Communicating Personalized Risk Factors for Lifestyle Coaching

George Drosatos

, Kyriakos Bakirlis

, Pavlos S. Efraimidis

and Eleni Kaldoudi

School of Medicine, Democritus University of Thrace, Dragana, Alexandroupoli 68100, Greece

Department of Electrical and Computer Engineering, Democritus University of Thrace, Kimmeria, Xanthi 67100, Greece

Keywords: eHealth, Personal Health Systems, Risk Factors, Healthy Behaviour, Lifestyle Related Disease, Mobile

Application, Privacy by Design.

Abstract: Chronic non-communicable diseases such as diabetes, chronic cardiorenal and respiratory disease and

cancer, are serious, burdensome and costly conditions that share a common characteristic: they heavily

depend on common behavioural risk factors, such as physical activity, diet, stress, and substance abuse.

Despite concerted efforts it has been remarkably difficult to change such lifestyle related disease

determinants, as behavioural change is a complex process requiring significant personal responsibility. In

this paper we propose a personal mobile eHealth application to communicate personalized lifestyle related

health risks and understand their individual impact on personal health condition and disease progression.

1 INTRODUCTION

Lifestyle related diseases are multifactorial, chronic,

costly diseases that create a world-wide burden on

individuals, and society. However, they share a

common characteristic: underlying disease

determinants hugely depend on lifestyle choices and

environmental factors, thus can be modified so that

disease can be prevented and its progression can be

halted.

Current trends in health informatics reflect an

important contemporary shift towards citizen

engagement for health, wellbeing and disease

prevention, as opposed to disease management.

However, broadcasting generic health messages (e.g.

‘do this, don’t do that’) has limited effects unless

there is a convincing, easily perceived and

personally customized body of evidence to back

healthy choices.

In this paper we propose a novel mobile eHealth

application to communicate personalized health risks

and coach the user to understand their impact on

various aspects of personal health and plan changes

in lifestyle related risk factors. The proposed

personal application consumes publicly available

open data on health risk evidence via standardized

programming interfaces and meshes these data with

private personal health information. The output is

graphical interactive representation of common

health risks as personalized to each individual.

Privacy is engineered into the design of the

application. The ultimate goal is to provide the

means to foster understanding of the complex

interdependent nature of personal lifestyle choices as

a disease determinant.

2 BACKGROUND

In the second global status report in a triennial series

published in 2014, WHO reports that chronic non-

communicable diseases such as cardiovascular

disease, cancer, diabetes and chronic respiratory

disease, were responsible for 38 million deaths per

year (68% of all deaths) and are projected to

increase from 38 million in 2012 to 52 million by

2030 (WHO, 2014).

For example, the total number of people with

diabetes has risen from 108 million in 1980 to 442 in

2014 (WHO, 2014), reaching epidemic proportions

globally. Diabetes is profoundly impacted by

lifestyle options including dietary intake, exercise,

stress, sleep, and use of alcohol. Interventions which

change each of these lifestyle behaviours are shown

to improve the health of diabetic patients (Spruijt-

Metz, 2014).

Another major category of life style related

diseases includes cardiovascular diseases, including

heart attacks, strokes, coronary heart disease,

cerebrovascular disease and peripheral arterial

Drosatos, G., Bakirlis, K., Efraimidis, P. and Kaldoudi, E.

Communicating Personalized Risk Factors for Lifestyle Coaching.

DOI: 10.5220/0006660405710578

In Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2018) - Volume 5: HEALTHINF, pages 571-578

ISBN: 978-989-758-281-3

571

disease. Behavioural risk factors are responsible for

about 80% of cardiovascular disease instances4 and

include unhealthy diet, physical inactivity, and

tobacco and alcohol use. Overall cardiovascular

disease is currently estimated to cost the EU

economy almost €210 billion a year; of the total cost

of cardiovascular disease in the EU, 53% (€111

billion) is due to health care costs, 26% (€54 billion)

to productivity losses and 21% (€45 billion) to the

informal care of people with cardiovascular disease

(European Heart Network, 2017).

