Provision of Personalized Data via Mobile Web Services in eHealth

Scenarios

Marc Jansen

1,2

, Abbas Siddiqui

and Oliver Koch

Computer Science Institute, University of Applied Sciences Ruhr West, Bottrop, Germany

Department of Media Technology, Linnaeus University, Växjö, Sweden

Keywords: Personalized Data, Context-awareness, Service Provision, Service Consumption, Middle-ware, eHealth,

mHealth.

Abstract: In societies where the demographic change leads to a more and more unbalanced state between the elderly

and all other different age groups, there, health management becomes one of the most significant problems,

e.g., in order to allow what is often called successful aging. Ubiquitous use of smart mobile devices enables

gathering of personalized data enriched with contextual information. This information can ideally be used in

modern eHealth scenarios, resulting in mobile health (mHealth) scenarios. This paper describes how highly

personalized and sensitive information, such as vital signs enriched with contextual information of a patient,

can be stored at mobile devices and provided via modern web technologies for later-on analysis and health

monitoring. Therefore, this paper presents a number of different scenarios in which such mobile technology

provides certain benefits, discusses benefits, drawbacks and challenges of such an approach and describes

an example implementation of an mHealth scenario.

1 INTRODUCTION

Nowadays, smartphones and other mobile devices

are used tremendously and this trend is still growing

rapidly. Here, the functions of smartphones in use

are much more than simple voice calling and text

messaging. The mobile users use mobile phones to

access their mail accounts, to share their calendars or

to even use their devices as a routing device (i.e., the

ones equipped with the GPS). Therefore, smart

mobile devices collect various information about

their users such as personalized information,

personal/business contacts, messages (emails, short

text messages like SMS, …), social interaction and

social contacts, calling history, work history and

footprints of internet usage (e.g., the browsing

history of a mobile web browser). Additionally, the

set of applications, a user has installed on his device,

provide a rough profile of usage of the device itself.

Two major advantages from the increased usage

of mobile devices are: On the one hand, the devices

are mobile and could therefore ideally be used to

follow their users round the clock and ubiquitously.

And on the other hand, a modern internet connection

is provided with most of the smart mobile devices,

this makes such kind of mobile devices a powerful

and valuable tool. The second reason for the success

of smart mobile devices is, that these devices are

usually equipped with a large set of different

sensors, allowing to contextualize the users’ current

tasks, e.g., by using the GPS sensor of a modern

mobile device, the current position of a user could

also be determined accurately, so that additional

information about the task could be gathered, either

to provide better support to users during the task or

to store additional contextualization data for later

analysis. In addition to GPS sensors, other sensors

like acceleration sensor, motion sensors, a digital

compass and alike, could be used on modern mobile

devices. In the context of eHealth scenarios,

additional sensors for monitoring current vital signs

of the user could be connected to a mobile device.

The idea presented in this paper is twofold, on

the one hand an architecture is presented that allows

to store the data on ones’ own device and not in

some central database. Here we argue that this is a

big advantage (in contrast to modern Cloud

Computing and Web 2.0 scenarios in which more

and more data is stored centrally), having in mind

that health data is always personalized data with a

big demand for security. The second idea is provided

in this paper is to combine vital data (gathered by

441

Jansen M., Siddiqui A. and Koch O..

Provision of Personalized Data via Mobile Web Services in eHealth Scenarios.

DOI: 10.5220/0004949704410446

In Proceedings of the 10th International Conference on Web Information Systems and Technologies (WEBIST-2014), pages 441-446

ISBN: 978-989-758-023-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

specialized eHealth sensors connected to a standard

mobile device) and contextualization data (gathered

by standard sensors of a mobile device) and to

provide it in an authentic and secure way, which

without the usage of mobile devices could hardly be

achieved. Additionally, two scenarios are presented

in detail to demonstrate the use of that kind of

technology. For one of the scenarios, an already

existing implementation is described. Here, both on

the implemented middleware and on the client side,

web based technologies have been used in order to

provide a user friendly implementation.

The rest of this paper is organized as follows:

first we provide an overview of related work in the

next section. After this we describe our approach in

section three, followed by a discussion of the

presented approach in section four. Section five

describes two scenarios in which the presented

approach provides benefits and section six describes

the web based implementation of one of these

scenarios. Finally, section seven provides a

conclusion and an outlook to future work.

