mHealthSwarm: A Unified Platform for mHealth Applications
Ben Philip
1 a
, Yasmeen Alshehhi
1 b
, Mohamed Abdelrazek
1 c
, Scott Barnett
1 d
,
Alessio Bonti
1 e
and John Grundy
2 f
1
Deakin University, Burwood, VIC, Australia
2
Monash University, Clayton, VIC, Australia
Keywords:
mHealth Apps, mHealth Platform, micro-mHealth Apps.
Abstract:
Mobile health (mHealth) applications are ubiquitous and offer several benefits such as easier access to one’s
health and wellness data through smartphones. However, their growth in popularity has also introduced several
challenges for both end-users and developers. Users face challenges around the poor user experience (UX)
introduced by the need to install several apps and limited customizability which is exacerbated by the limited
control over app functionality. While features common across different apps may not directly affect individual
developers, the fact that one app may provide a better implementation of features leads to wasted developer
effort. A single platform that can satisfy all user needs does not currently exist, and given the diversity of
the mHealth domain, a single app would be far too complex if it existed. In this paper, we present a new
approach and a platform for mHealth apps - micro-mHealth apps, and we discuss them as an alternative to the
current mHealth app development model, which we believe could improve the current state of mHealth app
development and adoption. We are currently evaluating our prototype with several micro-mHealth apps built
using common features in the mHealth apps available in commercial app stores.
1 INTRODUCTION
mHealth has been defined as the “medical and public
health practice supported by mobile devices, such as
mobile phones, patient monitoring devices, personal
digital assistants (PDAs), and other wireless devices”
(WHO, 2011). Today, several health, fitness and well-
ness applications for uses such as drug dosage and
medical reference (Ventola, 2014; Berauk et al., 2017;
Sezgin et al., 2017) and weight management (Hig-
gins, 2015) are popular among users as part of their
daily routines (Berauk et al., 2017; Higgins, 2015).
Despite the increase of mHealth apps and interest
in adopting them from a great many end users, there
are many limitations around existing mHealth app us-
ability, feature relevance, and data sharing. There are
an increasing number of users needing to install many
mHealth apps, a key challenge as there is no app that
delivers a complete set of features that a user might
need (Berauk et al., 2017). Users dislike downloading
multiple apps as they unnecessarily increase the num-
a
https://orcid.org/0000-0001-9438-4532
b
https://orcid.org/0000-0002-9432-9477
c
https://orcid.org/0000-0003-3812-9785
d
https://orcid.org/0000-0002-3187-4937
e
https://orcid.org/0000-0003-2240-0454
f
https://orcid.org/0000-0003-4928-7076
ber of applications on their smartphones (van Velsen
et al., 2013). Similarly, many users stop using health
apps due to the burden of repeated data entry, hidden
costs and confusion around several apps (Krebs and
Duncan, 2015; Wiechmann et al., 2016).
Data visualization is also a key component of
health tracking apps that helps users to track their
goals, habits, and achievements (Lee et al., 2020).
However, studies have shown diversity in data visu-
alization audiences (Jena et al., 2021), which in turn
introduce challenges related to the user experience
around mobile data visualization (Gu et al., 2018)
(Katz et al., 2017). For example, users generally de-
sire more control over charts to better monitor their
health progress and goals (Alshehhi et al., 2022a).
Further, another user survey around the strengths
and weaknesses of mHealth app visualizations high-
lighted the lack of interaction, such as zooming capa-
bilities, as a major limitation (Alshehhi et al., 2022b).
The ability to customize personal goals on charts was
also highlighted as one of the desired features. This,
in turn, indicates that the current data visualization
guidelines lack support for inclusive design, resulting
in a poorer experience.
From the engineering side, developing apps re-
quires writing large amounts of boilerplate code (Bar-
nett et al., 2015) and doing so for several different
Philip, B., Alshehhi, Y., Abdelrazek, M., Barnett, S., Bonti, A. and Grundy, J.
mHealthSwarm: A Unified Platform for mHealth Applications.
