Development and Evaluation of a Sensor-based System for Remote

Monitoring and Treatment of Chronic Diseases

The Continuous Care & Coaching Platform

Harm op den Akker

1,2

, Monique Tabak

1,2

, Mihai Marin-Perianu

, Rianne Huis in’t Veld

Valerie M. Jones

, Dennis Hofs

, Thijs M. T

onis

1,2

, Boris W. van Schooten

Miriam M. R. Vollenbroek-Hutten

1,2

and Hermie J. Hermens

1,2

Roessingh Research and Development, Cluster Telemedicine, Enschede, The Netherlands

University of Twente, Telemedicine Group, Enschede, The Netherlands

Inertia Technology, Enschede, The Netherlands

Keywords:

Telemedicine, User-centred Design, Physical Activity, Lifestyle Monitoring.

Abstract:

The steep increase in the number of people with chronic diseases in the Western world is a long recognised

problem which forces us to look for less labour intensive and more cost effective methods to deliver health-

care. Wider use of telemedicine services seems to be a solution that can contribute substantially to reduction

of costs and maintenance of health service quality at a high level, but its implementation in regular healthcare

practice often fails. This is no surprise since designing, building and implementing a successful telemedicine

application is no trivial task. The reasons for failure are many, and include technical problems and low us-

ability, however there are also organisational factors such as the often disruptive effect the introduction of new

technologies can have on existing healthcare processes. This paper presents the experience gathered when

building our system: the Continuous Care & Coaching Platform (C3PO). This platform was designed to mon-

itor patients with chronic diseases during daily life and provide them with the support they need to develop

and maintain an active lifestyle in order to improve quality of life and offset deterioration of health condition.

In this paper we share the lessons learned and present results of evaluation studies using earlier versions of the

platform. Innovations in the ﬁeld of wireless sensor technology as well as changing demands from the users

led us to redesign and develop a new version of the platform. This process is described, starting with the gath-

ering of requirements using a user-centred design approach and derivation of speciﬁcations. The components

of the platform are described as well as results of recent evaluation studies. Our evaluations and experience

with working with patients and healthcare professionals indicate that there is currently a great need to add in-

telligence to the platform. The augmentation of the platform with of a smart autonomous health- and lifestyle

coach is the current focus of our research which, we believe, will be a crucial next step towards the future of

intelligent telemedicine applications.

1 INTRODUCTION

The world’s population is rapidly ageing, as a result

of both longer life expectancy and declining fertil-

ity rates. Worldwide, the proportion of people aged

over 60 is growing faster than any other age group,

and by 2050 there will be some 2 billion elderly per-

sons. A person’s health typically deteriorates with

increasing age and chronic diseases emerge, induc-

ing greater demand for long-term care. According

to the United Nations report “World Population Age-

ing 150-2050” (United Nations, Department of Eco-

nomic and Social Affairs, Population Division, 2002),

the mortality, morbidity and disability attributed to

the major chronic diseases — heart disease, stroke,

cancer, chronic respiratory diseases and diabetes —

is expected to rise to 73% of all deaths and 60% of

the global burden of disease by 2020 . Finding in-

novative approaches to making our healthcare system

affordable and sustainable is needed.

Regular physical activity is beneﬁcial for healthy

aging as it substantially delays the deterioration of

health status and helps prevent the development of

secondary chronic diseases. A focus on (self-) man-

Tabak M., Tönis T., van Schooten B., Marin-Perianu M., Jones V., Huis in ’t Veld R., Vollenbroek M., Hofs D., den Akker H. and Hermens H. (2012).

Development and Evaluation of a Sensor-based System for Remote Monitoring and Treatment of Chronic DiseasesThe Continuous Care & Coaching

Platform.

In Proceedings of the Sixth International Symposium on e-Health Services and Technologies and the Third International Conference on Green IT

Solutions, pages 19-27

DOI: 10.5220/0004474700190027

 SciTePress

agement of physical condition will become increas-

ingly important in future chronic care and in the

support of the healthy elderly population. Informa-

tion and Communication Technologies play a cru-

cial role in supporting greater independence and self-

management of lifestyle and disorders. We need to

identify patient management approaches that will en-

sure appropriate monitoring and treatment remotely.

