Developing Accessible eHealth Portals: A Human-centred Approach

Kwangsu Choi

, Hourieh Khalajzadeh

, Humphrey O. Obie

, Jennifer Mcintosh

and Zhidian Lin

Faculty of Information Technology, Monash University, Clayton, Victoria, Australia

Keywords:

Human-centric eHealth Portal, Usability, Accessibility, Personalisation, Design Prototype, Web Design.

Abstract:

eHealth technology is being developed at a rapid rate in line with advances in Information and Communication

Technology (ICT). As eHealth technology increases, ensuring equitable accessibility and usability for all end-

users is becoming a signiﬁcant challenge. Considering human factors in the design phase of an application

will help ensure usability and accessibility gaps are addressed. We conducted a survey with a broad range of

users to investigate the relationship between different human factors and design elements that address these

human factors. From the survey results, we developed a high-ﬁdelity design prototype which we validated

with a usability study to enhance the usability and accessibility for diverse end-users.

1 INTRODUCTION

The rapid changes and advances in Information and

Communication Technology (ICT) have had a huge

impact on the modern healthcare system (Demiris

et al., 2008). eHealth services can improve access

to health service providers, thereby improving clin-

ical decision making, diagnostic accuracy and treat-

ment particularly with remote health care (AIHW,

2021). According to the biennial report 2020 by the

Australian Institution of Health and Welfare (AIHW)

(AIHW, 2021), eHealth services have enabled users

to manage their health and well-being through bet-

ter monitoring and communication with health pro-

fessionals. Internationally, the World Health Organ-

isation (WHO, 2015) has reported a steady increase

throughout a broad cross-section of countries; 64.4%

of WHO members are aware of the importance of

eHealth services and 58% of members have initiated

eHealth systems in their countries (WHO, 2016).

Currently, most mobile and web-based eHealth

applications are designed to support people to man-

age their health when it is not necessary for patients

to physically consult with a doctor or visit a hospital,

e.g., self-managing weight loss. People can access

https://orcid.org/0000-0001-6005-2863

https://orcid.org/0000-0001-9958-0102

https://orcid.org/0000-0002-6322-2984

https://orcid.org/0000-0002-6655-0940

https://orcid.org/0000-0003-3271-8458

many health services using various devices, such as

mobile phones, tablets, and laptops (Kipping et al.,

2016) to receive almost the same level of healthcare

for many conditions using evidence-based internet-

based healthcare services (Das and Svanæs, 2013).

Furthermore, eHealth has been developed to enable

people to make medical appointments, obtain health

information, or order medications, without directly

accessing a health professional, simplifying many of

the processes between patients and health profession-

als (B ˛akała and Korczak, 2010; Sarkar et al., 2011).

Indeed, guidelines or information given in the eHealth

applications are more detailed and provided more fre-

quently (Kipping et al., 2016).

The beneﬁcial aspects of eHealth applications are

not universally available to every user (Sarkar et al.,

2011). eHealth applications are internet-based tech-

nologies, and therefore access to eHealth applications

are dependent on a person’s ability to use the internet

(Zambianchi et al., 2019). This is particularly per-

tinent for elderly users, who have less exposure to

the internet and therefore internet-based eHealth tech-

nology (Report, 2021; Anderson and Perrin, 2017).

As seen in the statistics released by the United Na-

tions (UN, 2017), there is a steady increase in inter-

net usage by the elderly, but the usability and acces-

sibility of eHealth applications have failed to accom-

modate the trend. There may be several reasons for

this gap. One reason is that application developers’

and end-users’ expected outcomes for eHealth appli-

cations vary (Searl et al., 2010). As shown in studies

Choi, K., Khalajzadeh, H., Obie, H., Mcintosh, J. and Lin, Z.

Developing Accessible eHealth Portals: A Human-centred Approach.

DOI: 10.5220/0010981400003176

In Proceedings of the 17th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2022), pages 38-49

ISBN: 978-989-758-568-5; ISSN: 2184-4895

by Das and Svanæs (2013), the age group of the de-

velopers is generally young, and they focus more on

innovativeness and elegance of the technology con-

tained in the application rather than considering the

characteristics of healthcare providers and users.

The gap between developers and users can be

remedied by conducting a person-centred process to

engage relevant stakeholders in the design and de-

velopment of the eHealth application (Demiris et al.,

2008); guidelines exist that provide methods for de-

veloping person-centred applications (Power et al.,

2012), such as Web Content Accessibility Guidelines

(WCAG) 2.0

and WCAG 2.1

. However, the fac-

tors that determine the usability of an application can

vary signiﬁcantly from user to user (Reiners et al.,

2019). Even people with similar accessibility issues

may have other human factors that inﬂuence the way

they use eHealth (Ali et al., 2021). In addition, since

eHealth service customers and providers are highly

likely to have different preferences on both the design

and functional features of an application, the evalua-

tion of usability and user experience of the application

can also signiﬁcantly vary (Tieu et al., 2015).

We aim to identify which human factors to con-

sider in designing an eHealth web portal and to ﬁnd

appropriate methods to apply them to the design of

the application. This also includes identifying which

patients or user groups are likely to be discriminated

against when using eHealth applications. We con-

ducted a preliminary survey with 145 participants

from various age and language groups to explore the

relationships between human factors and design pref-

erences. We provided participants with qualitative

and quantitative questions to collect data regarding

preferences on the colour tone of the portal, typogra-

phy, navigation style, and icon/image style. Based on

the results, we developed a high-ﬁdelity design pro-

totype for an eHealth web portal for people needing

clinical access to Occupational Therapists (OT). A us-

ability study of the prototype veriﬁed the effective-

ness of the design factors. Our methods contribute to

providing the direction of design in the development

of other human-centred eHealth applications.

http://www.w3.org/TR/WCAG20/

http://www.w3.org/TR/WCAG21/

2 BACKGROUND

2.1 eHealth Portal and Human-centred

Design

For any eHealth application development, identify-

ing the characteristics and requirements of the poten-

tial users or patient groups should be prioritised (Tieu

et al., 2015). In the case of an eHealth portal, it is

vital to use effective and familiar design features for

targeted user groups (Kildea et al., 2019). This is be-

cause the portal is in general used by non-health pro-

fessionals, such as patients or their guardians in con-

junction with their healthcare providers to acquire rel-

evant and accurate health information while building

and maintaining a relationship between clinician and

patient (B ˛akała and Korczak, 2010). Well-designed

eHealth services not only facilitate patient-carer inter-

actions but also enable to ease the digital divide and

to equalise the knowledge balance between them so

that patients can manage their health more effectively

(Kildea et al., 2019).