Therefore, WHO has developed programmes to

reduce lifestyle related disease incidence by limiting

risk factors such as tobacco, physical inactivity,

dietary factors, obesity and overweight, alcohol use,

and environmental pollution. Overall, large, multicen-

tre cohort studies (Knoops, 2004; Moreno, 2008;

Haveman-Nies, 2001) have shown that even small

changes in lifestyle can make an important difference

towards health improvement and disease reduction.

Changing such lifestyle related disease

determinants represents the single biggest

opportunity to improve health outcomes while

bringing costs under control. But the ‘easy stuff’ is

far from easy. Despite concerted efforts by policy

makers, providers and payers — not to mention the

best intentions of individuals — it has been

remarkably difficult to effect behavioural change

(Ernst & Young, 2012).

Behavioural change is a complex process

requiring significant personal responsibility. Without

personalized predictive information, it is not

possible to leverage on this. It is also important to

understand your health risks in order to benefit from

news and research about specific diseases and plan

preventing monitoring.

2.1 Risk Evidence in Medicine

A health risk factor is a variable (demographic,

genetic, behavioural, medical, or even environmental)

which when present in an individual increases the

probability of a (usually) negative outcome to occur.

This risk association between the exposure agent and

the outcome is conveyed via a relative risk value.

For example, medical evidence suggests that

obesity is a risk factor for coronary arterial disease

(Guh, 2009). In this particular systematic review and

meta-analysis, men with a body-mass index (BMI)

between 25 and 30 kg/m

were found to have a

relative risk ratio of 1.29 to develop coronary arterial

disease (as compared to normal male of a BMI 18.5

to 25 kg/m

). A different risk association (for the

same risk factor) is found for men of a BMI greater

than 30 kg/m

, who present an elevated relative risk

ratio of 1.72. The same evidence source shows that

women with a BMI between 25 and 30 kg/m

have a

risk ratio of 1.80 to develop coronary arterial disease

and when their BMI is above 30 kg/m

the risk ratio

is elevated to 3.10. Thus, in this example, four

different associations are described between obesity

(and age) and the outcome of coronary arterial

disease.

Risk factors are derived from population

statistical studies. Systematic reviews select similar

studies and via statistical meta-analysis combine

results to improve the estimates of risk ratios and

increase the level of evidence. Several risk

associations are then used to build total risk

estimation models for the prediction of certain

outcomes that are affected by multiple risk factors at

the same time. For example, in cardiovascular

disease, several estimation systems exist, that vary

considerably in terms of the population size and

characteristics, their statistical considerations,

validation, and set of risk factors considered, e.g.,

cardiovascular risk, include the Framingham

equation (Sheridan, 2003), the Joint British Societies

(JBS) formula (Boon 2014), the ASSIGN score

(Woodward, 2007) and the SCORE risk charts

(Perk, 2012).

Conventionally, risk factors form the basis of

medical guidelines and are routinely communicated

to health care professionals to formulate the basis of

clinical patient management.

Evidence on risk is customarily published in

medical scientific press. Indicative of current

research emphasis on health risk prediction is the

exponential increase of published research papers in

the field. To illustrate this, we have placed a crude

PubMed query on risk prediction as follows:

"health risk appraisal"[TIAB] OR

"prediction rule"[TIAB] OR

"prediction rules"[TIAB] OR

"prediction model"[TIAB] OR

"prediction models"[TIAB] OR

"prediction score"[TIAB] OR

"prediction scores"[TIAB] OR "risk

score"[TIAB] OR "risk scores"[TIAB]

OR "risk factor"[TIAB] OR "risk

factors"[TIAB]) AND

("0001/01/01"[PDAT]:

"2016/12/31"[PDAT])

To derive the total number of PubMed

publications, we placed the following query:

"0001/01/01"[PDAT]:

"2016/12/31"[PDAT]

HEALTHINF 2018 - 11th International Conference on Health Informatics

572

Figure 1 shows the number of retrieved

publications per year plotted as a percentage of the

total number of PubMed indexed publications and as

an absolute number. The query retrieved more than

460.000 related publications that amount to almost

2% of the total PubMed indexed publications.

