Intelligent Provision of Tailored, Easily Understood, and Trusted

Health Information for Patient Empowerment

Marco Alfano

1,4 a

, John Kellett

, Biagio Lenzitti

and Markus Helfert

1,4 d

Innovation Value Institute, 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

Lero, Science Foundation Ireland Research Centre for Software, Ireland

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

Intelligence, Responsible AI, Big Data, Machine Learning.

Abstract: Although digital transformation in healthcare is accelerating, there is still a disconnect between current

healthcare, focused on disease management, and a more holistic approach that looks at the health and

wellbeing of the whole person. The latter approach aims at empowering patients and other health information

seekers by improving their comprehension of their health so that they can manage it better. Currently, few

stand-alone applications for patient empowerment exist and they seldom help users to understand health

information. Thus, health information seekers often interact with the Web through generic search engines,

which often produce results that are overwhelming, too generic, and of poor quality. This paper shows how

the use of Artificial Intelligence (AI) in a responsible way may provide patients and others with health higher

quality information that empowers them to improve their health and wellbeing. It presents an AI engine that

extracts health content from the Web and provides the user with health information that is relevant,

trustworthy, and easy to understand. The AI engine has been used to create an Intelligent Empowering Agent

(IEA) that dialogues with users in simple language to provide customised information on symptoms and

diseases, which helps them form their own evidence-based opinion on health matters that concern them.

1 INTRODUCTION

Although digital transformation in healthcare is

accelerating, “the existing care pathways and care

models rely on prescriptive approaches of health

professionals who assess and direct care and

treatment for patients, rather than creating care

models designed and informed by patients to achieve

personal goals and health outcomes” (Snowdon,

2020; Snowdon et al., 2014). The latter approach aims

at empowering patients by improving their

comprehension of their health so that they can

manage it better (Alfano et al., 2019; European

Health Parliament, 2017; WHO, 2016). Patient's

health literacy, information-seeking behaviour, sense

of meaning, shared decision-making, and self-

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

care/self-management are among the most important

elements that characterise person/patient

empowerment (Bodolica and Spraggon, 2019; Cerezo

et al., 2016; Fumagalli et al., 2015). To be empowered

a person/patient must:

1. have the necessary knowledge and self-

awareness to understand health conditions and

treatment options;

2. be able to make informed and conscious health

choices (i.e., decide);

3. actively manage, with or without advice from

medical professionals, their health and well-

being (i.e., act).

Currently, few stand-alone applications for patient

empowerment exist (Snowdon, 2020; Bodolica and

384

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

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment.

DOI: 10.5220/0011694300003414

In Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 5: HEALTHINF, pages 384-391

ISBN: 978-989-758-631-6; ISSN: 2184-4305

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

Spraggon, 2019) and patients and other users often go

on the Web to search for health information (Alfano

et al., 2019; Finney Rutten et al., 2019). However,

Web information is mainly obtained through generic

search engines and is often overwhelming, too

generic, outdated and of poor quality (Finney Rutten

et al., 2020; Alfano et al., 2019). Although Artificial

Intelligence (AI) could play an important role in

health empowerment (Iatraki et al., 2018; Kondyalkis

et al., 2013), it often empowers machines rather than

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

doctors and keep patients as passive recipients;

Davenport and Kalakota, 2019; Fast and Horvitz,

2017; Jiang et al., 2017).

This paper presents the principles of the

responsible use of AI for person/patient

empowerment and describes an AI engine that can be

used to extract and process health content from the

Web. It provides the user with health information

tailored to their specific needs, which is simple to

understand and trustworthy. This AI engine powers

an Intelligent Empowering Agent (IEA) that

dialogues with users in simple language and provides

customized information on symptoms and diseases,

which helps them form their own evidence-based

opinions on whatever health matters concern them.

2 RESPONSIBLE AI FOR

PATIENT EMPOWERMENT

Responsible Artificial Intelligence is the “practice of

designing, building, deploying, operationalizing and

monitoring AI systems in a manner that empowers

people and businesses, and fairly impacts customers

and society – allowing companies to engender trust

and scale AI with confidence” (Responsible Artificial

Intelligence Institute, 2021). Although Responsible

AI, Ethical AI, and Trustworthy AI are often used

interchangeably (Responsible Artificial Intelligence

Institute, 2021), in the context of our research to

enhance patient empowerment, we prefer the term

Responsible AI.