According to ISO 9241-210 (2019) (ISO 9241-

210:2019, 2019), human-centric design is deﬁned as:

“Human-centred design is an approach to in-

teractive systems development that aims to

make systems usable and useful by focusing on

the users, their needs and requirements, and

by applying human factors/ergonomics, and

usability knowledge and techniques.”

The four design phases for a human-centric design

plan, according to (ISO 9241-210:2019, 2019), are:

• Phase 1: Understanding and specifying the con-

text of use;

• Phase 2: Specifying the user requirements;

• Phase 3: Producing design solutions;

• Phase 4: Evaluating the design

Capturing considerable human factors during the re-

quirements phase (Phase 2) for an eHealth portal

might include basic demographic information as well

as computer competency, internet accessibility, phys-

ical attributes, preferences, and/or habits (Rot et al.,

2017). Not all factors are easy to understand as they

are different depending on individuals’ characteris-

tics (Castilla et al., 2016). It is important to consider

that even if the functional requirements are fully cap-

tured fulﬁlled, the design elements may adversely af-

fect the evaluation of usability and accessibility of the

application (Boulos, 2003; Holzinger, 2002). There-

fore, the design plan for a human-centred eHealth

Developing Accessible eHealth Portals: A Human-centred Approach

portal should be organised according to the correla-

tion between key human factors and design elements

(Kalimullah and Sushmitha, 2017).

2.2 Human Factors

The absence of consideration of human factors, which

may include Age, Gender, Race/Ethnicity, Education

attainment, Disability, Culture/Language, and Emo-

tional state, can have a profound impact on the ac-

cessibility and usability of a web application (Rein-

ers et al., 2019). however, it is difﬁcult to correctly

deﬁne and standardise which human factors should

be prioritised given the diversity of human factors in

each individual (Tieu et al., 2015). Human factors

that should be mainly considered for web application

development may change depending on the purpose

of the application and target user group (Das et al.,

2015). In the case of eHealth applications, different

researchers (Goel et al., 2011; Sarkar et al., 2011; Ali

et al., 2021; Reiners et al., 2019; Tieu et al., 2015;

Ancker et al., 2011; Kahn et al., 2009; Shamsujjoha

et al., 2021; Yong et al., 2012) cited human factors as

age, gender, race/ethnicity, education attainment, dis-

ability, culture, language, and emotional state as im-

portant human factors to consider; age and language

being two of the main ones.

Age, as a human factor, has the greatest inﬂuence

on physical capability changes and older age is the

biggest risk factor for chronic diseases or disabilities,

and therefore has a signiﬁcant impact on eHealth ser-

vices (LaMonica et al., 2017). According to a fact

sheet released by Report (2021), more than half of

the elderly (deﬁned as 65 years or older) responded

that they are not able to use new electronic devices or

applications without assistance from others. In Aus-

tralia, 87% of the elderly with disabilities use the In-

ternet, but only 9% of them use eHealth services as

they encounter barriers (Ali et al., 2021). This inter-

generational adoption gap is in part because physi-

cal characteristics of the elderly are not properly or

sufﬁciently reﬂected in the development stage, but

also because their non-physical characteristics, such

as preferences, habits, and emotions, are not consid-

ered (Castilla et al., 2016). In the same vein, con-

sidering human factors in design phases of an appli-

cation for people with disability, including for those

with visual impairment, hearing impairment, and cog-

nitive impairment including dementia, may also pos-

itively inﬂuence the usability of eHealth application

(Das et al., 2015).

Language is another major factor that adversely

impacts the usability and accessibility of eHealth ap-

plications. People often experience difﬁculty using

web applications in languages other than their usual

language, or people if they have low health literacy

or low language proﬁciency. People with language

barriers ﬁnd sourcing information online difﬁcult and

may be forced to ﬁnd information only through direct

access with face-to-face health services, limiting their

access to online eHealth options.

2.3 Vulnerable Groups

Vulnerable groups include the elderly, people living

with disabilities and recent immigrants who are likely

to have cognitive, physical, language and cultural bar-

riers, or poor health literacy (LaMonica et al., 2017;

Power et al., 2012; Ferati et al., 2012; Yong et al.,

2012). Although there are guidelines, such as WCAG

2.0 and WCAG 2.1, to improve usability for people

with physical disabilities, developers are often reluc-

tant to comply with them for reasons such as difﬁcul-

ties in identifying human factors precisely, the cost of

implementation, and implementing the corresponding

functionalities practically. Power et al. (2012) demon-

strated that following those guidelines does not al-

ways mean the absolute improvement of usability, but

that websites complying with the guidelines outper-

form in certain areas, such as contents recognition and

contents navigation, for people with disabilities.

3 APPROACH

Our method consists of 5 main stages: (1) Preliminary

survey; (2) Analysis of the participants’ responses; (3)

Development of a high-ﬁdelity design prototype; (4)

User study with a group of participants; (5) Analysis

of user study results. In this section, we discuss our

survey and prototype design.