To deal with this problem, Hermens and Vollenbroek-

Hutten (Hermens and Vollenbroek-Hutten, 2008) pro-

pose to develop remote monitoring and treatment sys-

tems. These should integrate ambulant sensing to

measure relevant bio signals and context information

with secure data handling and appropriate clinical de-

cision support functionality to assist in both technical

and clinical decision making. In addition, the sys-

tem should also provide feedback to both patients and

care providers. In other words, such systems should

provide continuous monitoring of health status, with

the promise of coaching, or continuous motivational

aid aimed at achieving behavioural change, whenever

required, and efﬁcient and effective individually tai-

lored treatment anywhere anytime.

These systems do promise to make healthcare

more quantitative, more efﬁcient and effective and

less costly, which is desirable for patients, healthcare

professionals and health insurance companies alike.

Although the number of available systems is rapidly

increasing, still around 75% of the telemedicine ini-

tiatives fail during the operational phase after a pilot

(Berg, 1999) and, despite numerous trials, very few

are transferred successfully into routine health care.

Inadequate technical performance as well as low us-

ability of these systems is considered to be among the

major barriers for successful implementation (Broens

et al., 2007).

Our aim was to develop a remote monitoring

and treatment platform for elderly and patients with

chronic disorders, to support them in developing

and maintaining an active lifestyle and improving

their physical condition, either independently or su-

pervised remotely by their healthcare professionals.

The platform, focussed on continuous monitoring and

feedback to the users is called the Continuous Care &

Coaching Platform or C3PO. In this paper, we de-

scribe the development of the platform and its com-

ponents: sensors, Smartphone, server and web por-

tal. Section 2 starts with a short overview of the cur-

rent state of the art of telemedicine platforms and re-

search. We then describe the process of our user-

centred design approach, starting with the lessons

learned from previously developed remote monitor-

ing systems, the development of functional and tech-

nical requirements, the platform design, and system’s

evaluation. Finally, we discuss the ongoing work in

improving the platform and discuss the future of re-

mote monitoring and treatment systems.

2 BACKGROUND

The ﬁeld of telemedicine research is large, and rapidly

expanding. A systematic review of reviews by Eke-

land et al. (Ekeland et al., 2010) found 1593 reviews

of telemedicine systems published since 1998. Hav-

ing included 80 reviews in their analysis, the authors

found only 21 reviews that concluded telemedicine

to be effective, and 18 that concluded that evidence

is “promising but incomplete”. Similar results can

be found when looking into telemedicine applications

targeting physical activity interventions (an important

focus of C3PO). A systematic review by LaPlante

and Peng (Laplante and Peng, 2011) analyzed 31 e-

Health intervention publications targetting physical

activity between 2006 and 2010. Four of the 7 stud-

ies that used pure control groups showed support for

e-Health, while the others showed no signiﬁcant dif-

ferences. The authors conclude that although it is dif-

ﬁcult to argue for the beneﬁts of e-Health compared to

classic interventions, none of the evaluated platforms

fared worse than its control setting. Thus it seems that

more research in the area of telemedicine, as well as

more standardized and thorough evaluations of its ef-

fects are necessary. This conclusion is also found in

a review conducted by Polisena et al. who studied

home telehealth systems for Chronic Obstructive Pul-

monary Disease (COPD), one of the important target

groups for the platform. The authors reviewed 9 stud-

ies, and concluded that although emergency depart-

ment visits and hospitalization rated declined, other

outcome parameters varied between studies (Polisena

et al., 2010).