The basic principles of Responsible AI in health

are analysed in the WHO guidance on “Ethics &

Governance of Artificial Intelligence for Health”

(2021). It shows that new technologies that use AI

hold great promise to improve diagnosis, treatment,

health research and drug development. AI

technologies can also support government run public

health functions, such as disease surveillance and

outbreak control, provided ethics and human rights

are at the heart of their design, deployment, and use.

The “Ethics guidelines for trustworthy AI” by the EU

High-Level Expert Group on Artificial Intelligence

(HLEG, 2019), sets out a list of seven key

requirements on the AI systems:

• Human agency and oversight;

• Technical robustness and safety;

• Privacy and data governance;

• Transparency;

• Diversity, non-discrimination, and fairness;

• Societal and environmental wellbeing;

• Accountability Trustworthy.

Responsible AI requires a holistic and systemic

approach, encompassing the trustworthiness of all

actors and processes that are part of the technical and

socio-technical context. The most important actors

are human beings, and they must be able to interact

with AI systems at any stage. In particular, the first

requirement, “Human agency and oversight”, states

that AI systems should empower human beings by

allowing them to make informed decisions and

safeguarding their fundamental rights. Users should

be given the knowledge and tools to comprehend and

interact with AI systems to a satisfactory degree and,

where possible, be enabled to reasonably self-assess

or challenge them. AI systems should support

individuals in making better, more informed choices,

in accordance with their goals (HLEG, 2019).

There cannot be truly Responsible AI in health

without the direct intervention of “empowered”

human beings (health professionals and patients) who

must be able to understand both the AI-system

process and its outcome, and directly participate in

the decisions and actions that originate. For

person/patient empowerment, the first step is

understanding health information. A previously

published research and a literature review (Alfano et

al., 2019, 2021) identified the following metrics of

health-information comprehension for empowerment

purposes:

• language complexity;

• information quality

• information classification/customization

(tailoring).

Thus, person empowerment becomes both a

requisite and an outcome of Responsible AI

(Fig. 1) and to have Responsible AI in health:

• AI systems must behave responsibly by

applying ethical principles such as the ones

discussed above (technical requirements);

• AI systems must help people to become

empowered by providing health information

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment

385

that presents characteristics that facilitate its

comprehension (socio-technical requirements).

The application of these principles leads to the

“virtuous circle” shown in Fig. 1. A Responsible AI

system facilitates person empowerment and an

empowered person exerts “human agency and

oversight” on the AI system.

Figure 1: “Virtuous” interaction between Responsible AI

and an empowered person.

3 USING AI FOR EXTRACTING

AND CLASSIFYING WEB

HEALTH INFORMATION

As shown in Alfano et al. (2020) and in Fox et al.

(2013), the most searched type of health information

on the Web is about medical problems and diseases.

Therefore, we decided to provide the user with web

information on symptoms and diseases organised in

sections. This is similar to what many other health

web sites do, but in a more structured,

comprehensive, and interactive way. Initially, we

made a visual analysis of fifty health web pages of

symptoms and diseases to understand what kind of

information is provided and how it could be best

divided into sections. The headings of each web page,

their semantic equivalents and themes were analysed

to determine those which were most frequently

provided; these were then grouped together into the

following initial classification:

• Overview (summary, what is it, what are the

types, definition, description, consideration,

types, types of, deeper types of, basics, basic

information).

• Alternative name (synonyms, substitutes).

• Symptoms (what are the symptoms, clinical

presentation, what you feel, related complaints,

what you cannot do anymore).

• Causes (what is main cause, it is hereditary,

causes, possible causes, most common causes,

most serious causes, what causes, other causes,

culprit, causes of, health conditions).

• Diagnosis and test (how is it evaluated and

diagnosed, what test will be done to diagnose

it, diagnosing, what to expect at a medical visit,

location of your pain, type and intensity,

history of your pain, other medical history,

other types of diagnosis, how is it diagnosed).

• Risk factor (who gets it, what are the risk

factors, who’s at risk, who does it affect).

• When to see a doctor (when to contact

professionals, when to contact a medical

professional, when should I call my healthcare

provider, when to see a doctor, what symptoms

require medical care, when to worry, symptoms

that require urgent care, when should I see my

healthcare provider).