3.1 Preliminary Survey

The preliminary survey was conducted anonymously

and distributed using Google Form. The participants

we recruited for this survey were from various age

groups (from 18 to 65 and over). As the authors were

ﬂuent in multiple languages, the survey was translated

and distributed in Korean and Chinese as well as En-

glish. The survey included general descriptive char-

acteristics, such as Gender, Age, Current residence,

Nationality and Language, Internet access and us-

age, Education attainment, Occupation, and Income

and also questions to capture the perception and us-

age of eHealth services. In addition, questions were

included to explore the design elements may be inﬂu-

enced by the individual’s human factors to investigate

ENASE 2022 - 17th International Conference on Evaluation of Novel Approaches to Software Engineering

Figure 1: eHealth web application sam-

ple 1 (https:// www.mydr.com.au/ ).

Figure 2: eHealth web application sam-

ple 2 (https:// www.betterhealth.vic.gov

.au/a-z).

Figure 3: eHealth web application sam-

ple 3 (https:// kidney.org.au/).

the design preferences of eHealth applications. For

these design-related questions, sample images were

provided to support uniformity in the interpretation of

questions and the accuracy of the responses. In case

any of the participants with less experience in the use

of eHealth services, we also provided three sample

eHealth websites as shown in Fig. 1-3. Fig. 1 is a

sample representing a web application consisting of

warm colours and images, Fig. 2 cool colours and

text-only icons, and Fig. 3 dark colours and texts.

All the websites had different styles, such as overall

colour tones, functionalities, icon styles, structures,

etc., so that we could identify greater discrimination

between participants’ responses. Using the same ex-

amples, participants were also asked to select speciﬁc

design elements that they felt would be suitable for

eHealth applications. The survey questions in Korean

and Chinese and distributed to participants who use

these languages in order to obtain more diverse and

accurate responses.

3.2 High-ﬁdelity Design Prototype

Based on the analysis of the survey data, a high-

ﬁdelity design prototype was created using Figma

Adee

and Able

plugins were installed and used to

check the WCAG 2.1 compliance. Interactions were

also created for all the page elements and functional-

ities for the user study. All images used in this pro-

totype, including icons and photographs, were either

self-created, free or open source images from online

image archives

and were cross-checked with both

plugins to satisfy AAA grade for the success crite-

ria stated in WCAG 2.1. As shown in the study con-

ducted by Castilla et al. (2016), the preferences on

the navigation styles by age groups are distinctly dif-

ferent. To maximise the usability of the prototype, we

strived to provide as diverse but intuitive and concise

navigation as possible. For instance, when required

https://www.ﬁgma.com

https://www.adeeapp.com

https://www.ﬁgma.com/community/plugin/73469388

8346260052/Able-

to show the additional contents that are related to the

current page, pop-up windows or notiﬁcations were

used to avoid loading a new page. This can mitigate

the of users experiencing confusion or failing to con-

tinue their task due to too frequent navigations from

page to page. One key functionality of this prototype

is that the page is customisable using an Accessibil-

ity Helper. The Accessibility Helper button is pre-

sented at the ﬁxed position, bottom left of all pages

so that the user can access it from anywhere in the

prototype. The available options for this functionality

are depicted in Fig. 4.The functions included in the

Accessibility Helper were designed based on the pre-

liminary survey and reviews of other studies regard-

ing the relationship between design features and us-

ability/accessibility for elderly/disabled people. The

design of this function also follows WCAG 2.1.

Figure 4: Options included in the Accessibility Helper.

3.2.1 Accessibility Helper

Options contained in the Accessibility Helper are:

1. Options for colour.

Images are taken from (https://icons8.com/illustratio

ns), (https://www.freepik.com/), and (https://icons8.com/)

The prototype is accessible via https://bit.ly/2THOQ

8K, and details of pages, design elements, funcitonalities,

and interactions via https://bit.ly/2W4kJcO

Developing Accessible eHealth Portals: A Human-centred Approach

Figure 5: Main page with cool tone (top) and warm tone

(bottom).

– Change Theme (Fig. 5): This enables users to

change colour from a cool tone to a warm tone

or vice versa.

– Reverse Contrast (Fig. 6) & Desaturate (Fig.

7): These options enable users to change the

contrast and saturation of the page.

Figure 6: Option: Reverse Contrast.

Figure 7: Option: Desaturate.

2. Options for text.

– Enlarge Text (Fig. 8): This increases the size of

all texts.

– Change Text Spacing & Line Spacing (Fig. 9):

This resizes spaces between characters.

Figure 8: Option: Enlarge Text.

Figure 9: Option: Change Text Space & Line Space.

3.2.2 Navigation Types

Fig. 10 shows different navigation types used in the

prototype.

1. Primary navigation (top menu) – this is ﬁxed at

the top of the page and enables users to access all

available contents,

2. Side navigation: this enables users to access dif-

ferent pages within the same level,

3. Sub-navigation (Hierarchical navigation): this en-

ables a user to see the paths and to navigate back

to previous paths,

4. Sub-navigation: this enables users to navigate

within a section, and

5. Footer navigation: this presents all accessibility

menus as well as important information.

4 RESULT

This section presents the preliminary survey results.

ENASE 2022 - 17th International Conference on Evaluation of Novel Approaches to Software Engineering

Figure 10: Navigation types used in the prototype.

4.1 Demographics

A total of 145 participants participated in this survey;

89 (61%) were female, 55 (38%) were male and 2

(1%) did not disclose their gender. All participants

were categorised into ﬁve different age groups. The

distributions of each age group are shown in Fig. 11.

Figure 11: Distribution of age groups.

Among all participants, 32 (21.9%) responded

that they currently suffer from at least one chronic

disease, and 9 (6.2%) participants were unsure about

their chronic conditions. According to the responses,

participants’ education attainment was high with 77

(53%) participants having completed at least a Bach-

elor’s degrees, but only 12 (8.2%) participants did not

have a high school diploma. Only 11 (4.8%) par-

ticipants responded that they spoke English as their

main language and the majority of the participants

133 (91.7%) identiﬁed as Asian and spoke Korean,

Chinese, Thai, etc., as their main language. All par-

ticipants except one had access to the Internet, and

131 (89.7%) of people self-evaluated their computer

skills as more than moderate.

