., Nakitende, I., Nabiryo, J.,

Namujwiga, T., & Kellett, J. (2020). Presenting

symptoms, diagnoses, and in-hospital mortality in a low

resource hospital environment. QJM: An International

Journal of Medicine, May, 1–7.

https://doi.org/10.1093/qjmed/hcaa169

World Health Organization. (2016). Framework on

integrated, people-centred health services: Report by the

Secretariat. World Health Assembly, (A69/39), 1–12.

Scale-IT-up 2021 - Workshop on Scaling-Up Healthcare with Conversational Agents

824