DOI: 10.5220/0011985600003464
In Proceedings of the 18th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2023), pages 605-612
ISBN: 978-989-758-647-7; ISSN: 2184-4895
Copyright
c
2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
605
platforms greatly increases cost and development ef-
fort (Pinto and Coutinho, 2018), which cross-platform
frameworks aim to simplify. Given the fragmentation
among the mHealth apps and current platforms, ef-
forts have been made to overcome these challenges.
This includes frameworks for cross-platform develop-
ment such as Ionic, Cordova and React Native. Al-
though these (limited) frameworks can be used for
developing mHealth apps (Banos et al., 2015), they
still do not solve the problem of data sharing between
apps, integration of wearables and peripherals, a uni-
form UX/UI and app features customization (Anas-
tasiadou et al., 2019). While recent works (Inupaku-
tika et al., 2020) have attempted to develop cross-
platform mHealth solutions that focus on simplify-
ing app development, they fall short on customizabil-
ity and integration with other services. The develop-
ment of a “super-app” that covers all major aspects of
health was previously suggested by Higgins as a so-
lution to such problems. However, this has its own
set of challenges, such as many unused features, di-
verse user needs and great maintenance issues (Hig-
gins, 2015).
There is thus a need for a new approach and plat-
form to accelerate and support building better collab-
orative mHealth apps with focused capabilities that
can be customized as needed. We propose an ap-
proach we term “micro-mHealth apps”, and we hy-
pothesize that they will offer a better experience for
both end users and mHealth app developers saving a
lot of effort, time and money. This paper discusses
our motivation and past work supporting this position,
a high-level solution architecture and platform de-
tails, and discusses the details of developing a micro-
mHealth app prototype.
2 MOTIVATION
2.1 Motivating Example
Consider a blood pressure management scenario.
Many blood pressure tracker apps are available in
commercial app stores. Unfortunately, users often
end up downloading different apps with several un-
used features, and feature overlaps between them
(Philip et al., 2022a; Alshehhi et al., 2022b). In the
case of data visualization, users can find that their
apps present similar visualizations to track their blood
pressure (figure 1) an example of redundant fea-
tures. In addition, related information such as weight
and glucose measurements are also often presented
separately. In such cases, this case, users may need
to switch between two or more apps to find mean-
ingful insights derived from different metrics. A pre-
vious study highlighted these challenges when users
need to use more than one mHealth app, and user ex-
pectations, suggesting the need for feature customiza-
tion that would allow end users to tailor mHealth apps
to their needs (Philip et al., 2022a). Our solution to
this problem proposes combining several separate app
features micro mHealth apps into one, to provide
a highly customizable, unified mHealth platform for
end users.
Figure 1: App screenshot showing overlapping charts with
single unique feature.
2.2 User Challenges and Expectations
Alshehhi et al. conducted a review of user perspec-
tives around visualizations in mHealth apps that indi-
cated several difficulties in chart customizability, in-
teractivity and functionality (Alshehhi et al., 2022a).
A subsequent anonymous survey allowed them to
further investigate these issues, which indicated that
the needs of several users have to be considered to
improve the acceptance of mHealth data visualiza-
tions (Alshehhi et al., 2022b). Similarly, Philip et
al. (Philip et al., 2022a) conducted an online sur-
vey of people having experience with mHealth apps
to understand challenges when using several mHealth
apps and expectations from future apps. While poor
design was highlighted as one reason for poor app
adoption, the participants indicated several other chal-
lenges such as additional, unnecessary features, fea-
ture overlaps between different apps, and repeated
manual data entry that can inadvertently drive users
away The survey showed that users expect a lot more
from current apps and prefer a single platform that
allows feature customization to build an application
meeting all their needs without any additional bloat.
These studies motivated us to consider the develop-
ment of a unified mHealth platform to combine to-
gether micro mHealth app features, and we hypothe-
size that such platforms can solve both the developer
and end-user challenges.