4.2 Analysis Methods

Descriptive statistics were used to analyse the data

collected from the survey. A Chi-square test was

conducted to investigate the association between key

sociodemographic factors and design elements. Age

was considered the major sociodemographic factor of

interest, and thus the statistical signiﬁcances between

age and design elements were mainly investigated.

For the group of participants who had never used an

eHealth application (100 of 145, 68.5%), emotional

states or the ﬁrst impressions from the given sample

eHealth applications were also investigated using the

Chi-square test. This was conducted to clarify the sig-

niﬁcance of design elements for eHealth applications

from potential users as well as relationships between

users’ emotions and design elements. To determine

the association between more than three variables,

the Multinomial Logistic Regression model was used.

Two sociodemographic factors were selected as inde-

pendent variables to determine the relationships with

design elements and the usage of eHealth applica-

tions. Using the model, we identiﬁed the relationships

of design choice with age and gender, the relationship

of colour choice with age and chronic disease, the re-

lationship of eHealth usage with age and gender, and

the relationship of eHealth usage with age and com-

puter skill. The alpha levels for both methods were set

to 0.05. Microsoft 365 Excel was used for all analy-

sis and Real Statistics Add-in (Zaiontz, 2021) was in-

stalled to calculate the Multinomial Logistic Regres-

sion.

4.3 Association between Human Factors

and Design Elements

We determined “Age” as the key human factor as it is

highly related to other factors, such as the likelihood

of having a chronic disease, computer skills, prefer-

ences, etc., and thus it was mainly used to investigate

the association with design elements (Ali et al., 2021;

Reiners et al., 2019; Kalimullah and Sushmitha, 2017;

LaMonica et al., 2017; Schaller et al., 2016).

4.3.1 Design Type vs. Age (+ Gender)

Figure 12: Design types selected by participants by age

groups.

The preference on the design type is clearly distin-

guishable between generations. Younger people were

more likely to choose Material over Flat and Skeuo-

morphism designs while older people were more

likely to prefer Skeuomorphism to other designs as

seen in Fig. 12. According to the Chi-square test, the

associations between Age and design types were sta-

Developing Accessible eHealth Portals: A Human-centred Approach

tistically signiﬁcant with χ

of 22.1710 and p-value

of 0.0143 at α = 0.05.

We also tested how participants’ design selections

were associated with age and gender through a Multi-

nomial Logistic Regression model as gender is an-

other human factor that often determines preferences

(Goel et al., 2011; Reiners et al., 2019; Ancker et al.,

2011). The model showed that gender was not a sta-

tistically signiﬁcant factor that determined a design

type as seen in Table. 1. For any given gender, the

tendency to select the design type was highly depen-

dent on a participant’s age. The model forecasts that

the likelihood of choosing Material design decreases

as age increases: the 45 years and over participants

were 42.43% more likely to choose the Skeuomor-

phism design rather than Material design. However,

the older generations did not show clear design pref-

erences over the younger generation. Less than 23%

of younger people would be more likely to choose

Skeuomorphism design while over 34% of older peo-

ple would be expected to choose Material design.

Table 1: Summary of Associations between design ele-

ments and human factors.

Chi-square p-value Alpha Signiﬁcance

Design type vs. Age 22.1710 0.0143 0.05 Yes

+ Gender 10.6338 0.1004 0.05 No

Icon types vs. Age 18.9094 0.0414 0.05 Yes

Font size vs. Age 64.4726 5.11E-10 0.05 Yes

Colour vs. Age 10.2934 0.4151 0.05 No

+ Gender 2.9792 0.5613 0.05 No

+ Chronic disease 0.8354 0.9337 0.05 No

Emotional state vs. Age

Sample 1 25.0458 0.0493 0.05 Yes

Sample 2 22.5657 0.0938 0.05 Yes

Sample 3 11.5834 0.7103 0.05 No

4.3.2 Icon Types vs. Age

Participants’ preferences for icon types were highly

tilted towards “Both icon and text”. It was observed

that 80.82% of participants chose icons with both im-

age and text across all age groups. As shown in Table.

1, the result of the Chi-square test showed that age

and icon types were not independent of each other and

the relationship between them was statistically signif-

icant with χ

of 18.9094 and a p-value of 0.0414 at

α = 0.05.

4.3.3 Font Size vs. Age

The relationship between age and font size was also

statistically signiﬁcant with χ

of 64.4726 and a very

low p-value of 5.11E-10 at α = 0.05 as shown in Ta-

ble. 1. However, this result may not be accurate be-

cause the image we provided in the Google form au-

tomatically responded to the screen size it was dis-

played on. This might imply that the font sizes shown

in mobile phones and computer monitors would be

different. Therefore, we used this result to investigate

participants’ general preferences for font size rather

than ﬁnding out the exact font sizes they preferred.

4.3.4 Colour vs. Age (+ Gender, + Chronic

Disease)

The ‘cool’ tone was preferred over the ‘warm’ tone

(54.8% compared to 37.7% respectively) for eHealth

applications, however, there was no statistically sig-

niﬁcant relationship between colour tone and Age

(χ

= 10.2934 and p-value = 0.4151 at α = 0.05).

These results were not changed by applying gender

and chronic disease along with age to the Multino-

mial Logistic Regression model. The model shows

that the relationship of Colour with Age and Gen-

der is not statistically signiﬁcant (χ

is 2.9792 and p-

value is 0.5613 at α = 0.05). With Age and Chronic

disease, there was no statistical signiﬁcance between

those variables with colours (χ

is 0.8354 and p-value

is 0.9337 at α = 0.05).

4.3.5 Emotional State (Sample 1, 2, and 3) vs

Age

Figure 13: Summary of participants’ emotional states.

Participants were asked to choose up to 3 different op-

tions for the ﬁrst impression on each sample website.

Options given to the participants were joy, trust, sur-

prise, neutral, anticipation, disgust, fear, and sadness.