2 RELATED WORK

Providing Web Services on mobile devices was

probably presented first by IBM (McFadden, et al.,

2003). In this work a solution is presented for a

specific scenario in which Web Services are hosted

on mobile devices. A more general approach for

providing Web Services on mobile devices is

presented in (Srirama, et al., 2006 and AlShawan, et

al., 2010). In (Li and Chou, 2011) the authors

suggest a different approach, focusing on the

optimization of the HTTP protocol for mobile Web

Services provisioning. Importantly, none of the

mentioned approaches manages to overcome certain

limitations of mobile devices, e.g., permanently

changing networks, IP addresses from networks with

network address translation (NAT) or the fact that

mobile devices are usually not designed to be always

online (might be switched off, might have not

network connection, …).

An additional approach that manages to

overcome these problems, is presented in (Jansen,

2013a). This approach utilizes a central proxy

infrastructure, that allows on the one hand to tackle

the mentioned challenges and on the other hand

provides a stable infrastructure for mobile device to

provide standardized Web Services. Furthermore,

the central proxy provides the major part of the

technical infrastructure that allows to use

standardized protocols for the deployed Web Service

at the mobile devices.

Additionally, the work presented in (Jansen,

2013b) argues for a new perspective to Web

Services especially if those services are deployed to

mobile devices. Here, the author argues that one of

the major benefits of mobile devices is the fact that

they are mobile. In connection especially to mHealth

scenarios, this fact allows to gather different kinds of

data from the sensors of a mobile devices that allows

to easily contextualize the user (in most mHealth

scenarios, the patient) in his/her current

task/situation.

3 APPROACH AND DISCUSSION

The approach we used for the implementation of the

example scenarios is based on a standardized Web

Services. As already explained, a number of

different approaches exist that allow the provision of

Web Services on mobile devices. Since from our

point of view, the approach described in (Jansen,

2013a), solves the most pressuring problems, we

decided to base our implementation on this

approach.

In addition to this, we implemented a

middleware layer allowing a loose coupling between

the provider of the mobile Web Services and the

consumer service in our case a web-based

application that visualizes the data measured by

different sensors and in different training sessions.

Another advantage of the approach described in

(Jansen, 2013a) is that it allows to store requests at

the central proxy, e.g., if the device that provides the

mobile Web Service is not connected to a network.

As soon as the device is connected again, the request

could be performed by the device and the result is

available for the client.

Additionally, this approach implements Web

Services for mobile devices based on standardized

protocols, so that developers who are familiar in

using Web Services do not need to learn a whole

new technology, but can rely on knowledge they

already have. Furthermore, also on the mobile Web

Service provider side, the approach utilizes

annotations pretty much in the same way as for the

implementation of Web Service on non-mobile

devices, providing again a very flat learning curve

for the new technology.

The chosen approach provides a number of

benefits, e.g., since the measured eHealth data only

remains on the mobile device of the patient, there is

no need for a central server infrastructure. This

results in lower management costs, easier to achieve

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

442

security and no additional costs for central backups.

Additionally, there is no need for central personnel,

space, hardware and software provisioning and no

central costs for energy.

On the other hand, this de-centralized approach

also provides some drawbacks. Since the data is not

stored centrally, hence, no central backup strategy is

applicable, but the patient him-/herself is responsible

for his/her data. Also security management would be

easier to accomplish in an architecture with central

data storage. Additionally, new questions with

respect to the security of the data come into play

while the data remains on mobile devices: what

happens if the device gets lost or stolen? Therefore,

of course, the personal data should be stored

encrypted on the device of the patient. Still,

encryption would just solve the problem of privacy

(no one else but the owner of the mobile device

would be able to access the data), but not the

problem of availability of the data. Once the device

is no longer in reach of its owner, the data would no

longer be accessible for future analysis.

Besides the challenges that come from the

described drawbacks, an additional challenge comes

into play considering the target group for this kind of

mHealth applications. The usual target group for this

kind of applications consists of elderly. Since this

target group is usually not specifically well used to

handle mobile devices, a clear and precise

explanation of the necessary steps and the used

technology is more than ever necessary.

4 TWO EHEALTH SCENARIOS

Key strengths of mHealth system architectures based

on mobile Web Services encompass the following

aspects:

1. On-demand transmission of data to Health

professionals: the doctor receives only the

data that really interests him and especially at

the point in time he is interested in.

