Mobile Health App for Biofeedback Response in Physiotherapy

Development and Validation

Gonc¸alo Telo

1,3

and Hugo Gamboa

1,2

Faculty of Sciences and Technologies, NOVA University of Lisbon, Lisbon, Portugal

LIBPhys-UNL, Lisbon, Portugal

PLUX - Wireless Biosignals, S.A., Lisbon, Portugal

Keywords:

mHealth, mHealth App, Biofeedback, Mobile, Physiotherapy.

Abstract:

This work consists in developing an electromyographic biofeedback system in the form of a user-friendly

mobile application which is simple to use, as it was designed to be an auxiliary component in physiother-

apy sessions. This was achieved by implementing a framework that allows the integration of multi-platform

plugins, as well as a web view based user interface, which assures the best of the designs allied to the speciﬁ-

cations of the native APIs. The communication between the native and the JavaScript methods was tested, as

the validation of the application was made internally.

1 INTRODUCTION

Bioelectricity is the branch of Bioelectromagnetism

that studies electrical phenomena in biological tis-

sues, including areas like Electromyography (EMG)

that allow the study of the functioning of muscles and

motorneurons(Malmivuo and Plonsey, 1995).

Biofeedback is the information transfer that char-

acterizes a certain state of a biological process, there-

fore allowing the training of several physiological ac-

tivities. For this purpose, the level of activity of the

organ in study is provided by an electronic instrument,

that acquires a signal that is presented to the user in

the form of some kind of feedback (visual or sound).

This technique can, for instance, be applied to phys-

iotherapy, being able to provide a number of several

exercises and a better evaluation of the progression

and training of a patient (Bray, 1998).

Surface Electromyography (sEMG) is a non-

invasive technique that measures the electric signals

from the muscle contraction on the skin, in a phe-

nomenon that is controlled by the nervous system.

This signal is composed of the contribution of each

motor unit (MU) that activates in order to apply a

force and therefore the amplitude of the signal is pro-

portional to the quantity of MUs that participate in it

(Kamen and Gabriel, 2009; Chowdhury et al., 2013).

Electronic Health (eHealth) incorporates all the

tools and services that use information and commu-

nication technologies (ICTs) with the purpose of as-

sisting the healthcare environment either in diagnosis,

treatment or monitoring activities (European Comis-

sion, 2012). In ICT, parallel development trends

such as wireless technologies and ambient intelli-

gence have become some of the most important fac-

tors for the evolution of eHealth (Saranummi, 2011).

In 2012, a World Bank report established that

75% of the world’s population has access to a mo-

bile phone (Tomlinson et al., 2013). Mobile health

(mHealth) covers medical and public health practice

supported by mobile phones, patient monitoring de-

vices, personal digital assistants, and other wireless

devices (European Comission, 2014).

The creation of an mHealth app that can provide

the best of the latest trends both in the health context

and in the ICTs is therefore an important step towards

the enhancement of medical devices.

This mobile application was developed in a web

view basis, which will conjugate the android native

(Java) with the web developing languages (JavaScript,

HTML, CSS). That option will allow freedom of cre-

ation of the Graphical User Interface (GUI) and make

sure that it has all the characteristics that we aim to

achieve. This means that this is a freestyle mobile ap-

plication that allows its user to guide his session of

physiotherapy independently, with the purpose of en-

hancing the app’s responsiveness and easy adaptation

to the several exercises that can be performed.

502

Telo, G. and Gamboa, H.

Mobile Health App for Biofeedback Response in Physiotherapy - Development and Validation.

DOI: 10.5220/0005825405020506

In Proceedings of the 9th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2016) - Volume 5: HEALTHINF, pages 502-506

ISBN: 978-989-758-170-0

2 MOBILE APPLICATION

The developed software (physioplux lite) is an elec-

tromyographic (EMG) biofeedback system designed

for physiotherapists to use both in monitoring and

treatment contexts. By providing real-time muscular

feedback, it is intended for analysis of the patient’s

muscular activity which is monitored by one or two

medical devices. The latter correspond to muscle-

BAN devices, wireless electromyographic apparatus,

designed and assembled at PLUX - Wireless Biosig-

nals, S.A.. It is a component of a more complex sys-

tem, but it is also a medical device by itself.