We recategorised the selections into Positive, Neu-

tral, with Negative: Positive included joy, trust, and

surprise

; Neutral included neutral and anticipation

;

Negative included disgust, fear, and sadness. The

summary of participants’ emotional states is shown in

Fig. 13. The chart illustrates the summary of partici-

pants’ ﬁrst impressions and emotions for each sample

website shown in Fig. 1, Fig. 2, and Fig. 3. After cat-

egorising all data, we conducted a Chi-squared test to

ﬁnd if there was a relationship between age and emo-

tions for each sample website design. As shown in

ENASE 2022 - 17th International Conference on Evaluation of Novel Approaches to Software Engineering

Table. 1, the associations between Age and Sample 2

and 3 were not statistically signiﬁcant. However, our

results showed that the proportion of positive emo-

tions that responded to sample 3 (26.4%) were signif-

icantly lower than those of sample 1 (42.45%) and 2

(35.48%). These results are in line with the partici-

pants’ comments about each sample website. Partic-

ipants mentioned that they feel comfortable with the

bright tone, regardless of colour, and they preferred

a website with sufﬁcient white spaces, to help them

to distinguish contents and functionality more easily.

Participants who left comments on sample 3 said that

they feel uncomfortable because the page was con-

structed with too much text and small font sizes, mak-

ing it stressful to read the content.

5 EVALUATION

We designed the prototype presented in Section 3.2,

according to the survey results. To ensure the us-

ability of the prototype, we conducted a user study

with nine participants. We recruited the participants

among those who answered positively to the survey

question that asks the intention of participating in

the user study. The age distribution of participants

included four people in their 20s, two in their 30s,

and three people in their 60s. Among younger par-

ticipants, no one had chronic diseases or disabilities,

however, two out of three participants aged over 60 re-

ported that they cannot properly read contents in nor-

mal text size, which is generally 16 ft, without mag-

nifying reading glasses due to eyesight impairment.

The ﬁrst author conducted the user studies as a

mediator. During the user study, the mediator pro-

vided participants with the necessary information and

observed participants’ behaviours. The behaviours

and time spent on each task were recorded as well.

The user study began by providing a brief explanation

about the project and provided instructions to install

and execute the prototype. Each participant was then

given some basic tasks and corresponding question-

naires. Tasks were categorised into two parts. In the

ﬁrst part, participants were asked to navigate page to

page so that they can understand the structure of the

prototype as well as the design elements used in the

prototype. The second part of the tasks asked par-

ticipants to use the functionalities contained in the

prototypes, which included personalising features of

The surprise emotion is in general considered as a pos-

itive emotion and anticipation emotion as a neutral emotion

in Korean and Chinese. As the majority of our participants

are originally from those countries, we classiﬁed surprise as

a positive and anticipation as neutral emotion.

“Accessibility Helper”. After ﬁnishing each section,

participants also completed the corresponding ques-

tionnaires

. Many participants shared their thoughts

regarding the prototype with us and provided invalu-

able feedback as well.

This user study was conducted using Zoom and

held for 30 to 35 minutes on average, including time

for providing instruction and explanation about the

project. All participants were asked to join the user

study using either desktop or laptop. To complete all

the tasks, participants spent approximately 13 to 15

minutes. Overall, participants’ responses to the pro-

totype were signiﬁcantly positive. Participants scored

4.625 out of 5 for satisfaction on the user interface

(design components), 4.875 for the overall function-

ality, and 4.875 for the overall satisfaction.

We could not observe any signiﬁcant problem in

using navigation features, however, we were able to

identify how participants varied in the way their nav-

igation between generations. A participant in his

20s, performed almost all navigation-related tasks us-

ing footer navigation, while another participant in her

20s, used side and sub-navigation actively. According

to this participant’s comment on the navigation func-

tionalities, she found that the structure of the proto-

type was very intuitive as it was comfortable to navi-

gate between pages and contents using different navi-

gation options. In contrast to the use of various navi-

gation by young participants, it was observed that par-

ticipants over 60 tried to perform most of the tasks

using only primary navigation. Even when complet-

ing tasks using side navigation, they ﬁrst looked at

the primary navigation to ﬁnd the corresponding func-

tionality. For example, although the functionality for

converting displayed information from Map view to

List view was distinctly presented on the side menu,

all participants in their 60s tried to ﬁnd the relevant

feature from the top menu. A 62-year-old female par-

ticipant commented that she looks habitually at the

top menu ﬁrst, regardless of tasks. Details of each

navigation type are shown in Fig 10.

By observing participants’ performances for com-

pleting given tasks, we could identify barriers they

encountered. The most crucial factor that adversely

affects the usability and accessibility of the prototype

was linguistic differences. Due to the limitations of

the design prototype, we were not able to provide an

option to change to different languages. The design

prototype that we provided was all in English, thus

participants with low English proﬁciency had difﬁ-

culty completing the tasks. This was especially ev-

ident when accessing speciﬁc information that con-

The details of tasks and questionnaires provided to the

participants can be found at https://bit.ly/36SoQdW.

Developing Accessible eHealth Portals: A Human-centred Approach

tained unfamiliar words, such as medical terms as

highlighted by one participant.

"It was difﬁcult for me to access certain in-

formation as I have never heard of such

terms, like dementia and dietetics. But, for

me, even with the same medical terminology,

the one with the picture was easier to ﬁnd."

[P5/Female/20s]

One of the participants, who was originally Korean, in

his 60s and was not well versed in English, could not

perform the task properly without support due to the

language barrier. The mediator translated both tasks

and contents in the prototype into Korean to help him

process the user study. This barrier was observed not

only in non-English speakers but also in participants

who use English as their second language. When car-

rying out the task of ﬁnding speciﬁc information con-

taining medical terms, three participants had to spend

more time compared to other participants who speak

English as their main language. The time spent on

each task for all participants is shown in Table 2.

Table 2: User study participants and time spent on each task.