ENASE 2023 - 18th International Conference on Evaluation of Novel Approaches to Software Engineering
606
2.3 Mini Apps
Mobile apps can be classified into three categories
native apps, mobile web apps and hybrid apps (Inu-
pakutika et al., 2020) with their development largely
depending on their use. Web and hybrid apps are sug-
gested for better cross-platform compatibility, while
native or hybrid apps are suggested for better periph-
eral support (Serrano et al., 2013). mHealth apps can
be built using any approach and although the three
different formats of mobile apps offer some advan-
tage over each other, they are restricted to single apps
- i.e. each package or URL is a single application that
runs or renders on the user’s smartphone. Hybrid apps
offer the benefit of using a web application wrapped
inside a native container, which is potentially more
beneficial for developers as the web component can
be shared across platforms. However, end-users may
not see any difference or obtain any direct benefits
through such designs.
The concept of “Mini-apps” was introduced in a
working draft white paper (W3C, 2019) submitted to
the W3C as a new format of mobile apps based on
web technologies that integrates with and behave like
native apps. Based on hybrid apps, they offer bene-
fits such as being free from installation and storage
constraints with the ability to access platform capa-
bilities through native container apps. These apps can
be considered across two layers - (1) the Logic Layer
responsible for handling the business logic, internal
state and integration with system data and services;
and (2) the View Layer for rendering the user inter-
face (UI) and handling user interaction.
This structure can be compared to a traditional
web application, with the views constructed using
HTML and CSS, and the app logic handled using
JavaScript. UIs can be rendered using native compo-
nents such as WebViews which also act as an interface
between mini apps and the native container app. Ac-
cess to resources such as storage and hardware (cam-
era, GPS etc.) cannot be provided directly through
the web interface, but through custom APIs provided
by the container that allows controlled access to sys-
tem resources. By extending hybrid applications into
such containers, a whole ecosystem of mini apps can
be deployed in a way that facilitates data sharing be-
tween them and allows a user to create custom apps
with features tailored to their individual needs.
We build on top of this idea for health/fitness/well-
ness apps by introducing a single, highly customiz-
able super-app and provide a much better health app
experience by avoiding the challenges of full apps
(Liu et al., 2019; Liu et al., 2020).
3 OUR PLATFORM
To address the requirements identified in the previ-
ous section, we have developed a novel mHealth-
Swarm unified mHealth platform. This provides a
framework for developing health and wellness ser-
vices and abstracts device features and low level oper-
ations into simple high level functionality, providing a
highly customizable environment that can be person-
alized by adding/removing single-function mHealth
services/apps.
Our mHealthSwarm platform uses a registration
mechanism that records app permissions when linked
to a user’s account. Each time a service app needs
user data or access to underlying device hardware, it
needs to make a request to the platform using exposed
APIs. These requests are then checked, and then the
appropriate resource is provided to the requesting ap-
plication. The following section presents a very high-
level overview of the main architectural components.
Our mHealthSwarm platform architecture is out-
lined in Figure 2. It is divided into four components,
which are briefly described below.
Application Manager: The Application Manager
combines the view and logic layers and can be con-
sidered as a task manager that keeps track of active
micro-apps and provides an interface to the container
(i.e., our micro-mHealth app platform) APIs.
Data Manager: A gateway for all health data pro-
duced and consumed by hosted micro-apps, along
with raw data collected from health peripherals. Al-
though this component currently only supports local
data storage, future extensions are planned to support
remote storage and management.
Device Manager: The component is responsible for
handling sensors connected to the smartphone and
provides an interface for collecting raw data produced
by these sensors.
Conversation Manager: Although our platform could
reduce the need for several health apps, users may still
have to navigate through the micro-apps.
The ecosystem of conversational interfaces have
been expanding rapidly
1
and studies around their use
have also suggested their usefulness in performing
simple tasks such as scheduling appointments (Palan-
ica et al., 2019). Given the availability of numerous
health metrics on our platform, we believe it is a suit-
able environment for a chatbot to help users interact
with.