• Management and treatment (how is it

treated, medication for it, treating the

underlying condition, how can I get rid of it,

home care, care and treatment, medication,

medications, how to ease it yourself, treatment

for, medical treatment, home treatment,

surgery).

• Prevention (how can I prevent it, beware of,

can it be prevented).

• Outlook and prognosis (can it be cured, what

happen after I start treatment, what should I

expect in the long term, complications, what is

the outlook if you have it, what’s the outlook,

living with).

Next, we created a list of 474 symptoms derived

from a classic textbook on signs and symptoms

(MacBryde, R.S. Blacklow, 1970), review of the

literature, and expert opinion, and a list of 801

diseases derived by the commonest diagnoses

encountered in primary care (Finley et al., 2018) and

the diseases most often associated with in-hospital

death (Kellett and Deane, 2007).

Machine learning was then applied to

automatically classify the headings of web pages

related to the symptoms and diseases to the classes

seen above. This was done by using the following

functions:

• Url_Generator (k, n) uses a Google

API to

retrieve n URLs of keyword k (e.g., a

symptom). For each symptom and disease, 40

URLs were retrieved.

HEALTHINF 2023 - 16th International Conference on Health Informatics

386

• BaseScraper (URL) takes each URL and

extracts all <h> headings.

• MachineLearner uses a TensorFlow

algorithm to create an AI model that allocates

headings to classes. TensorFlow is an open

source library for numerical computation and

large-scale machine learning. It allows to create

dataflow graphs—structures that describe how

data moves through a graph, or a series of

processing nodes. Each node in the graph

represents a mathematical operation, and each

connection or edge between nodes is a

multidimensional data array, or tensor (Abadi

et al., 2016). The AI model has been trained

with 1500 headings manually allocated to

different classes, reaching a ~ 85% precision.

• Sorter uses the AI model to classify headings

by indicating the probability a heading belongs

to a class. It uses the TextVectorization()

function, provided by the Keras framework

to transform text into a vector by also removing

capital letters and punctuation to avoid similar

words counted as different vectors.

As a last step, for each symptom or disease, the 40

web pages indicated by Url_Generator were

downloaded. The HTML contents, related to the

various headings, were extracted, using the DOM

the HTML page, and stored in a DB. The most fitting

web information, for each symptom/disease section,

is chosen based on:

• Information Relevance

The proximity of the symptom/disease with

the title of the page, S

, is computed as the

maximum of the cosine of similarity between the

symptom/disease and its synonyms (A) and the

title of the page (T), as follows:

𝑆



(

𝐴,𝑇

)

=𝑚𝑎𝑥

∑





∙









∑













∑











(1)

where A

and T

are the occurrences of the

same word present in the two texts and n is the

number of words present in A and T.

The symptom/disease occurrence index,

occ

, is computed as the maximum of the ratio

between the average occurrences of the

symptom/disease and its synonyms (A) and the

number of occurrences of the most frequent word

https://www.tensorflow.org/

https://keras.io/

in the page (P), as follows:

𝐼



(

𝐴,𝑃

)

=𝑚𝑎𝑥

(

∑











)

(



)

(2)

For improved results, we have applied a

stemming process to all words and we have

evaluated the similarities between words by

using the normalized Levenshtein distance

(Yujian and Bo, 2007).

• Information Quality

The quality of a web page P is computed by

counting the meta data of the page that can be

associated to the information quality (Alfano et

al., 2021; WHO, 2003; Eysenbach, 2002) and

grouping them in four categories:

o Temporal indicators T(P) (Last-Modified,

revised, PublishDate, ...);

o Spatial indicators S(P) (og:email,

og:phone_number, og:street-address, ...);

o Descriptive indicators D(P) (abstract,

summary, description, og:description, …);

o Author indicators A(P) (author,

citation_author, …).

o A fifth indicator O(P) refers to the most

popular health web sites.