Overall, there is an increase in the absolute number

of retrieved publications over the years. When seen

as the percentage of the overall PubMed corpus

growth, linear regression for the last 40 years shows

a statistically significant increasing trend (regression

coefficient = 0.000898, p-value < 0.001, R-squared

= 98.5%).

Figure 1: Number of PubMed indexed publications per

year retrieved via the query: “health risk appraisal” OR

“prediction rule” OR “prediction rules” OR “prediction

model” OR “prediction models” OR “prediction score”

OR “prediction scores” OR “risk score” OR “risk scores”

OR “risk factor” OR “risk factors” in the fields of title and

abstract, plotted as a percentage of the total PubMed

indexed publications and as an absolute number.

A number of models have been proposed for

capturing various aspects of the clinical research

process that aims to generate new risk evidence.

Examples include the Ontology-Based eXtensible

data model (OBX) (Kong, 2011) and the Ontology

of Clinical Research (OCRe) (Sim, 2014); both aim

to support the process of generating new scientific

knowledge in medicine, rather than actual evidence

on risk prediction. Scientific knowledge on risk

factors is captured by CARRE risk factor ontology

(Third, 2015). This ontology has been used to create

an open data repository of structured information on

risk factor evidence in the medical domain of

cardiorenal disease and comorbidities, as part of a

European Commission funded project CARRE

(FP7-ICT-611140). The repository was developed

by medical experts following a methodology based

on selecting recent highest evidence sources (meta-

analyses where available) and allowing for peer

review of manually entered risk factor data. This

RDF repository is freely available via a SPARQL

endpoint, and currently holds structured descriptions

of more than 100 risk factors, corresponding to a

total of 250 different risk associations between more

than 50 medical conditions.

2.2 Risk Prediction and Patient

Empowerment

Recent years have seen a paradigm shift towards

patient centred care and health management. Under

this view, patient empowerment has emerged as an

approach that can help improve medical outcomes

while lowering costs of treatment by facilitating self-

directed behaviour change. The concept seems

particularly promising in the prevention and

management of chronic diseases, and it is directly

connected with personalized patient services,

education and preventive measures.

At the First European Conference on Patient

Empowerment held in Copenhagen, Denmark on

11–12 April 2012 (WHO, 2012), R. Johnstone of the

International Alliance of Patient Organizations

claimed that: “What needs to happen is for doctors

to come down off their pedestal and for patients to

get up off their knees”. Currently, the research

community appreciates that improving a person's

ability to understand and manage his or her own

health and disease, negotiate with different teams of

health professionals, and navigate the complexities

of health systems is crucial to achieving better health

outcomes (The Lancet Editor, 2012).

A recent pilot study deployed a self-monitoring

intervention for cardiorenal patients coupled with

rich health risk visualizations as derived from the

CARRE repository (Zhao, 2017). The intervention

requires the patient to use a set of personal and

wearable sensors to monitor health status and

couples these real-time data with medical

information retrieved from personal health records.

Personal private data are then coupled with risk

evidence to create medical alerts for the

management and prevention of disease progression

in cardiorenal patients or those at risk of this disease.

A randomized controlled trial (Kizlaitis, 2016)

showed that the intervention statistically increased

empowerment by a 12.4% in metabolic syndrome

patients, and health literacy by 21.3% in patients

with heart failure of chronic kidney disease.

Communicating Personalized Risk Factors for Lifestyle Coaching

573

3 COMMUNICATING

PERSONALIZED HEALTH

RISKS

In this paper we propose a personal eHealth

smartphone application to communicate health risks

to healthy citizens or people at increased risk of

chronic disease. The aim is to provide a personalized

assessment of lifestyle related risks that create the

necessary knowledge background for people to

make educated choices towards healthier lifestyle

adjustments. To achieve this goal, the Risk Coach

application merges open data on health risk evidence

with private health related information.