Our current prototype aims to provide end users
a better experience by providing complete customiz-
ability of functions through micro-health-apps. It is
1
https://www.businessinsider.com/chatbot-market-stats
-trends?r=AU\&IR=T
mHealthSwarm: A Unified Platform for mHealth Applications
607
Figure 2: Architecture overview of our platform.
implemented as an Android application, with the sup-
porting framework for micro-mHealth apps imple-
mented as a back end service using Node.js. The plat-
form is designed to facilitate access to device hard-
ware such as Bluetooth and GPS modules along with
other embedded sensors (IMUs, camera) through ab-
stract functions. However, given the very limited use
of external sensors by currently available apps (Philip
et al., 2022b), the current implementation has limited
support for peripheral devices, with plans for support-
ing more in the future. The micro-mHealth apps do
not have any control over the device hardware, but
are only passed the data from these sensors via the
platform.
Data storage currently uses the Room Persis-
tence Library for predefined data types, with addi-
tional support for app-specific storage in separate
files. The platform interface allows all registered
micro-mHealth apps with the appropriate permissions
to fetch and store the required data. Data transfer
between the app and the remote service uses HTTP
methods with the data wrapped in JSON objects.
Our mHealthSwarm platform provides a set of
mini apps, which form the basis of our micro-mHealth
apps. These are structured like web applications,
which reduces the learning curve required for de-
veloping mHealth apps, while also making them
platform-independent. Interaction with the platform
is handled through exposed APIs called from the logic
layer, a brief example of which is presented in the next
section.
While native container apps for different mobile
operating systems will have to be developed, they of-
fer the benefit of cross-platform support for the micro-
mHealth apps they host. To ‘install’ services, users
will just need to link those micro-mHealth apps avail-
able on the platform to their accounts, during which
permissions will be established. Launching a micro-
mHealth app will launch the feature droplet in the
container, which then handles app requests and pro-
vides data to the running micro-mHealth app after val-
idating permissions. The overall implemented flow of
data and platform events are outlined in figure 3.
4 USAGE EXAMPLE
Consider an app such as MapMyWalk
2
with the core
feature of tracking walks. A micro-mHealth app can
be built with the same core feature, eliminating ad-
ditional, unnecessary functionality, allowing users to
choose the app functionality they need. To demon-
strate this, we developed a micro-mHealth app on our
platform
3
where the service keeps track of a user’s lo-
cation, step count and heart rate (if available) during
a workout session and allows them to review the data
later. The following code snippet from the JavaScript
logic layer outlines a very basic request for the user’s
location. It shows a call to the platform API for log-
ging the GPS coordinates in the database and setting
a callback reference to the function that handles the
location data. This data is then used by the micro-
2
https://play.google.com/store/apps/details?id=com.ma
pmywalk.android2
3
https://github.com/benphilip1991/del-container
ENASE 2023 - 18th International Conference on Evaluation of Novel Approaches to Software Engineering
608
Figure 3: An overview of our micro-application container data-flow.
mHealth app to provide the walk tracking feature.
Figure 5 (right) shows the micro app implementing
this feature, and figure 4 shows the overall control
and data flow for this app after it is launched. This
snippet also indicates how access to sensor data and
other resources can be greatly simplified through our
proposed approach.
1 function a pp In it () {
2 let a pp Re q ue st = {
3 // i nj ec te d on l au nc h
4 " ap pId " : this. appId ,
5 " r eq ue st s " : [{
6 " reso ur ce " : " lo ca ti on " ,
7 " call ba ck " : " pr oc es s Lo c " ,
8 " t og gl e " : true
9 }]
10 };
11
12 // Log da ta in the d at ab as e
13 D el Ut il s . se t Se n s or L og g er R equ e st (
14 JS ON . s tr ing if y ( a p pR eq ue s t ));
15
16 // R eg is te r c al lb a ck
17 D el Ut il s . se t Ca l lb a ck R eq u es t (
18 JS ON . s tr ing if y ( a p pR eq ue s t ));
19 }
20
21 // R eg is te r ed cal lb ac k for GPS da ta
22 function p ro c es sL oc ( dat aTy pe , l oc ) {
23 // p ro ce ss loc at io n da ta
24 }
Listing 1: App data request registration example.