Telemedicine platforms come in many shapes and

forms, some focussing more on vital sign monitoring

for e.g. elderly (Czabke et al., 2011) or cardiac pa-

tients (Kumar et al., 2008; Jones et al., 2009); oth-

ers focus more on the human-computer interaction

through television (Burkow et al., 2008), smartphones

(op den Akker et al., 2010; Wieringa et al., 2011;

ahl et al., 2008; Chen et al., 2009), web-based com-

munities (Lewis et al., 2008) or web portals (Her-

mens and Vollenbroek-Hutten, 2008). A full review

of all the related aspects of telemedicine platforms

would be out of the scope of this paper. It is how-

ever important to recognize their complexity in both

technical and societal sense, and we strongly believe

that careful attention to all aspects, as well as broad

user involvement throughout all the stages of devel-

EHST/ICGREEN 2012

opment can contribute to the potential success of any

telemedicine application.

3 FIRST LESSONS LEARNED

In 2006, development started on an ambulant activity

monitoring system. In this ﬁrst version of the plat-

form, a 3D motion sensor (XSens MT9) was used to-

gether with a data logger. The sensor contained three

separate uni-axial piezoelectric accelerometers with a

range of 2g, and was connected over a wire to the

data logger. Both devices were worn on a belt on the

waist. In one of our ﬁrst studies on daily physical

activity, we used the MT9 to measure the activity lev-

els of chronic low back pain (CLBP) patients in their

everyday lives and compared the results of the ex-

perimental group with those of healthy controls (Van

Weering et al., 2009). Participants wore the devices

for 7 consecutive days from 7:00 am until midnight

for a maximum of 17 hours per day. After the mea-

surement week, the memory card from the data log-

ger was read out and the data was analyzed. These

ﬁrst studies gave us valuable insights in the activity

patterns of CLBP patients, showing for example their

tendency to be much more active in the mornings and

much less active in the evenings compared to healthy

controls. But it had an obvious disadvantage in terms

of its size and was not comfortable to wear for long

periods of time. Therefore, we started the develop-

ment of a new version in 2008.

The second itertation of the platform consisted of

a wireless 3D accelerometer (XSens MTx-w) and an

HTC Smartphone. The sensor connects wirelessly via

Bluetooth to the Smartphone, which can be worn on a

belt clip or in a pocket. The use of a wireless connec-

tion between the sensor and the Smartphone meant

the system was easier to operate and more ﬂexible

in terms of its placement on the patient’s waist. The

use of a Smartphone also provided us with the op-

tion of designing real-time feedback on the stream-

ing input data and consequently allowed us to de-

sign fully ambulatory interventions. The Smartphone

could be used to show the patient a real-time graph

of his daily activity compared to reference values and

based on this provide timely feedback messages de-

signed to encourage the patient to change his phys-

ical activity behaviour. The platform was used in

a number of studies for monitoring everyday activ-

ity patterns in patients with different conditions in-

cluding chronic low back pain, chronic fatigue syn-

drome (CFS) and chronic obstructive pulmonary dis-

ease (COPD); it was also trialled on cancer survivors

and obese subjects. In each of these groups, the pa-

tient’s physical activity plays a role in the progres-

sion of his condition and general quality of life. In

COPD, for example, patients are symptomatic even

when performing normal daily life activities. This

causes the patient to avoid activity, leading to a de-

crease in physical ﬁtness which causes a rise in the

incidence of COPD exacerbations. In this case, im-

proving the physical activity behaviour can offset this

downward spiral that is so commonly seen in COPD.

The Smartphone application also prompts patients to

answer questions presented at ﬁxed time intervals dur-

ing the day on self-perceived activity performance, to

rate their level of pain (in case of CLBP), dyspnoea

(in case of COPD), fatigue (in case of CFS, COPD)

or to provide information on food intake (in case of

obesity) by means of visual analogue scales (VAS).

The feedback intervention studies showed that use of

the system for a period of several weeks resulted in

decreased symptoms and increased activity levels.

Although the second version of the platform was

an obvious improvement in terms of usability and

ﬂexibility, new issues emerged during the use of this

system in the various studies. The wireless Bluetooth

connection between the sensor and Smartphone was a

drain on the batteries of both devices. For the Smart-

phone we had to use extended battery packs, making

the devices more bulky than they needed to be, and for

the sensor, the drain on the batteries meant the device

would often run out of power after 12 hours of opera-

tion. Another issue was data loss. Because the sensor

sends out its raw sampled values continuously, each

time the Bluetooth connection is dropped, data is lost.