The core of the application is the current

evidence on health risks. Such evidence is extracted

from the public CARRE risk factor repository

available at https://devices.duth.carre-project.eu/

sparql. Data on risk evidence is in a structured RDF

format and contains information on exposure and

outcome, risk ratio value (with confidence intervals)

and the conditions under which this risk ratio value

is valid. This condition is based on the

characteristics of the population group used to study

the particular risk factor and also includes levels of

exposure. For example, the repository contains

evidence on obesity as a risk factor for cholelithiasis

as reported in a recent meta-analysis (Guh, 2009).

Evidence on this risk factor reveals four different

risk associations, each corresponding to a different

condition and presenting different risk ratio values

(Table 1). In this example, the condition under

which obesity is a risk factor for cholelithiasis is a

combination of the sex and the value of body mass

index (BMI). It is evident that obese females (body

mass index ≥ 30) are at an increased risk of

cholelithiasis (relative risk of 2.32) as compared to

overweight females (25  body mass index < 30)

and as compared to obese males, who appear to have

about the same risk as overweight females. The

exact form of the condition (the variables and their

values) depends on the clinical study and its

population groups. In fact, risk associations in

CARRE repository at present use a total of 95

different variables to formulate conditions, and this

number will most likely increase as new risk factors

are entered in the repository.

In order to personalize risk communication, the

appropriate personal information needs to be known

and compared to the known risk factor conditions in

the public repository. This is ensured via a Risk

Coach application module that allows the user to

input appropriate personal information. To increase

efficacy and reduce privacy concerns, the

application dynamically recognizes different

variables available in the CARRE risk factor

repository and actively builds the personal

information input fields at real time.

By default, the user is required to input some

basic demographic and biometric information such

as age, sex, height and weight. This initial

information is used to calculate any personal risks, if

these apply based on the specific values describing

the user. To investigate whether further risk factors

apply for the particular user, the application queries

the external risk factor repository for all known

variables and dynamically constructs the appropriate

fields for the user to fill in personal values. Personal

risks are calculated by checking which of the risk

association conditions available in the external

database are valid for the user inserted personal data.

Table 1: A risk factor description example showing four different risk associations for the same risk factor, each

corresponding to a different condition and presenting different risk values (main description fields included).

Description fields

Risk association #1

Risk association #2

Risk association #3

Risk association #4

Risk factor:

obesity

[is an issue in]

cholelithiasis

obesity

[is an issue in]

cholelithiasis

obesity

[is an issue in]

cholelithiasis

obesity

[is an issue in]

cholelithiasis

Condition:

(body mass index ≥ 25

AND body mass index

< 30)

AND sex = 'male'

body mass index ≥ 30

AND sex = 'male'

(body mass index ≥

25 AND body mass

index < 30)

AND sex = 'female'

body mass index ≥ 30

AND sex = 'female'

Ratio type:

relative risk

Ratio value:

1.09

1.43

1.44

2.32

Confidence Interval:

0.87 – 1.37

1.04 – 1.96

1.05 – 1.98

1.17 – 4.57

PubMed ID:

19320986

HEALTHINF 2018 - 11th International Conference on Health Informatics

574

3.1 An Architecture to Preserve

Privacy by Design

The Risk Coach application is designed to be used

by the citizens for self-management of health and

disease mainly outside the formal healthcare context.

Based on the requirements for personal data

communication and an analysis of privacy issues in

personal e-health systems (Drosatos, 2016), we can

identify the following functionalities that may

introduce privacy issues in Risk Coach application:

(1) personal data storage and processing; (2)

integration of personalized public data.

When both storage and processing of personal

data are located on a user personal device, privacy

can generally be achieved by default. This approach

is followed in Risk Coach where personal data and

personal risk calculations are stored locally on the

application. Also, all risk calculation is performed

locally on the smartphone device, as shown in

Figure 2 which presents an overview of the

application architecture. factors.