Gender Age Main Language Eng. Prof.

P1 Male 60s Korean VL 4’5” 8’48”

P2 Male 30s Korean H 3’00” 4’3”

P3 Female 30s English VH 3’25” 4’27”

P4 Male 20s Chinese M 4’15” 4’15”

P5 Female 20s Japanese H 4’15” 4’59”

P6 Female 20s English VH 2’35” 3’30”

P7 Female 60s Korean L 7’3” 4’1”

P8 Male 60s Korean M 4’4” 5’36”

P9 Female 20s English, Cantonese VH 3’16” 3’36”

Eng. Prof. stands for English proﬁciency. VH, H, M, L, and VL stand for Very

High, High, Moderate, Low, and Very Low, respectively.

T1: Time spent on completing task 1

T2: Time spent on completing task 2

Similarly, we could ﬁnd evidence of how the use of

certain images enhanced accessibility. Another par-

ticipant was not able to understand the purpose of

icons yet he could infer the functions through the

icon’s shape, location, and image. This was consis-

tent with the result of the preliminary survey that we

conducted. According to the survey results, more

than 80% of participants responded that they pre-

ferred icons with both text and image as those icons

are more intuitive and effective to convey meanings.

In this user study, we also received various feed-

back related to the Accessibility Helper. In general,

young participants responded that it was interesting

and new to them, while older participants emphasised

the need for this functionality. A male participant in

his 60s revealed considerable satisfaction in changing

theme colour function. As per his comment:

"If a health-related web application is com-

prised of too many different colours, I feel re-

jected because its atmosphere is too hectic.

Also, in the case that the colour used in such a

web application is too dark, I hesitated to use

it as it gives me some negative feelings. I think

the Accessibility Helper will contribute to util-

ising this application more precisely and fre-

quently because users can choose the colour

to suit their personal taste." [P1/Male/60s]

His comment was also consistent with participants’

responses regarding the colour of the eHealth applica-

tion as a number of participants also pointed out that

the factor that causes the bad ﬁrst impression is the

colour used in the application. They felt discomfort

if a health application is too colourful or too dark. In

addition, a participant with poor eyesight mentioned

that the function that enabled her to change the font

size and text space was very impressive and helped

her to read and understand contents. She also pro-

vided feedback that such functionalities are essential

for older populations like herself as vision impairment

is one of the most common problems that the elderly

have. A male participant in his 60s emphasised that

the guidelines for the functionalities should be pro-

vided to maximise the usability:

"I have never seen this function, called Ac-

cessibility Helper, before. I think this func-

tion will be very useful for the elderly, those

with disabilities, and even ordinary users. It

can be especially helpful for people with vi-

sion impairment, old people like me. But for

the elderly, it is hard to even know if there is

such function or not unless a clear explana-

tion is given." [P8/Male/60s]

One common feedback received from elderly par-

ticipants was that personalisation features should be

saved individually without having to change them ev-

ery time they access the application. This was an im-

portant point to consider for future work.

6 DISCUSSION

For healthcare-related web applications, it is impor-

tant to precisely consider human factors that have a

signiﬁcant impact on the usability and accessibility

to design elements because the main stakeholders,

such as the elderly, disabled people, and other pa-

tients, are highly likely to have a different physical

or psychological characteristics. As shown in the re-

sults of our survey responses from 145 diverse par-

ticipants, the preferences for design elements, such

as font sizes and icon types, are distinguishable by

ENASE 2022 - 17th International Conference on Evaluation of Novel Approaches to Software Engineering

age groups. Although the relationship between age

and colour of the web application was not statistically

signiﬁcant, most participants responded that their im-

pressions and emotions towards an application were

determined by the colour and therefore it should be

considered as one of the main design elements.

We developed a Figma-based high-ﬁdelity design

prototype for an eHealth web application. This pro-

totype targets those who regularly seek occupational

therapists due to their physical or mental impair-

ments, or their guardians. To narrow the gaps re-

sulting from the difference in human factors between

our participants, we designed a functionality called

Accessibility Helper. This feature accommodates the

users’ needs by allowing them to change each design

element based on their preferences. Participants who

used this feature in the user study did not reveal any

negative emotions regarding the design elements that

they had from the sample web application we pro-

vided in the survey. Moreover, in the user study, we

found that the usability of web applications for each

participant was exacerbated by different design ele-

ments. We hypothesised that the enhancement of the

usability and accessibility resulting from the modify-

ing design elements is dependent on the distinctive

characteristics people have. This ﬁnding was consis-

tent with the evidence we explored initially as shown

in the results of the preliminary survey.

In particular, we were able to clearly see the cor-

relation between people’s characteristics and prefer-

ences on the design elements in the responses pro-

vided by participants aged less than 35 and partici-

pants aged over 55. We found that younger partici-

pants focused on the novelty and rarity of the Accessi-

bility Helper rather than considering the effectiveness

of this function. In other words, we could observe that

they focus heavily on functional factors of the proto-

type when evaluating the usability rather than consid-

ering how the function might be beneﬁcial to them.

This is likely because compared to the older partic-

ipants, younger participants have more opportunities

to use a variety of web applications and have less hes-

itation to adopt new technology [8]. Moreover, an-

other reason why they may prioritise the functional

aspects is that they are less likely to have disabilities

or other physical impairments that hinder them from

accepting design elements consisting of web applica-

tions [23]. Based on the observation from the user

study, the only factor that impressed them was the

“Changing theme” option that affects the overall at-

mosphere of the prototype; the other design options

they stated that it would be useful for those who are

visually impaired or disabled.