The Information Quality, Iq, is then computed as

follows:

𝐼



(

𝑃

)

()∙()∙()∙()∙()



(3)

• Language Complexity

The Language Complexity, L

of a web page P, is

computed with the following formula:

𝐿



(

𝑃

)

∑

(



)∙







∑









(4)

where W

are the words of a text P, WF(W

) is

the Word Familiarity, i.e., the number of Google

results of W

(Alfano et al., 2021; Kloehn et al.,

2018, Leroy et al., 2016), and A

are the occurrences

of W

The most fitting information is computed by using

the following weighted formula:

𝐹𝑖𝑡𝐼𝑛𝑓𝑜 = 𝛼 ∙ 𝑆



+ 𝛽∙𝐼



+𝛾∙𝐼



+ 𝛿∙𝐿



(5)

Where α + β + γ + δ = 1 and allow to differently

weigh the information relevance, information quality

https://en.wikipedia.org/wiki/Document_Object_Model

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment

387

and language complexity. A Support Vector Machine

(SVM, Cortes and Vapnik, 1995) has been trained

with human classified web pages to provide the best

combination of the weights and then the most

appropriate information for each section.

4 INTELLIGENT EMPOWERING

AGENT

The AI engine, presented in the previous section, has

been implemented as part of an Intelligent

Empowering Agent (IEA) that provides health

information tailored to the users’ needs, which is

intelligible, current, accurate, trustworthy, valuable,

and usable. The IEA model is shown in Fig. 2. Its

components have been described in detail (Alfano et

al., 2022) and can be summarized as follows:

a. User Query. The user selects a complaint from a

list or directly enters it as free text.

b. User Profile. The user provides some

information about him/her (currently age range

and sex).

c. Search Engine. The search engine retrieves the

top 40 Google results for each symptom/disease.

d. AI Algorithm. The AI algorithm uses the AI

model shown above to select the health

information according to following criteria:

o Custom Information, to provide users with

tailored content for the symptom/disease of

interest, organized in sections.

o Information Quality, to provide users with

current, accurate, trustworthy, and

unambiguous information.

o Language Complexity, to provide users

with information that they can easily

understand.

e. Output Presentation. Tailored relevant health

information is provided on complaints, diseases,

medical tests, when to see a doctor, treatment,

prevention, and scientific articles according to

the section headings presented above. The

scientific-articles section uses a PubMed

TM4

API

to retrieve the most relevant articles on the topic

of interest.

A “traffic-light” colouring coding (i.e., red,

amber, or green), that implies the need for an

urgent consultation with a healthcare

professional, is also provided.

In an IEA prototype, the Conversational Health

Agent for Person Empowerment (CHAPE -

http://cohealth.ivi.ie/chape/), users input their age and

sex; CHAPE then provides a list of possible

complaints that can be easily understood, such as pain

or discomfort, breathlessness, and weakness or

fatigue. Depending on the complaint selected and the

user’s profile characteristics, a further sub-list of

possibly related complaints is presented, to help

define the primary complaint more precisely (Fig. 3).

Alternatively, the user can directly type in any

complaint in a free text area.

Figure 2: IEA model.

https://pubmed.ncbi.nlm.nih.gov/

HEALTHINF 2023 - 16th International Conference on Health Informatics

388

Figure 3: CHAPE interface allows users to specify their

complaints in an easy and natural way.

An information window is then displayed (Fig. 4)

and it contains:

• Complaint name with a background colour

(red, amber, or green), which indicates the

health risk.

• Complaint description.

• Alternate names of the complaint.

• Related complaints.

• Disease(s) associated with the complaint.

• Tests commonly used to further define the

complaint.

• When to see a doctor

• Treatment.

• Prevention.

• Scientific articles.

Figure 4: Output window containing information about the

searched element.

When a related complaint, disease, or test is clicked

on, a new information window for that element is

opened. The list of related complaints, diseases and

tests of the new window is ordered so that the

elements that are in common with the previous

searches are shown first (Fig. 5). This further

customizes the provided information by following the

user search path.

Figure 5: Related complaints and diseases are ordered by

their correlation with previous searched elements.

Some subjective tests have been carried out to

assess the effectiveness of the IEA in terms of

usability, user experience and perceived value

(Alfano et al., 2022). In terms of usability, most

respondents found CHAPE interface clear and helpful

in identifying health information about complaints,

diseases, and tests. In terms of user experience, most

respondents could better understand their complaints

and related diseases and tests. In terms of perceived

value, the majority of respondents found that CHAPE

would improve communication with their doctors.

Expanding the user profile and improving the

system’s interaction interface were the main

recommendations.