In order to meet the previous description, the app

had to fulﬁl a number of criteria:

1. Provide real-time biofeedback;

2. Help physiotherapists to deﬁne goals;

3. Report and track exercise progress, including the

primary objective measurements;

In addition, it was also necessary to establish us-

ability and performance requirements, concerning the

smartphone’s Hardware and Operative System (OS).

The list below provides some examples of the selected

characteristics:

1. The physioplux lite should be used in Android

OS tablets or smartphones with minimum require-

ments: OS version above 4.3, 1GB RAM, Dual-

core 1GHz processor, and integrated bluetooth

with touch screen

2. The physioplux lite should be used in its core

functions with or without internet connectivity.

Overall, the application can be divided in three

main sections: settings, display and report.

In Settings the user selects the device whose feed-

back he aims to receive. The mobile device scans the

local area to ﬁnd all available muscleBAN devices,

providing the MAC Addresses the user can choose

from. Depending on the number of devices found,

the app automatically deﬁnes the use of one or two

devices, which reﬂects on the display. It is also possi-

ble for the user to deﬁne the name that is shown in the

display identifying each device. This should represent

the muscle the device is applied to. All the other ac-

quisition parameters are static, in order to assure the

quality of the treatment.

The main page is where the data from the devices

is displayed. Here the user must deﬁne objectives in

two steps: ﬁrstly, a percentage value is required. This

will be used to calculate a threshold, as a result of

the selected percentage of the reference value. Then,

the user must deﬁne if an objective is achieved either

by the measured value being higher or lower than this

threshold. By default, the reference value is 1 V, how-

ever it is possible to establish this by performing a

calibration. This is done by measuring the patient’s

Maximum Voluntary Contraction (MVC), which al-

lows the ﬁxing of the maximum value for each device

at the MVC value. As the objective of each bar is met,

a goal is achieved and the timer starts counting. This

stops when one of the objectives is no longer met. If

the new value registered by the timer is higher than

the previous, the max value is updated, otherwise it

only counts as a goal achieved. Each session allows

the execution of as many exercises as the user wants,

and a new exercise begins when the user clicks on the

exercise display.

The Report is mostly ﬁlled automatically. In the

report page the user can only deﬁne the user’s id and

his condition, as well as adding descriptions and ad-

ditional notes to each exercise. The session’s starting

time and the elapsed time are both displayed, along

with the MVC value of each device. Besides that, all

of the exercises performed are registered, just as the

corresponding thresholds, the number of goals that

were met and the maximum time recorded. When the

session is ﬁnished the user can export the report, as a

PDF ﬁle, to his email account or to a cloud.

The mobile application can be described with re-

spect to two different areas: the multi-platform plugin

and the application itself. In this ﬁrst phase of devel-

opment the platform of choice was the Android OS,

because of its major market quote. However the app

is ready to receive APIs for other operative systems.

2.1 Multi-platform Plugin Design

Apache Cordova is an open source framework with a

suite of APIs that allow developing mobile apps us-

ing JavaScript with access to native device functions.

This tool enables the creation of a web view app that

is developed just with HTML, CSS and JavaScript

in a user interface (UI) framework environment like

Intel

XDK. Despite the web technologies that are

used, the app is hosted locally (Apache Cordova,

2015).

This software has the enormous advantage of

being consistent across multiple device platforms,

which permits the migration to other device plat-

forms without considerable changes. This system

is available in operative systems such as Android,

iOS, Blackberry or Windows Phone (Apache Cor-

dova, 2015).

Developers can also create their own plugin or

access the database and use the third-party plugins

that the developers community shares, in his own app

(Apache Cordova, 2015).