A potentially important and usually elusive

privacy issues may arise when at run time public

data are requested from external sources (in this case

the risk factor repository). Although the data are

publicly available, just the act of linking particular

data to a specific user may cause a privacy violation,

by revealing the user’s health condition. There are a

number of proposed techniques to conceal user

requirements by altering the initial request, e.g. by

expanding and generalizing the request for public

data. These techniques fetch a large amount of data

to the user application and then a second round of

local processing extracts the specific data relevant to

the user (Efraimidis, 2016). Other emerging

approaches require the cooperation of a group of

users in the system to conceal one another’s requests

(Romero-Tris, 2015). An alternative is to use

anonymous network technologies that protect the

physical address of user from the public service. A

representative example is the TOR service

(Dingledine, 2004), which creates a network of

proxies over the internet and allows recursive

message encryption along the chain of proxies. In

the current approach, the entire risk factor dataset is

locally retrieved, so as to hide user specific requests.

3.2 Implementation

The Risk Coach application is currently implemented

for Android mobile devices using Android Studio

(version 2.3.3) and it is compiled for Android 7.1

(API level 25) with backward compatibility until the

Android 4.4w (API Level 20). Animated graphs are

produced using the MPAndroidChart library

(https://github.com/ PhilJay/MPAndroidChart) and

the retrieval of data throw the CARRE SPAQL

Endpoint is performed using the OkHttp library

(https://square.github.io/ okhttp/).

The application supports the following main

functionalities (implemented as individual Android

application fragments): (a) visualize the generic risk

factor evidence available in the public CARRE risk

factor repository, as an overall repository summary

Figure 2: An overview of the Risk Coach mobile application functional modules and architecture.

Communicating Personalized Risk Factors for Lifestyle Coaching

575

(a)

(b)

(c)

Figure 3: Screenshots of Risk Coach mobile Android application.

and browser; (b) create of a local copy personal

health related information, with fields dynamically

compiled based on what information is required to

customize available evidence on risk associations;

and (c) view and interact with personalized risk

information.

Indicative screenshots are shown in Figure 3.

Personalized risk visualization allows the user to

initially explore a summary of personal risk factors.

Each risk can then be expanded in a graph chart to

show all potential outcomes and the particular risk

ratio information. The user can also interact with

personal health information to set lifestyle goals and

observe the expected results in terms of risk

reduction; for example increase or decrease weight

to realize its impact on obesity related risk factors

4 DISCUSSION AND FURTHER

WORK

Patient empowerment is about designing and

delivering health and social care services in a way

that they are inclusive and enable citizens to take

control of their health care needs. According to the

European Network for Patient Empowerment

(ENOPE), the first mandate for an activated patient

is to be able to understand their health condition and

its effect on their body.

The Risk Coach application presented in this

paper addresses directly this basic requirement by

communicating state-of-the-art medical evidence on

common lifestyle related health risk determinants as

personalized to the individual. The application

enables the individual to have an overall view of

his/her own health status; understand potential

health deterioration and disease progression based

on current health status; and visually and

quantitatively investigate the impact each health

determinant may have on various health conditions.

Work in progress involves a randomized

controlled trial to access the efficacy of

communicating health risks to the general public via

the personalized application that preserves privacy

by design. The aim of the study is to assess user

satisfaction and efficacy of the application to

empower people and coach them towards a healthier

lifestyle. Primary objectives are to increase health

literacy, and increase level of empowerment.

Secondary objectives are to improve lifestyle habits

(smoking, physical activity, adherence to self-

monitoring and therapy) and test for application

acceptability and user satisfaction. The study

considers two different population groups: healthy

volunteers and chronic heart disease patients. The

HEALTHINF 2018 - 11th International Conference on Health Informatics

576

control arm will be given general written

information on common lifestyle related risks; the

intervention arm will be given access to the Risk

Coach mobile application.

Health literacy, before and after intervention,

will be assessed via the European Health Literacy

Questionnaire (Sørensen, 2013) and the Lipkus

Expanded Health Numeracy Scale (Lipkus, 2001).