These limitations as well as further advancements in

the ﬁeld of wireless sensor technology and mobile de-

vices led to a further cycle of design and development

and a new version of the platform, described in the

next section of this article.

4 REQUIREMENTS AND DESIGN

In the following two sections we will explain the

process of requirements elicitation for the new plat-

form, called the Continuous Care & Coaching Plat-

form (C3PO) as well as the design of each of its indi-

vidual components.

4.1 Requirements

Early versions of the platform were mainly designed

for research purposes, but the experimental studies in-

dicated that the system had potential utility as a be-

havioral intervention tool. However, it also became

clear that the system needed to be redesigned to bet-

Development and Evaluation of a Sensor-based System for Remote Monitoring and Treatment of Chronic Diseases - The

Continuous Care & Coaching Platform

ter meet the user’s needs in terms of being more user

friendly and being able to be integrated into the daily

work practices of the health care professionals. The

early involvement of patients and professionals in the

requirements analysis and the design process is cru-

cial for this. Therefore, for developing the require-

ments of C3PO, we chose to apply an iterative, user-

centred design approach, as shown in Figure 1. In this

approach the designers take the users as the starting

point for design and involve them in the evaluation of

design choices.

User knowledge

State of the art

Scenario refinement (FICS)

Scenario (PACT)

Functional requirements

Technical specifications

System design

Platform implementation

Conceptual

Functional

Technical

Evaluation

Figure 1: General iterative design process of the platform.

Exactly how to usefully involve users in the de-

velopment process is far from obvious because of

the knowledge gap between the users and developers

and their differences in language use which seriously

complicates effective communication (Huis in’t Veld

et al., 2010). Scenarios are considered an effective

technique to bridge these gaps (Carroll, 2000). A sce-

nario is a detailed storyline that describes the users

daily activity in the setting of the envisioned applica-

tion of the system to be designed, from which require-

ments can be elicited.

A scenario was developed following the People-

Activity-Context-Technology (PACT) framework.

Using this approach forces to think about all the

relevant stakeholders that will be using the new

system, the actions that they perform and the (med-

ical) context in which they perform them, as well

as the technological innovations that are necessary

for successful implementation. Incorporating the

principles of evidence-based medicine into PACT

scenario development provides a starting point for

more effective and efﬁcient design of telemedicine

applications (Huis in’t Veld et al., 2010). On the

one hand, we performed a state of the art literature

research, and on the other hand we assessed the

users medical knowledge and needs by means of

questionnaires, PACT tables and interviews. From

this approach the possibility to have more insight into

the health status of the patient for both patient and

professional became apparent and the clear need to

involve the healthcare professional in the treatment

emerged. Besides the user-centred perspective of

PACT, the “designer-centric” components related to

system use were added to the scenario: Function and

events, Interactions and usability issues, Content and

structure, Style and aesthetics (FICS) (Benyon and

Macaulay, 2002). The FICS approach is more system

descriptive and provides insights for technicians

to consider the technical speciﬁcations. It became

apparent that the platform needed a clear distinction

between sensors, the Smartphone, server applications

and web portals, as these are the high level building

blocks for each of our telemedicine applications.

Besides the clear separation of these building blocks,

each of the platform components should have a

modular set up as well, so they can be easily extended

or changed to serve different user groups. Finally,

from the requirements and analysis of the gathered

data, there were strong indicators that the way we

provide feedback to the patients was rather sub-

optimal in terms of its coaching ability. The scenario

was updated, containing PACT elements combined

with FICS elements. From the developed scenario,

requirements were elicited by group discussions in a

participatory design setting, for example conducted

during plenary project meetings where both users

and developers were present. For this purpose,

tables, textual descriptions, and mock-ups were used.