Mobile Health App for Biofeedback Response in Physiotherapy - Development and Validation

503

Figure 1: Apache Cordova Native Application Architec-

ture(adapted from (Wargo, 2013)).

2.1.1 Native API

The Native Application Programming Interface

(API), is the ﬁrst level above the OS system layer.

This interface enables the communications between

our app and the devices.

The communication between the devices and the

smartphone/tablet is established using Bluetooth Low

Energy. This service results in a lower energy con-

sumption when compared to the classic bluetooth,

also guaranteeing a faster form of communication us-

ing notiﬁcations or indications that ﬁre a callback

when the device sends data.

The basic native OS includes methods like con-

nect, disconnect and close. However, in order to

communicate with the device we have complemen-

tary methods like start, stop and description. Several

callbacks are constructed in order to guarantee a good

level of communication between the devices, they are

represented by data frames, events and command re-

sponses.

Data frames are the vehicle used to pass the EMG

samples from the device to the mobile application.

They consist in a sequence number and an EMG sam-

ple value.

Some examples of the events received are: is con-

nected, is disconnected, the battery stage and events

concerning the device’s correct or incorrect place-

ment. Just as the occurring errors with the command

that the API sent to the device represent the responses

that can be received.

2.1.2 Plugin Architecture

The plugin architecture is based on the communi-

cation between the device’s API and the native in-

terface of the Cordova framework, that is called by

the JavaScript object that communicates with the top

layer of the app.

The native interface is the main core of the plugin.

It is where the plugin activity is placed and where the

callback to the app interface is set.

There are two main source codes used in the plu-

gin structure. The ﬁrst is on the JavaScript side and an

example of how to call the plugin’s native interface is

deﬁned as follows:

MuscleBan.prototype.connect = function(

onSuccess, onError, address)

{

exec( onSuccess,

onError,

"MuscleBan",

"connect",

[address]);

}

The next method is what enables the interaction

with the device’s API by receiving inputs and return-

ing callbacks. The latter reﬂect either a success or an

error case, which are treated by a layer at a higher

level.

private void connect(JSONArray args,

CallbackContext callbackCtx) {

try {

String macAddress = args.getString(0);

if(mDeviceService.connect(macAddress)){

callbackCtx.success();

} else {

callbackCtx.error(

"Error - on try to connect D1");

}

} catch (Exception e) {

Log.e(LOG_TAG, "Error on connect: " +

e.getMessage());

}

In the matter of data frames or events, the callback

channel is open and remains in that state so that the

data ﬂow is enhanced.

2.2 Application

This web view based application was developed using

HTML5, JavaScript and CSS, as well as Apache Cor-

dova plugins like the one described in 2.1. Addition-

ally, we make use of other libraries, such as jQuery

mobile, Font-Awesome and jsPDF in order to fulﬁl

the requirements that were planned.

HEALTHINF 2016 - 9th International Conference on Health Informatics

504

2.2.1 Application User Interface

The UI was designed to provide a user-friendly expe-

rience and to be very responsive (Figure 2).

Figure 2: Application’s Display Page.

It consists in an HTML5 canvas based interface

where all the elements are responsive. The timers

(now and max) turn grey on click, with the purpose

of not interfering with the session training.

The user can also choose to receive audio notiﬁ-

cations in two different modalities: if the objectives

are being respected or if the objectives are not being

respected.

3 SYSTEM EVALUATION

3.1 Communication and Visualization

Tests

In order to evaluate the readiness of the bluetooth

communication with the selected framework, we

made a validation test to quantify the delay associated

with the biosignal acquisition and its communication

through the mobile phone to the user’s interface.

For this test we made a simple app, similar to the

described in 2, but in this case with two different de-

vices: a virtual one with a square wave behaviour (be-

tween zero and max) and a device similar to the ones

used to acquire the EMG but with a photodiode. The

main goal was for the photodiode to detect the move-

ment of the virtual device bar, counting the time until

the response of the stimuli arrived.