In the case of older participants, the deteriora-

tion of usability due to a decline in physical capabil-

ity is a factor that should be considered more impor-

tant than dissatisfaction from the functional aspects

(Kalimullah and Sushmitha, 2017). We found this

from the feedback provided by elderly participants

as well. One common disability or impairment that

elderly participants have is the eyesight impairment

due to ageing (AIHW, 2000). All three participants

in their 60s reported that they are unable to read con-

tent with small-sized texts which signiﬁcantly affects

usability and accessibility. In the same vein, they em-

phasised that the typography-related functions, such

as changing font size, text spacing, and line spac-

ing, are essential features for them. In addition, the

structure of the application is another signiﬁcant fac-

tor that should be considered. When designing the

prototype, we assumed that the elderly would have

relatively lower proﬁciency or frequency of using IT

technology compared to younger people. Therefore,

in line with the research conducted by Castilla et al.

(2016), we designed the structure of the prototype

as intuitive and concise as possible by using pop-up

windows and notiﬁcations and by providing various

navigations. What we found from the observation

of the user study was consistent with the assumption

as older participants who are aged over 60 could not

properly perform the tasks that require using different

navigations rather than linearly navigating between

pages. They completed most of the tasks using only

the primary navigation, and in the case of tasks that

could not be performed with the primary navigation,

they tried to ﬁnd the way from the menu closest to the

contents they have.

Regardless of the age groups, some participants

encountered difﬁculty in using the prototype due to

linguistic differences. The usability and accessibility

gaps caused by the absence of health literacy or dig-

ital literacy are difﬁcult challenges for developers to

solve as it is least likely to estimate the universal level

of those literacies of their stakeholders (Tieu et al.,

2015). As shown in the work of Holzinger (2002), one

efﬁcient approach to solve these disparities is to attach

appropriate images to the contents that users may not

be able to understand easily. In our user study, we

found that participants tried to perform the tasks by

analogy through images for several tasks containing

medical terminologies. They also commented that

they were able to complete the tasks by assuming

through images shown in the icons and contents even

though they had no idea of the meaning of the words.

This proves that usability and accessibility can be en-

hanced by visualising contents appropriately.

Developing Accessible eHealth Portals: A Human-centred Approach

7 LIMITATION

An important limitation of our work is that we did

not have people with substantially severe disabilities

in the participants. Although the Accessibility Helper

contains options to change the contrast and saturation

of the contents, we could not verify their effectiveness

because none of our participants had colour blindness.

However, we veriﬁed that all contents used in this pro-

totype are visible for those with colour blindness or

colour deﬁciencies by cross-checking with the colour

simulator

website and Figma plugins.

Secondly, the quantitative results from our work

were likely to have confounding bias. We relied

mainly on univariate statistics and the largest number

of variables included in one relationship was three.

Moreover, the fact that the age distributions of partici-

pants favoured particular groups is likely to lead to an-

other bias. The third limitation is that the sample pop-

ulation for both the preliminary survey and user study

did not include participants aged under 18. According

to a report released by AIHW (2020), approximately

9% to 10% of people who are in general classiﬁed

as children and adolescences have disabilities, which

is higher on average than the age group between 20

and 35 who was one of the main participant groups

involved in this work.

Finally, since our work is based on the high-

ﬁdelity design prototype, the usability and accessi-

bility we investigated may differ from actual applica-

tions with practically runnable functionalities. With

the design prototype we have, it is difﬁcult to estimate

the usability changes resulting from the frequency of

the use and the time spent for each functionality.

8 CONCLUSIONS

We identiﬁed the correlation between human factors

and design elements through a preliminary survey

and demonstrated the effectiveness of this correlation

through a Figma-based high-ﬁdelity design prototype.

In the user study to investigate the effectiveness of the

prototype, we found barriers that hinder usability. In

addition, by looking at the preference or priority by

age groups, we were able to understand which de-

sign elements should be primarily considered to en-

hance the usability for each age group. This study

contributes in understanding how human factors can

be applied to eHealth applications. In future, we will

aim to focus on implementing this prototype on a web

application by reﬂecting on ﬁndings from this work to

Colblinder: https://www.color-blindness.com/coblis-c

olor-blindness-simulator/

enhance the usability and accessibility of the applica-

tion. Moreover, the resulting web application will be

tested and veriﬁed by more diverse user groups.

ACKNOWLEDGMENT

Support for this work from ARC Laureate Program

FL190100035 and Discovery Project DP200100020

is gratefully acknowledged.

REFERENCES

AIHW (2000). Disability and ageing: Australian population

patterns and implications. Report, Australian Institute

of Health and Welfare.

AIHW (2020). People with disability in australia. Report,

Australian Institute of Health and Welfare.

AIHW (2021). Digital health. Report, Australian Institute

of Health and Welfare.

Ali, M. A., Alam, K., Taylor, B., and Ashraf, M. (2021).

Examining the determinants of ehealth usage among

elderly people with disability: The moderating role of

behavioural aspects. International Journal of Medical

Informatics, 149:104411.

Ancker, J. S., Barrón, Y., Rockoff, M. L., Hauser, D.,

Pichardo, M., Szerencsy, A., and Calman, N. (2011).

Use of an electronic patient portal among disad-

vantaged populations. Journal of general internal

medicine, 26(10):1117–1123.

Anderson, M. and Perrin, A. (2017). Tech adoption climbs

among older adults. Report, Pew Research Center.

B ˛akała, A. and Korczak, K. (2010). Accessibility of e-

health services for people with disabilities. Infor-

matyka Ekonomiczna, 18.

Boulos, M. N. K. (2003). The use of interactive graphi-

cal maps for browsing medical/health internet infor-

mation resources. International Journal of Health Ge-

ographics, 2(1):1–14.

Castilla, D., Garcia-Palacios, A., Miralles, I., Breton-

Lopez, J., Parra, E., Rodriguez-Berges, S., and

Botella, C. (2016). Effect of web navigation style

in elderly users. Computers in Human Behavior,

55:909–920.

Das, A., Faxvaag, A., and Svanæs, D. (2015). The impact

of an ehealth portal on health care professionals’ in-

teraction with patients: qualitative study. Journal of

medical Internet research, 17(11):e267.

Das, A. and Svanæs, D. (2013). Human-centred methods in

the design of an e-health solution for patients under-

going weight loss treatment. International journal of

medical informatics, 82(11):1075–1091.