Patient empowerment will be assessed via the

SUSTAINS instrument (Unver, 2013) which

considers three axes: enabling and strengthening

empowerment of patients; enabling better medical

results; and enabling a more efficient use of

healthcare resources and containing costs. Finally,

following the ISO standard ISO 9241 Part 11,

system usability will be measured by taking into

account the context of use of the system — i.e., who

is using the system, what they are using it for, and

the environment in which they are using it. Along

with an informative system evaluation, we will also

deploy the System Usability Score instrument

(Brooke, 1996).

Communicating health risks via interactive

personalized risk visualizations is expected to

increase health awareness, motivate persons to adopt

a healthier lifestyle, and contribute towards

increasing public health literacy and informed

shared decision making.

ACKNOWLEDGEMENTS

The work presented in this paper was partly

sponsored by the FP7-ICT project CARRE (Grant

No. 611140), funded in part by the European

Commission and Greek National Matching funds.

REFERENCES

Boon, N., Boyle, R., Bradbury, K., Buckley, J., Connolly,

S., Craig, S., ..., Wood, D., 2014. Joint British

Societies’ consensus recommendations for the

prevention of cardiovascular disease (JBS3), Heart,

100(Suppl 2), ii1-ii67.

Brooke, J., 1996. SUS: a “quick and dirty” usability scale.

Usability Evaluation in Industry, Taylor and Francis,

189-194.

Dingledine, R., Mathewson, N., Syverson, P., 2004. Tor:

the second-generation onion router. In Proc. of the

13th USENIX security symposium, 303-320.

Drosatos, G., Efraimidis, P.S., Williams, G., Kaldoudi, E.,

2016. Towards Privacy by Design in Personal e-

Health Systems. In Proc. Proceedings of the 9th

International Joint Conference on Biomedical

Engineering Systems and Technologies (BIOSTEC),

5(HEALTHINF), 472-477.

Efraimidis, P.S., Drosatos, G., Arampatzis, A.,

Stamatelatos, G., Athanasiadis, I., 2016. A privacy-by-

design contextual suggestion system for tourism.

Journal of Sensor and Actuator Networks, 5(2), 10.

Ernst & Young, 2012. Global Life Sciences Report,

Global Lifesciences Center Ltd.

European Heart Network AISB, 2017. European

cardiovascular disease statistics – 2017 Edition.

Guh, D. P., Zhang, W., Bansback, N., Amarsi, Z.,

Birmingham, C. L., Anis, A. H., 2009. The incidence

of co-morbidities related to obesity and overweight: a

systematic review and meta-analysis, BMC Public

Health, 9(1), 88.

Haveman-Nies, A., Tucker, K.L., de Groot, L.C.P.G.M.,

Wilson, P.W.F., Van Staveren, W.A., 2001.

Evaluation of dietary quality in relationship to

nutritional and lifestyle factors in elderly people of the

US Framingham Heart Study and the European

SENECA study. European Journal of Clinical

Nutrition, 55(10), 870.

Kizlaitis, R., Bileisiene, N., Puronaite, R., Stundys, D.,

Juozalenaite, G., Smagurauskas, T., Papazoglou, D.,

Zagkas, K., Roumeliotis, S., Passadakis, P., Pouliliou,

S., Drosatos, G., Makris, N., Kaldoudi, E., 2016.

Deliverable 7.4: Use case deployment & evaluation,

CARRE FP7-ICT project No. 611140, 1-117.

Knoops, K.T., de Groot, L.C., Kromhout, D., Perrin, A.E.,

Moreiras-Varela, O., Menotti, A., Van Staveren, W.

A., 2004. Mediterranean diet, lifestyle factors, and 10-

year mortality in elderly European men and women:

the HALE project. JAMA, 292(12), 1433-1439.

Kong, Y.M., Dahlke, C., Xiang, Q., Qian, Y., Karp, D.,

Scheuermann, R.H., 2011. Toward an ontology-based

framework for clinical research databases, J. Biomed.

Inform., 44(1), 48-58.

Lipkus, M., Samsa, G., Rimer, B.K., 2001. General perfor-

mance on a numeracy scale among highly educated

samples. Medical Decision Making, 21(1), 37-44.