Figure 4: Main interaction between the platform and a
micro-app after launch.
The sample application is built using JavaScript,
HTML and CSS with the OnsenUI framework
4
for
creating a native-like interface. Although the core
business logic for the micro-mHealth apps would re-
quire a similar effort as developing native code for
similar features, the benefit of this approach is ap-
4
https://onsen.io
mHealthSwarm: A Unified Platform for mHealth Applications
609
Figure 5: Prototype with three “installed” micro-apps (left);
our micro-app implementation of a walk tracker (right).
parent when considering the single-function nature of
these app droplets and the overall simplification of ac-
cess to device resources and shared user data.
The novelty of the platform for mHealth comes
from a highly customizable environment where each
micro-mHealth app provides a single function and can
be added/removed as required, enabling end users to
create personalized health apps. To evaluate our plat-
form and its feature customizability with end users,
we implemented core functions (i.e., workout track-
ing, meal tracking, and mood tracking) inspired from
3 popular apps. Our evaluations are still ongoing, and
preliminary results are discussed in the next section.
To further enhance the customizability, we also in-
troduced semi-customizable charts for different use-
cases such as tracking one’s meals (figure 5 right),
and while the current implementation is very lim-
ited, further extensions to visualization components
are planned to extend the range of personalization on
the platform.
This design is expected to remove unnecessary de-
velopment overheads from developers, who no longer
have to worry about low level details such as device
integration and data management. This may also open
up the domain to a wider developer community in ad-
dition to mobile app developers, allowing for more
innovation in the domain. The hybrid app design is
also expected to simplify updates as a verified service
update will be made available to all users on release
without the need for manual search and installation.
5 DISCUSSION
Although the Android and iOS operating systems of-
fer native support for health data aggregation with
frameworks such as Google Fit and Apple Health,
they still suffer from a common issue - the reliance
on more than one health app. However, given the
popularity of hybrid applications and wide commer-
cial success of mini-app frameworks for messaging
and eCommerce (e.g., WeChat
5
), they appear to be a
suitable alternative to native services. This approach
offers the benefits of hybrid applications while pro-
moting an open, customizable platform of mHealth
services. Despite the advantages, this customizabil-
ity comes with some drawbacks, such as an added
overhead for learning the platform functionality and
slightly lower micro-app performance. However, we
believe that these are minor issues considering the
overall benefits the platform offers.
In terms of functionality, our mHealthSwarm plat-
form is currently limited to an external heart rate mon-
itor, built-in sensors, GPS and camera. The platform
is also currently limited to Android, with a future ex-
pansion planned for supporting iOS. We are currently
also exploring ways to improve data visualization
in micro-mHealth apps by using platform-provided
components. This is expected to simplify the devel-
opment aspects and provide more flexibility to end-
users. Similarly, given the highly specific and limited
nature of the platform and its capabilities, a domain-
specific language for developing micro-mHealth apps
is also being considered as a future extension to fur-
ther simplify app development by generating micro-
mHealth app components. However, despite the po-
tential advantages, the platform currently suffers from
limited support for existing apps and business models,
and requires further investigation.
Our mHealthSwarm platform is currently being
evaluated with end-users as well as mobile app de-
velopers to get their feedback. Preliminary user eval-
uations have indicated a preference for this approach,
with the participants stating they like the feature cus-
tomizability offered by the platform. This is high-
lighted by one participant’s comment - “What I def-
initely like about the platform is the customization -
that I can select the features that I use”. So far, almost
all users have indicated that the platform was easy to
learn, with emerging themes from participant inter-
views highlighting the ease of use, the benefits of fea-
ture customization, and the ease of data management
on a single platform. Similarly, ongoing developer
evaluations have also seen mostly positive feedback,
with the main highlight being the reduced develop-
5
https://www.wechat.com/
ENASE 2023 - 18th International Conference on Evaluation of Novel Approaches to Software Engineering
610
ment effort. A developer’s comment - “...this seems
to be a bit more like web development, so yeah [it’s]
quite simple... with less boilerplate code, compared to
regular mobile apps” - highlights the potential ben-
efits for app developers. However, major concerns
such as the limitations of the platform in its current
state (including limited support for testing and de-
bugging micro-mHealth apps, and limited documen-
tation) when compared to current mobile app devel-
opment frameworks were also highlighted.