2. Data remains in the direct disposal area of the

patient: the need for informational self-

determination is thus optimally fulfilled.

The following two scenarios, describe the use of

Web Service based mHealth applications by

illustrating the two aforementioned strengths.

4.1 Scenario 1: Surgical Preparation

The central source of revenue for hospitals is the

surgical area. This is the actual place where the

decision is taken whether a hospital makes profit or

deficit. Therefore, usually all hospital processes are

aligned to the optimal utilization of the surgical area.

Any delay in the process or even surgeries that need

to be canceled or postponed at short notice, cause a

more or less considerable economic loss.

Prior to surgery, patients usually must follow

certain rules of conduct and take or discontinue

certain medicine for several days or weeks. Has

he/she not done so (e.g., medication for blood

clotting was not discontinued), the sugery has to be

canceled at a short notice. This may lead to an

underutilization of the surgical area and

consequently the associated economic losses.

A 78-year-old patient is to receive a new right

hip joint. She suffers from a lighter hypertension and

type II diabetes. At the same time she tends to water

retention, which manifests in an abrupt increase in

her weight. Normally hip replacement operations

rank among "standard" operations with less risk

involved. But, because of the patients’ secondary

diagnoses, however, there is a significant surgical

risk. Prior to the operation her risk factors must

therefore be closely monitored. To shorten the

period of rehabilitation after the surgery, the patient

should move as much as possible in advance to the

surgery and complete a specific exercise program to

strengthen certain muscle groups.

For several months already, the patient has a

smartphone. At her last doctor's appointment, four

weeks before the surgery, a mobile mHealth

application for collecting and managing various

sensor data (in this case, blood pressure meter, blood

glucose meter and scale connected via bluetooth), is

installed and configured on her mobile phone. The

data is provided via a mobile Web Service, which is

also installed on the device. At the same time the

female patient receives an introduction into a sensor

package that will be handed out to her. Part of the

mHealth application are also information films, in

which the preliminary physical exercises are

provided to her and she gets also informed about the

risks and the progress of the medical intervention.

In the following four weeks, the female patient

regularly measures her blood pressure, blood sugar

and weight. This data is automatically enriched with

location, time, and motion information from her

mobile phones sensors (GPS, accelerometer,

proximity sensor, etc.), allowing to contextualize the

data gathered by the medical sensors.

Thus, noticeable measured values (e.g. increased

heart rate and blood pressure due to sports activities)

can be interpreted, according to the context in which

they are occurred.

ProvisionofPersonalizedDataviaMobileWebServicesineHealthScenarios

443

The data is generated using the internal sensors of

smart phones and possibly even enriched by data

from external web services (e.g. historical weather

data). By using a feedback function, the patient may

give feedback, for example how she felt during the

exercise and the measurements. The logged

feedback and sensor data can be queried by the

hospital on-demand. When retrieving the data the

hospital doctors can decide if they want to get all

data or only aggregated data in terms of a status

report (surgery possible / not possible). In addition,

the mHealth application detects emergency

situations (e.g. very high blood sugar levels) and

alerts the hospital if necessary. The surgical

management can check at any time whether the

surgery will be possible, a change in behavior of the

patient is necessary or an adaptation of the surgical

plan is required.

Even at the stage of rehabilitation the Web

Service on the patient’s smartphone can provide on-

demand information about the patient compliance

(e.g., lack of physical exercise), allowing the timely

intervention of the treating physicians.

4.2 Scenario 2: Regular Data

Measurement by Patients

In the second scenario we imagine a 33 years old

male patient diagnosed of having a genetic disease

that disallows his body to effectively fight against

blood lipids. Therefore, he is likely to have an

increased risk of an attack. In order to lower this

risk, besides a special medication, the patient is

asked to have 24 hours blood pressure measurements

at least once a year. A couple of years ago, this was

more or less unhappy day every year, in which the

patient had to wear a heavy blood pressure monitor

over a whole day. Lately, his doctor told the patient

that with new upcoming technology in the mHealth

sector, the process for 24 hours blood pressure

monitoring could be eased a lot.

By connecting a small and light blood pressure

sensor to his smartphone, the patient is easily able to

collect the necessary data in a convenient way.

Furthermore, by collecting additional data from the

mobile device of the patient, e.g., the current

geolocation and/or the acceleration at which the

patient is currently moving, the gathered data could

be contextualized in order to ease the task for later

analysis.