The requirements were prioritized following the

MoSCoW method (must have, should have, could

have, wouldnt have). A must have for the platform

included for example: “The system must continuously

monitor the amount of activity, the activity pattern,

and activity intensity”. The addition of an emergency

alarm function on the other hand was an example

of a could have. Following this iterative design

approach

, the platform was designed, implemented

and evaluated. The evaluation covered conceptual

(lessons learned), functional (e.g. usability) and

technical (e.g. battery time) aspects, and the scenario

and requirements were updated accordingly. More

details regarding the evaluation are given in Section

4.2 The Platform

From applying the iterative approach described

above, it emerged that the activity monitoring system

had to be highly ﬂexible and conﬁgurable to accom-

modate the various requirements coming from differ-

ent research projects and research goals. Each study

employs its own experiment protocol and the different

Multiple groups of end-users (totalling around 50),

covering a variety of patient populations where involved

throughout a number of different projects.

EHST/ICGREEN 2012

protocols included different procedures such as use of

a web portal, or showing a graph to the patient, or pro-

viding regular feedback messages, or prompting the

user to answer various types of questions. We also

want to offer support for easy integration of different

sensor types and the possibility to use multiple sen-

sors simultaneously. Figure 2 gives a high level view

of the architecture.

Sensor

Smartphone

Syncer

Server

Syncer

Data

Base

DBServer

Web Portal

Patient

Care Provider

Server

Figure 2: High level architecture overview of the platform.

The individual components are explained in more

detail below.

Sensors. The ProMove-3D wireless sensor was de-

signed to provide a trade-off among performance,

computational and storage resources, wireless capa-

bilities, low-power operation and wearable form fac-

tor. The sensor node can capture, process and com-

municate wireless full 3D motion and orientation in-

formation. The main building blocks of the sensor

node are the following:

• The sensor modules, including accelerometer, gy-

roscope and magnetic compass. All chips are

single-chip, three-axial, fully digital sensors. The

possibility exists to include additional sensors.

• The master microcontroller, which implements all

sensor sampling and handling, synchronization

mechanisms and the application logic.

• The wireless modules. ProMove-3D implements

IEEE 802.15.4 compatible wireless communica-

tion in the 2.4 GHz license-free band as well

as Bluetooth communication through an optional

module.

Through this design, the ProMove-3D becomes a

versatile sensor node supporting a large range of ac-

tivity sensing applications. It can operate as a long-

term activity monitoring device, with low power con-

sumption, long sleep cycles and intermittent wireless

communication for periodic data transfers. At the

other end of the spectrum, it can support high data

rate, full 3D motion and orientation data capturing

and streaming wireless within accurate synchroniza-

tion bounds for accurate motion analysis. Figure 3

shows the sensor node from top, side and bottom. The

side cap is detachable and gives access for charging

the battery. When the cap is attached, the node is au-

tomatically switched on. The node can be worn by

means of a multi-functional clip holder, which allows

various attachments. Two modalities are depicted in

Figure 3: a belt clip, which is meant to be used for

wearing the node around the waist, and an elastic

band clip, which is meant to be used for wrapping the

node with an elastic band, for example around limb

segments.

Figure 3: ProMove-3D node top view with clip holder (left),

side view with belt clip (middle) and bottom view with elas-

tic band clip (right).

Smartphone. In order to ensure ﬂexibility — an

important system requirement — we designed a

Smartphone software framework in which applica-

tions can be deﬁned by linking together various mod-

ules. Each module performs a typically small, clearly

deﬁned task and delivers its output to a central com-

munication module, called the Hub. Other modules

can subscribe to this information and are notiﬁed

when new output becomes available. For example: a

BluetoothModule is tasked with opening, maintaining

and closing Bluetooth connections within the device.

As output it delivers a stream of data coming from the

sensor, as well as updates regarding the status of the

connection. A UserInterfaceModule can subscribe to

the status messages and provide the user with a warn-

ing if a Bluetooth connection is lost. Software de-

velopers are free to develop their modules without

speciﬁc limitations and, as long as they adhere to the

simple communication protocol with the central Hub,

they can share their output with other modules. In

order to provide the connection between the Smart-

phone and server side services, such as web portals

that provide information to healthcare professionals,

the AndroidSync module takes care of synchronizing

data between Smartphone and server.