Overall, our work in this domain focuses on pro-
viding a better framework and platform for micro-
mHealth apps as a means to achieve the following
goals: (1) To provide a consistent platform for bet-
ter collaboration and communication between the dif-
ferent apps; (2) To foster an ecosystem of micro-
mHealth apps built for specific functions, eliminat-
ing unnecessary components and bloat; (3) To reduce
the increasing device resource requirements and con-
tention of several large health apps on one device; and
(4) To provide a better user experience and increase
mHealth adoption.
6 SUMMARY
In this paper, we presented our mHeallthSwarm plat-
form of a micro-mHealth platform that addresses ma-
jor challenges with current mHealth apps such as the
need to install more than one app, unnecessary bloat
in the applications and challenges with data sharing
and management. We discuss implementation details
of our platform with a brief data flow sequence dia-
gram showing a high-level flow of app requests and
data. We then briefly discuss a sample app and how
it can be developed using the platform APIs. Overall,
our approach for developing micro-mHealth apps will
guide the design and development of novel mHealth
apps and platforms, and in the future have a positive
impact on the adoption of mHealth services.
ACKNOWLEDGEMENTS
Philip is supported by Deakin University Schol-
arship and ARC Research Hub IH170100013.
Grundy is supported by ARC Laureate Fellowship
FL190100035.
REFERENCES
(2019). MiniApp Standardization White Paper, W3C First
Public Working Draft.
Alshehhi, Y., Abdelrazek, M., and Bonti, A. (2022a). Anal-
ysis of personal data visualisation reviews on mobile
health apps. In ACHI 2022, The Fifteenth Interna-
tional Conference on Advances in Computer-Human
Interactions, pages 111–118. IARIA.
Alshehhi, Y. A., Abdelrazek, M., and Bonti, A. (2022b).
Needs and Challenges of Personal Data Visualisations
in Mobile Health Apps: User Survey.
Anastasiadou, D., Folkvord, F., Serrano-Troncoso, E., and
Lupia
˜
nez-Villanueva, F. (2019). Mobile Health Adop-
tion in Mental Health: User Experience of a Mo-
bile Health App for Patients With an Eating Disorder.
JMIR mHealth and uHealth, 7(6):e12920.
Banos, O., Villalonga, C., Garcia, R., Saez, A., Damas, M.,
Holgado-Terriza, J. A., Lee, S., Pomares, H., and Ro-
jas, I. (2015). Design, implementation and validation
of a novel open framework for agile development of
mobile health applications. BioMedical Engineering
OnLine, 14(2):S6.
Barnett, S., Vasa, R., and Grundy, J. (2015). Bootstrap-
ping mobile app development. In Elbaum, S. and Can-
fora, G., editors, Proceedings - 2015 IEEE/ACM 37th
IEEE International Conference on Software Engineer-
ing, ICSE 2015, volume 2, pages 657–660, United
States of America. IEEE, Institute of Electrical and
Electronics Engineers.
Berauk, V., Murugiah, M., Soh, Y., Sheng, Y., Wong, T.,
and Ming, L. C. (2017). Mobile Health Applications
for Caring of Older People: Review and Comparison.
Therapeutic Innovation & Regulatory Science, page
216847901772555.
Gu, J., Mackin, S., and Zheng, Y. (2018). Making sense:
an innovative data visualization application utilized
via mobile platform. In 2018 IEEE 20th Interna-
tional Conference on High Performance Computing
and Communications; IEEE 16th International Con-
ference on Smart City; IEEE 4th International Con-
ference on Data Science and Systems (HPCC/SmartC-
ity/DSS), pages 1105–1109. IEEE.
Higgins, J. (2015). Smartphone Applications for Patients’
Health & Fitness. The American journal of medicine.