Demiris, G., Afrin, L. B., Speedie, S., Courtney, K. L.,

Sondhi, M., Vimarlund, V., Lovis, C., Goossen, W.,

and Lynch, C. (2008). Patient-centered applications:

use of information technology to promote disease

management and wellness. a white paper by the amia

ENASE 2022 - 17th International Conference on Evaluation of Novel Approaches to Software Engineering

knowledge in motion working group. Journal of the

American Medical Informatics Association, 15(1):8–

13.

Ferati, M., Vogel, B., Kurti, A., Rauﬁ, B., and Astals, D. S.

(2012). Web accessibility for visually impaired peo-

ple: requirements and design issues, pages 79–96.

Springer.

Goel, M. S., Brown, T. L., Williams, A., Hasnain-Wynia,

R., Thompson, J. A., and Baker, D. W. (2011). Dis-

parities in enrollment and use of an electronic pa-

tient portal. Journal of general internal medicine,

26(10):1112–1116.

Holzinger, A. (2002). User-centered interface design for

disabled and elderly people: First experiences with

designing a patient communication system (pacosy).

In International Conference on Computers for Handi-

capped Persons, pages 33–40. Springer.

ISO 9241-210:2019 (2019). Ergonomics of human-system

interaction — part 210: Human-centred design for in-

teractive systems. Standard.

Kahn, J. S., Aulakh, V., and Bosworth, A. (2009). What

it takes: characteristics of the ideal personal health

record. Health affairs, 28(2):369–376.

Kalimullah, K. and Sushmitha, D. (2017). Inﬂuence of de-

sign elements in mobile applications on user experi-

ence of elderly people. Procedia computer science,

113:352–359.

Kildea, J., Battista, J., Cabral, B., Hendren, L., Herrera, D.,

Hijal, T., and Joseph, A. (2019). Design and devel-

opment of a person-centered patient portal using par-

ticipatory stakeholder co-design. Journal of medical

Internet research, 21(2):e11371.

Kipping, S., Stuckey, M. I., Hernandez, A., Nguyen, T., and

Riahi, S. (2016). A web-based patient portal for men-

tal health care: beneﬁts evaluation. Journal of medical

Internet research, 18(11):e294.

LaMonica, H. M., English, A., Hickie, I. B., Ip, J., Ireland,

C., West, S., Shaw, T., Mowszowski, L., Glozier, N.,

and Duffy, S. (2017). Examining internet and ehealth

practices and preferences: survey study of australian

older adults with subjective memory complaints, mild

cognitive impairment, or dementia. Journal of medi-

cal Internet research, 19(10):e358.

Power, C., Freire, A., Petrie, H., and Swallow, D. (2012).

Guidelines are only half of the story: accessibility

problems encountered by blind users on the web. In

Proceedings of the SIGCHI conference on human fac-

tors in computing systems, pages 433–442.

Reiners, F., Sturm, J., Bouw, L. J., and Wouters, E. J.

(2019). Sociodemographic factors inﬂuencing the use

of ehealth in people with chronic diseases. Interna-

tional journal of environmental research and public

health, 16(4):645.

Report (2021). Internet/broadband fact sheet. Technical

report, Pew Research Center.

Rot, A., Kutera, R., and Gryncewicz, W. (2017). De-

sign and assessment of user interface optimized for

elderly people. a case study of actgo-gate platform. In

ICT4AgeingWell, pages 157–163.

Sarkar, U., Karter, A. J., Liu, J. Y., Adler, N. E., Nguyen, R.,

López, A., and Schillinger, D. (2011). Social dispari-

ties in internet patient portal use in diabetes: evidence

that the digital divide extends beyond access. Jour-

nal of the American Medical Informatics Association,

18:318–321.

Schaller, S., Marinova-Schmidt, V., Setzer, M., Kondylakis,

H., Griebel, L., Sedlmayr, M., Graessel, E., Maler,

J. M., Kirn, S., and Kolominsky-Rabas, P. L. (2016).

Usefulness of a tailored ehealth service for informal

caregivers and professionals in the dementia treatment

and care setting: the ehealthmonitor dementia portal.

JMIR research protocols, 5(2):e4354.

Searl, M. M., Borgi, L., and Chemali, Z. (2010). It is time

to talk about people: a human-centered healthcare sys-

tem. Health research policy and systems, 8(1):1–7.

Shamsujjoha, M., Grundy, J., Li, L., Khalajzadeh, H., and

Lu, Q. (2021). Human-centric issues in ehealth app

development and usage: A preliminary assessment. In

IEEE International Conference on Software Analysis,

Evolution and Reengineering (SANER), pages 506–

510.

Tieu, L., Sarkar, U., Schillinger, D., Ralston, J. D., Ratana-

wongsa, N., Pasick, R., and Lyles, C. R. (2015). Bar-

riers and facilitators to online portal use among pa-

tients and caregivers in a safety net health care sys-

tem: a qualitative study. Journal of medical Internet

research, 17(12):e4847.

UN (2017). World population ageing 2017 – highlights.

Report, United Nations.

WHO (2016). Global diffusion of ehealth: making universal

health coverage achievable, report of the third global

survey on ehealth. Report, World Health Organiza-

tion.

Yong, J. Y., Stvilia, B., and Mon, L. (2012). Cultural inﬂu-

ences on seeking quality health information: An ex-

ploratory study of the korean community. Library &

Information Science Research, 34(1):45–51.

Zaiontz, C. (2021). Real statistics resource pack for excel.

Zambianchi, M., Rönnlund, M., and Carelli, M. G. (2019).

Attitudes towards and use of information and commu-

nication technologies (icts) among older adults in italy

and sweden: The inﬂuence of cultural context, socio-

demographic factors, and time perspective. Journal of

cross-cultural gerontology, 34(3):291–306.

Developing Accessible eHealth Portals: A Human-centred Approach