The results were very satisfying, with a delay

ranged between 100 and 150 ms.

3.2 Validation Tests

After the development stage being complete, the app

was tested internally by a therapist in a simulated en-

vironment. These tests where made using a vertical

approach, to assure all the methods and events correct

functioning.

Before validating the system in a clinical envi-

ronment, it was necessary to perform a risk analysis,

which is part of the software’s technical ﬁle. This

analysis includes explanations of the several issues

that could lead to the misuse of the mobile applica-

tion and whose severity was classiﬁed according to

their probability and gravity. This was followed by

the mitigation of such problems, as a way of guaran-

teeing a good quality performance and experience.

The application was also tested by the company’s

physiotherapist and other engineers, several scenarios

and issues where detected and some improvements

where made. The most important improvement was

the adjustment of the sub-sampling frequency, in or-

der to achieve a good clinical result with a good visual

feedback. Other issues concerned the area of the in-

formation given to the user, including the critical bat-

tery level warning and the misplacement of the device

on the skin.

The veracity of the collected data was also taken

into consideration during the evaluation, in order to

validate the method. This was accomplished by com-

bining the sensibility of the physiotherapist and direct

comparisons with other PLUX’s products results in

the physiotherapy area.

4 CONCLUSIONS

The main purpose of this work was to make a ro-

bust and steady software, with a user-friendly inter-

face that should allow an easy interaction between the

therapist and the patient.

Aiming for this outcome, the whole app was de-

veloped from scratch by creating the plugin and the

application section, as well as the web view based in-

terface.

The plugin is speciﬁc for this app, but it was writ-

ten so that it can easily be used with other purposes

and apps, giving much information to the developer.

Mobile Health App for Biofeedback Response in Physiotherapy - Development and Validation

505

The communication between the Java code (native

API) and the JavaScript (app code) was found to be

responsive and highly reliable.

When the UI design was established it took into

account all the good practices, going from the selec-

tion of the icons to the colors and the type of events

that are displayed.

Many of the decisions that inﬂuenced the design

of the app, were taken considering the feedback re-

ceived over time, as a way of guaranteeing the best

user experience possible.

As a continuation of this work, we will proceed

with the validation tests, concerning the market entry

strategy.

The next step to take will be to implement the

APIs from other operative systems such as iOS, in or-

der to guarantee the multi-platform compatibility of

the app, along with ﬁlling the requirements of the two

biggest OS in the mobile market.

REFERENCES

Apache Cordova (2015). Apache Cordova.

Bray, D. (1998). Biofeedback. Complementary Therapies

in Nursing and Midwifery, 4(1):22 – 24.

Chowdhury, R., Reaz, M., Ali, M., Bakar, A., Chellappan,

K., and Chang, T. (2013). Surface Electromyography

Signal Processing and Classiﬁcation Techniques. Sen-

sors, 13(9):12431–12466.

European Comission (2012). eHealth Action Plan 2012-

2020 - Innovative healthcare for the 21st century.

European Comission (2014). Green Paper on mobile health.

Kamen, G. and Gabriel, D. A. (2009). Essentials of Elec-

tromyography. Human Kinetics.

Malmivuo, J. and Plonsey, R. (1995). Bioelectromagnetism:

Principles and Applications of Bioelectric and Bio-

magnetic Fields. Oxford University Press, New York.

Saranummi, N. (2011). In the Spotlight: Health Infor-

mation Systems— \vskip-48pt Mainstream-

ing mHealth\hﬁll. IEEE Reviews in Biomedical Engi-

neering, 4:17–19.

Tomlinson, M., Rotheram-Borus, M. J., Swartz, L., and

Tsai, A. C. (2013). Scaling Up mHealth: Where Is

the Evidence? PLoS Medicine, 10(2):e1001382.

Wargo, J. M. (2013). Apache Cordova 3 programming.

HEALTHINF 2016 - 9th International Conference on Health Informatics

506