Happy to know about these new possibilities, the

patient is willing to take his next 24 hours blood

pressure measurements with this new technology.

Two weeks after the measurement was done, the

patient has a new appointment with his doctor in

which the data from the measurement should be

discussed. During this discussion, the doctor

recognizes a tremendous increase in the blood

pressure at 17:03 at the day of the measurement.

First, neither the patient nor the doctor could explain

this increase, but by looking at the contextual data

gathered along the blood pressure values, the reason

for the increase of the blood pressure became clear.

By analyzing the geoposition of the patient and his

acceleration at the time of the increasing blood

pressure, it turned out that the patient was watching

his famous soccer team playing against another local

team. Knowing this, the patient remembers that at

this exact time, the opponents of his famous team

managed to shoot a goal. Being upset about this

event clearly explains the increase in the patients’

blood pressure values.

Again here in this scenario, beside the two major

advantages of on the one hand privacy of the data

(the data remains with the patient and is not stored

centrally) and the just in time provision of the

interesting data to the right person (in this case the

doctor) could easily be seen. Furthermore,

advantages of using modern mobile devices in

mHealth scenarios, like the rich source for the

contextualization of the patient are also obvious.

5 IMPLEMANTATION OF AN

EXAMPLE SCENARIO

The prototype consists of two parts namely Web

Services (i.e., running on a mobile device) and a

web based client (i.e., running in a browser).

The following paragraphs describe the detail

about these aforementioned components of the

prototype followed by the interaction among the

components.

5.1 Mobile Application

Android based mobiles are selected for the

implementation of the concept as according to the

International Data Corporation (IDC) Worldwide

Quarterly Mobile Phone Tracker’s 2013 report,

android based mobiles are the market leaders. An

Android based application is developed to collect the

various sensors data.

There are two categories of sensors, the ones

integrated in a mobile phone and others which are

externally linked via well-known communication

technologies (e.g., Bluetooth, Wi-Fi).

In this particular implementation, the location of

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

444

a patient is determined by an integrated global

positioning system (GPS), and the speed of

movement is measured by a built-in accelerometer,

which measures the force of acceleration caused by

the gravity or weight to determine the speed. The

standard Android library is used to retrieve the

sensors data.

Besides the integrated sensors, to measure the

heart-rate and, to measure the instant speed of a

patient, a Zephyr sensor (Zephyr HxM BT – Heart

Rate Monitor) is used. The sensor is mounted to a

smart fabric based strap, which must be worn around

the chest. The data is periodically collected via

Bluetooth. The HxM sensor provides a Bluetooth

API to collect the data in a specific message format.

After the collection of the data from all those

aforementioned sensors, the data is saved by the

application shown in Figure 1 to a SQLite database.

Figure 1: A screenshot of the app collecting data from an

mHealth session.

A SQLite database runs without a server, which

makes it a suitable choice for the energy limited

mobile devices. Nevertheless, SQLite data is very

application related and dependent, once an

application is uninstalled then the data will no longer

be available. The database contains two tables

“sessions” and “sensorsData”.

Besides retrieval and storage of the sensors data,

another part of the application handles the incoming

request as a mobile Web Service. It fulfills the

authentic incoming request by sending the stored

data on the device to the client.

5.2 Web Client

The client runs in a browser and can be invoked

from everywhere. Most of the functionality is

implemented in JavaScript and HTML5.

Nevertheless, due to the same origin restriction,

some façade, according to the façade/proxy design

pattern (Gamma, et al., 1995) services needed to be

implemented on the server side. For this

implementation we have chosen PHP.

After a successful authentication (e.g., as a

doctor), the web application sends a Web Service

request to the mobile device of the patient in order to

receive a list of all available sessions. By choosing a

certain session, another Web Service request is send

to the mobile device that receives the data gathered

during this session. This data could then be

visualized in a tabular format or as a representation

in a Google Map, where, e.g., the data is located at

the correct geoposition and different parts of the data

could be visualized according to the geoposition at

which they appeared. An example of the visual

representation on the client side in a Google Map

could be seen in Figure 2.

Figure 2: Visual representation of gathered health data in a

Google Map.

5.3 How All Work Together

The functionality of the prototype is described from

ProvisionofPersonalizedDataviaMobileWebServicesineHealthScenarios

445

the two different perspectives namely a user (e.g., a

patient), and a user of the collected data such as a

general physician, a family member or patient

himself.