5 CONCLUSIONS

This paper shows how the use of AI in a responsible

way may help empower a person/patient better

manage their health and wellbeing. It presents an AI

engine that extracts health content from the Web to

provide the user with health information that is

relevant, trustworthy, and easy to understand. The AI

engine powers a prototype Intelligent Empowering

Agent (IEA) that dialogues with user in simple

language and provides tailored, trustworthy

information on symptoms and diseases, which help

users to form evidence-based opinions on health

matters that concern them. To our best knowledge,

this is the first attempt to create an intelligent

empowering agent that exploits the potential of AI

and the vast amount of health information available

on the Web to facilitate comprehension and action on

general complaints/diseases.

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment

389

6 FUTURE

RESEARCH/DIRECTIONS

Although third party subjective assessments are

encouraging, the user profile and the system

interaction need to be improved. To this end, the user

interface is being provided with a graphical

representation of the body, for complaint

identification and location, and more user profile

information, such as gait, body type, nutritional

status, comorbidities, are being added. Complaints

and diseases are being associated with Concept

Unique Identifiers (CUI) of the Unified Medical

Language System (UMLS)

to map them to standard

terms taken from medical-term classifications such as

ICD-9

, ICD-10

, or SNOMED

. How information

is gathered and filtered out will be improved and

explicitly explained to improve trustworthiness.

Although user input is anonymous, users will be

provided with an option to grant or withdraw

informed consent to use their data. Finally, the agent

is going to be tested on a wider demographic.

ACKNOWLEDGEMENTS

This work was supported, in part, by Science

Foundation Ireland grant 13/RC/2094_P2 and co-

funded under the European Regional Development

Fund through the Southern & Eastern Regional

Operational Programme to Lero - the Science

Foundation Ireland Research Centre for Software

(www.lero.ie). Thanks to Yvan Pannefieu, from

ESEO Grand École d'Ingénieurs, for his contribution

on the machine learning algorithm for the extraction

and classification of HTML headings.

REFERENCES

Abadi, M. et al. (2016). TensorFlow: A System for Large-

Scale Machine Learning. Proc. of 2th USENIX

Symposium on Operating Systems Design and

Implementation (OSDI ’16). https://doi.org/

10.5555/3026877.3026899.

Alfano, M., Kellett, J., Lenzitti, B., Helfert, M. (2022).

Characteristics and Subjective Evaluation of an

Intelligent Empowering Agent for Health Person

Empowerment. Studies in health technology and

informatics, 299, 177–182. https://doi.org/10.3233/

SHTI220979.

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

https://www.cdc.gov/nchs/icd/icd9cm.htm

Alfano, M., Lenzitti, B., Taibi, D., Helfert, M. (2021). On-

Line Retrieval of Health Information Based on

Language Complexity, Information Customization and

Information Quality. In Ziefle M., Guldemond N.,

Maciaszek L.A. (eds) Information and Communication

Technologies for Ageing Well and e-Health.

ICT4AWE 2020. Communications in Computer and

Information Science, vol 1387. Springer, Cham, 2021.

https://doi.org/10.1007/978-3-030-70807-8_1.

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

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, 2020,

Springer, Cham. https://doi.org/10.1007/978-3-030-

52677-1_5

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

Provision of Tailored Health Information for Patient

Empowerment: An Initial Study. In Proceedings of the

20th International Conference on Computer Systems

and Technologies (CompSysTech ’19). Association for

Computing Machinery, New York, NY, USA, 213–

220. https://doi.org/10.1145/3345252.3345301.

Bodolica, V., Spraggon, M., (2019). Toward patient-

centered care and inclusive health-care governance: a

review of patient empowerment in the UAE. Public

Health, 169(971), 114–124. https://doi.org/10.1016/

j.puhe.2019.01.017

Cerezo, P. G., Juvé-Udina, M. E., Delgado-Hito, P. (2016).

Concepts and measures of patient empowerment: A

comprehensive review. Revista Da Escola de

Enfermagem. https://doi.org/10.1590/S0080-6234201

60000500018

Cortes, C. and Vapnik, V. (1995). Support-Vector

Networks. Machine Learning, 20, 273-297.

http://dx.doi.org/10.1007/BF00994018

Davenport T, Kalakota R. (2019). The potential for

artificial intelligence in healthcare. Future Healthc J.

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

European Health Parliament. (2017). Patient empowerment

and centredness. https://www.healthparliament.eu/

patient-empowerment-centredness/#:~:text=There%

20is%20a%20widespread%20consensus,expanding%2

0burden%20of%20chronic%20diseases.