Inupakutika, D., Kaghyan, S., Akopian, D., Chalela, P., and
Ramirez, A. G. (2020). Facilitating the development
of cross-platform mHealth applications for chronic
supportive care and a case study. Journal of Biomedi-
cal Informatics, 105:103420.
Jena, A., Butler, M., Dwyer, T., Ellis, K., Engelke, U.,
Kirkham, R., Marriott, K., Paris, C., and Rajaman-
ickam, V. (2021). The next billion users of visual-
ization. IEEE Computer Graphics and Applications,
41(2):8–16.
Katz, D., Dalton, N., Holland, S., O’kane, A., and Price,
B. A. (2017). Questioning the reflection paradigm for
diabetes mobile apps. In eHealth 360°, pages 315–
326. Springer.
Krebs, P. and Duncan, D. T. (2015). Health app use among
us mobile phone owners: A national survey. JMIR
mHealth uHealth, 3(4):e101.
Lee, B., Choe, E. K., Isenberg, P., Marriott, K., and Stasko,
J. (2020). Reaching broader audiences with data visu-
mHealthSwarm: A Unified Platform for mHealth Applications
611
alization. IEEE Computer Graphics and Applications,
40(2):82–90.
Liu, Y., Xu, E., Ma, Y., and Liu, X. (2019). A First Look
at Instant Service Consumption with Quick Apps on
Mobile Devices. In 2019 IEEE International Confer-
ence on Web Services (ICWS), pages 328–335. IEEE.
Liu, Z.-Y., Liao, K., and Wang, J.-L. (2020). Research on
the influence of we-chat applet on improving user ex-
perience. In Proceedings of the Korean Society of
Computer Information Conference, pages 613–615.
Korean Society of Computer Information.
Palanica, A., Flaschner, P., Thommandram, A., Li, M., and
Fossat, Y. (2019). Physicians’ Perceptions of Chatbots
in Health Care: Cross-Sectional Web-Based Survey. J
Med Internet Res, 21(4):e12887.
Philip, B., Abdelrazek, M., Barnett, S., Bonti, A., and
Grundy, J. (2022a). Towards Better mHealth Apps:
Understanding Current Challenges and User Expec-
tations. In 2022 IEEE/ACM 9th International Con-
ference on Mobile Software Engineering and Systems
(MobileSoft), pages 33–37.
Philip, B. J., Abdelrazek, M., Bonti, A., Barnett, S., and
Grundy, J. (2022b). Data Collection Mechanisms
in Health and Wellness Apps:Review and Analysis.
JMIR mHealth and uHealth, 10(3).
Pinto, C. M. and Coutinho, C. (2018). From Native to
Cross-platform Hybrid Development. In 2018 Inter-
national Conference on Intelligent Systems (IS), pages
669–676.
Serrano, N., Hernantes, J., and Gallardo, G. (2013). Mobile
web apps. IEEE Software, 30(5):22–27.
Sezgin, E.,
¨
Ozkan-Yildirim, S., and Yildirim, S. (2017).
Investigation of physicians’ awareness and use of
mHealth apps: A mixed method study. Health Pol-
icy and Technology, (3):251–267.
van Velsen, L., Beaujean, D. J., and van Gemert-Pijnen,
J. E. (2013). Why mobile health app overload drives
us crazy, and how to restore the sanity. BMC Medical
Informatics and Decision Making, 13(1):23.
Ventola, C. L. (2014). Mobile devices and apps for health
care professionals: uses and benefits. P & T : a peer-
reviewed journal for formulary management, (5):356–
364.
WHO (2011). mHealth: New horizons for health
through mobile technologies: second global survey on
eHealth, volume 3 of Global observatory for eHealth
Series, 3. World Health Organization.
Wiechmann, W., Kwan, D., Bokarius, A., and Toohey,
S. L. (2016). There’s an App for That? Highlighting
the Difficulty in Finding Clinically Relevant Smart-
phone Applications. The western journal of emer-
gency medicine, (2):191–194.
ENASE 2023 - 18th International Conference on Evaluation of Novel Approaches to Software Engineering
612