From the mobile-health-moniters’ user (e.g., a

patient) perspective, a user starts the application

which collects the sensors information on his/her

mobile device. Before starting a new session to

record the sensors data, a user must wear a HxM

chest strap with mounted heart-beat sensor to collect

the real heart-beat of a user, otherwise a default

value (i.e., which is “0” in this case) will be stored.

Once the session is started, every few seconds the

new values from the location sensor, 3-axis

accelerometer and HxM are retrieved and stored in

the “sensorsData table in the “mhealth” database.

Let’s assume, a user started a session during the

workout, once the workout is over, the user decides

to discontinue the session and clicks the disconnect

button, once the session is disconnected a new entry

in the “sessions” table will be stored with start-time

and end-time of the session.

When a concerned person (e.g., a personal

doctor, a family member, a care take, or the

user/patient him-/herself), let’s assume, a house-

doctor wants to know about the patient who is using

the mobile-health-monitoring app. The house-

doctor will enter the web-link in a browser to invoke

the mobile Web Service running on the mobile

device. Before accessing the patients’ data, the

house-doctor must identify him-/herself with

credentials through a username and a password.

After the house-doctor is identified, all the stored

sessions on the patients’ mobile device are retrieved

via the mobile Web Serivce. Now, the house-doctor

can select any session to analyse the measured data.

6 CONCLUSIONS AND FUTURE

WORK

Wrapping up, this paper provides a discussion about

the usage of standardized Web Services on mobile

devices like smartphones in the context of mHealth

scenarios. We provided some example scenarios, a

discussion about the current state of the art and also

about benefits and drawbacks of the presented

approach. Last but not least, a web-based

implementation of one of the example scenarios was

presented.

Nevertheless, beside the obvious benefits of such

an approach, there are also some topics that need to

be considered much more intensively in future

research.

One of the major things to have in mind, is the

security of the data on the phone of a patient. More

likely, data security (especially data privacy) could

be much easier realized for a central database (i.e.,

administrated by IT specialists). On the other hand,

data security is anyway a topic that users of smart

devices have to deal with and there are already quite

powerful technologies available that increase the

level of security for mobile data. Here, an important

task for future research efforts will be to integrate

such technologies in the described scenarios.

Another aspect comes into play if we consider

that the different kinds of data possibly gathered by

mobile devices need to be arranged and visualized in

a way that allows doctors to easily interpret this

data. Since doctors are usually not specialists in data

analysis, additional efforts need to be considered for

the presentation and analysis of the data.

Last but not least, the reliability of the medical

data gathered by not necessarily medical devices is

an issue that needs to be tackled.

REFERENCES

F. AlShahwan, K. Moessner, “Providing SOAP Web

Services and REST Web Services from Mobile

Hosts”, In: Fifth International Conference on Internet

and Web Applications and Services (ICIW 2010), pp.

174-179.

E. Gamma, R. Helm, R. Johnson, J. Vlissides, Design

Pattern – Elements of Reusable Object-Oriented

Software, Addison-Wesley. 1995.

M. Jansen. Analysis and Improvement of Energy

Consumption for Providing Mobile Web Services.

In.International Journal of Soft Computing and

Software Engineering, DOI: 10.7321/jscse. 2013a.

M. Jansen. About the Necessity to Change the Perspective

for Mobile Web Services. In: Proceedings of the 15th

IEEE International Symposium on Web Systems

Evolution, 2013b.

L. Li, W. Chou, “COFOCUS – Compact and Expanded

Restful Services for Mobile Environments”, In:

Proceedings of the 7

International Conference on

Web Information Systems and Technologies, pp. 51-

60, Noordwijkerhout, The Netherlands, 2011.

S. McFaddin, C. Narayanaswami, M. Raghunath, “Web

Services on Mobile Devices – Implementation and

Experience”, In Proceedings of the 5

IEEE Workshop

on Mobile Computing Systems & Applications

(WMCSA’03), pp. 100-109, Monterey, CA.

S. Srirama, M. Jarke, W. Prinz, “Mobile Web Service

Provisioning”, In: Proceedings of the Advanced

International Conference on Telecommunications and

International Conference on Internet and Web

Applications and Services (AICT/ICIW 2006), p. 120,

Guadeloupe, French Caribbean.

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

446