Eysenbach G., Powell, J., Kuss, O., Sa, E.R. (2002).

Empirical studies assessing the quality of health

information for consumers on the world wide web: a

systematic review. JAMA J. Am. Med. Assoc,.

287(20), pp. 2691–2700, 2002. https://doi.org/

10.1001/jama.287.20.2691.

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

perception of artificial intelligence. In Proceedings of

31st AAAI Conference on Artificial Intelligence, AAAI

2017, pp. 963–969, January 2017. https://doi.org/

10.48550/arXiv.1609.04904

https://www.cdc.gov/nchs/icd/icd-10-cm.htm

https://www.snomed.org/

HEALTHINF 2023 - 16th International Conference on Health Informatics

390

Finley C.R., Chan D.S., Garrison S., Korownyk C., Kolber

M.R., Campbell S., Eurich D.T., Lindblad A.J.,

Vandermeer B., Allan G.M. (2018). What are the most

common conditions in primary care? Systematic

review. Canadian Family Physician 2018;64:832-840.

Finney Rutten L.J., Blake, K.D., Greenberg-Worisek, A.J.,

Allen, S.V., Moser, R.P., Hesse, B.W. (2019). Online

Health Information Seeking Among US Adults:

Measuring Progress Toward a Healthy People 2020

Objective. Public Health Reports. 2019;134(6):617-

625. https://doi.org/10.1177/0033354919874074.

Fox, S., Duggan, M.: Health Online 2013, Pew, Research

Center’s Internet & American Life

Project (2013). http://www.pewinternet.org/2013/01/15/

health-online-2013/

Fumagalli, L. P., Radaelli, G., Lettieri, E., Bertele’, P.,

Masella, C. (2015). Patient Empowerment and its

neighbours: Clarifying the boundaries and their mutual

relationships. Health Policy. 119(3), 384–394.

https://doi.org/10.1016/j.healthpol.2014.10.017

AI HLEG. (2019). Ethics guidelines for trustworthy AI.

Retrived 22 September 2022 from https://digital-

strategy.ec.europa.eu/en/library/ethics-guidelines-trust

worthy-ai.

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.

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

in healthcare: past, present and future. Stroke and

Vascular Neurology 2017;2: e000101. https://doi.org/

10.1136/svn-2017-000101.

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.

Kloehn N. et al. (2018). Improving consumer understanding

of medical text: development and validation of a new

subsimplify algorithm to automatically generate term

explanations in English and Spanish. J. Med. Internet

Res. 20(8), e10779. https://doi.org/10.2196/10779

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.

Leroy G. et al. (2012). Improving perceived and actual text

difficulty for health information consumers using semi-

automated methods. In: AMIA Annual Symposium

Proceedings, pp. 522–531

MacBryde CM, Blacklow RS, eds. (1970). Signs and

Symptoms: Applied Pathologic Physiology and

Clinical Interpretation. 5th edn. Philadelphia:

Lippincott, 1970. 27.

Responsible Artificial Intelligence Institute. https://

www.responsible.ai/faq (accessed on 10th September

2022).

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

Healthcare Transformation. HIMSS. Retrieved

September 2022 from https://www.himss.org/resour

ces/digital-health-framework-healthcare-transformati

on-white-paper.

Snowdon, A., Schnarr K., Alessi, C. (2014). It’s All About

Me : The Personalization of Health Systems. Retrieved

April 2022 from https://www.ivey.uwo.ca/media/

3467873/its-all-about-me-the-personalization-of-heal

th-systems.pdf.

World Health Organization. (2021). Ethics and governance

of artificial intelligence for health: WHO guidance.

Geneva, 2021.

World Health Organization. (2016). Framework on

integrated, people-centred health services: Report by

the Secretariat. World Health Assembly, (A69/39),

April 2016, pp. 1–12. https://apps.who.int/g

b/ebwha/pdf_files/WHA69/A69_39-en.pdf.

World Health Organization. (2003). Improving Data

Quality: A Guide for Developing Countries. 2003. pp.

1–74.

Yujian, L., Bo, L. (2007). A Normalized Levenshtein

Distance Metric," in IEEE Transactions on Pattern

Analysis and Machine Intelligence, vol. 29, no. 6, pp.

1091-1095, June 2007. https://doi.org/10.1109/T

PAMI.2007.1078.

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment

391