Moreno, L.A., Gonzalez-Gross, M., Kersting, M., Molnar,

D., De Henauw, S., Beghin, L., ..., Ortega, F.B., 2008.

Assessing, understanding and modifying nutritional

status, eating habits and physical activity in European

adolescents: the HELENA Study. Public health

nutrition, 11(3), 288-299.

Perk, J., De Backer, G., Gohlke, H., Graham, I., Reiner,

Z., Verschuren, M., Albus, C., Benlian, P., Boysen, G.

et al., 2012. European Guidelines on cardiovascular

disease prevention in clinical practice, Eur Heart J.,

33(13), 1635-1701.

Romero-Tris, C., Viejo, A., Castellà-Roca, J., 2015. Multi-

party methods for privacy-preserving web search:

Survey and contributions. In Advanced Research in

Data Privacy. Springer International Publishing, 367-

387.

Sheridan, S., Pignone, M., Mulrow, C., 2003.

Framingham-based tools to calculate the global risk of

coronary heart disease. Journal of general internal

medicine, 18(12), 1039-1052.

Communicating Personalized Risk Factors for Lifestyle Coaching

577

Sim, I., Tu, S.W., Carini, S., Lehmann, H.P., Pollock, B.

H., Peleg, M., Wittkowski, K.M., 2014. The Ontology

of Clinical Research (OCRe): an informatics

foundation for the science of clinical research. Journal

of biomedical informatics, 52, 78-91.

Sørensen, K., Van den Broucke, S., Pelikan, J. M., Fullam,

J., Doyle, G., Slonska, Z., Kondilis, B., Stoffels, V.,

Osborne, R.H., Brand, H., 2013. Measuring health

literacy in populations: illuminating the design and

development process of the European Health Literacy

Survey Questionnaire (HLS-EU-Q), BMC public

health, 13(1), 948.

Spruijt-Metz, D., O’Reilly, G.A., Cook, L., Page, K.A.

Quinn, C., 2014. Behavioral contributions to the

pathogenesis of type 2 diabetes. Curr Diab Rep,

14(4), 475.

The Lancet Editor, 2012. Editorial: Patient Empower-

ment – who empowers whom?, The Lancet,

379(9827), 1677.

Third, A., Kaldoudi, E., Gkotsis, G., Roumeliotis, S.,

Pafili, K., Domingue, J., 2015. Capturing Scientific

Knowledge on Medical Risk Factors, In Proc. of the

15th International Semantic Web Conference (ISWC),

Springer International Publishing, 566-580.

Ünver, Ö., Atzori, W., 2013. Questionnaire for Patient

Empowerment Measurement Version 1.0 (Document

D3.2), SUSTAINS: Support USers To Access

INformation and Services, EU CT PSP Grant

Agreement No 29720, 1-74.

Woodward, M., Brindle, P., Tunstall-Pedoe, H., 2007.

Adding social deprivation and family history to

cardiovascular risk assessment: the ASSIGN score

from the Scottish Heart Health Extended Cohort

(SHHEC), Heart, 93(2), 172-176.

World Health Organization (WHO), 2012. Empowering

Patients, http://www.euro.who.int/en/what-we-do/

health-topics/noncommunicable-diseases/sections/

news/2012/4/empowering-patients, accessed on

5/10/2017.

World Health Organization (WHO), 2014. Global Status

Report on Non-communicable Diseases 2014,

http://www.who.int/nmh/publications/ncd-status-

report-2014/en/, accessed on 5/10/2017.

Zhao, Y., Parvinzamir, F., Wilson, S., Wei, H., Deng, Z.,

Portokallidis, N., Third, A., Drosatos, G., Liu, E.,

Dong, F., Marozas, V., Lukosevicius, A., Kaldoudi,

E., Clapworthy, G., 2017. Integrated Visualisation of

Wearable Sensor Data and Risk Models for

Individualised Health Monitoring and Risk

Assessment to Promote Patient Empowerment,

Journal of Visualisation, 20(2), 405-423.

HEALTHINF 2018 - 11th International Conference on Health Informatics

578