Server. On the server side of the platform, the core

component is R2D2 (the Roessingh Research and

Development Database). R2D2 is designed to han-

dle a wide variety of data types such as streaming bio

signal data, questionnaire results, context information

and subject information. R2D2 offers a ﬁxed struc-

ture for storing data uniformly from multiple research

Development and Evaluation of a Sensor-based System for Remote Monitoring and Treatment of Chronic Diseases - The

Continuous Care & Coaching Platform

Figure 4: The old (left) and new (right) versions of the

Smartphones and graphical user interface.

projects and applications. This uniform way of stor-

ing data makes it possible to do simple comparison

studies on data gathered by different researchers in

different projects. The design of the data structure is

based on two core principles: (1) the subject is at the

top of the data structure, and (2) all measured data

is linked to a timeline. A subject is an anonymous

participant in a research project and all data that is

measured is subsequently linked to a subject. Fur-

thermore, for all data that is stored, the time at which

the data was gathered is the important factor in link-

ing various layers of data to each other. The subject’s

activity data, for example, is stored in a signal layer

as a constant stream of <time, value>-pairs that indi-

cate their performed activity intensity during the day.

Another layer is used for storing the time at which the

subject received feedback from the system and which

exact message was displayed to the user; these are

events that are stored in an event layer. Similar lay-

ers exist for storing interval data (in which a period of

time can be labelled as e.g. walking or lying), ques-

tionnaire data (in which questionnaire results can be

stored) and video data (if a camera is used for e.g.

gait analysis). Figure 5 shows a screenshot of a data

viewer application, the RRDToolkit, showing various

layers of measured data on a timeline. A server ap-

plication handles secure synchronous access to the

database. Password protected user accounts provide

access to the various sets of data that are stored in the

database, and locking mechanisms are in place to pre-

vent simultaneous write operations, that could result

in data corruption.

Web Portal. To provide a useful view on the data

for biomedical researchers, patients and healthcare

professionals, we designed a modular framework for

developing web portals, in much the same way as the

Figure 5: A screenshot of the RRDToolkit, a data viewer for

researchers, showing an activity graph (signal layer), feed-

back events (event layers) and weather data (interval layers)

for a day of measurement of a Chronic Low Back Pain pa-

tient.

Smartphone application. Different remote monitoring

and treatment applications focus on different types of

measurements and different types of feedback, but a

general component in many applications is a portal

that allows patients and healthcare professionals to

track disease progression or health status. The pri-

mary goal of the web portal is to display measurement

data to the patient or the healthcare professional in an

easy to understand way. Figure 6 shows a screenshot

of the web portal developed within the Dutch “Condi-

tionCoach” project, showing the view for COPD pa-

tients who are able to track their own levels of physi-

cal activity, and progression over the week, month or

year. But the portal can also be used for gathering

new data. For example, an electronic triage applica-

tion was developed for COPD patients where patients

receive a set of questions on their health status (Efﬁng

et al., 2009) on a daily basis via the portal and can give

their responses via the portal. In a trial with COPD

patients, subjects were asked daily if their symptoms

were more severe than normal. If the patient answered

“yes”, they were asked to report on all their symptoms

irrespective of whether the level of each was normal,

slightly increased, or greatly increased. A rule-based

system (an action plan) implemented on the web por-

tal decides when the patient needs to start taking med-

ication. This diary form was successfully used in clin-

ical care as part of a regime of self-treatment of exac-

erbations in COPD patients.

5 EVALUATION

Earlier versions of the platform, and especially the

second platform iteration was extensively evaluated

in trials with different patient groups and their health-

care professionals by means of questionnaires and in-

depth semi-structured interviews. An important out-

EHST/ICGREEN 2012

Figure 6: Screenshot of the Dutch CoCo (ConditionCoach)

project web portal, showing a weekly overview of activity

data (center) and links to all the separate available modules

(top) for a patient.

come was the perceived acceptability of the system,

especially by physical therapists.

For C3PO, a small-scale evaluation study was per-

formed to optimize the usability, and to assess inten-

tion of use of the patients. In the Netherlands, six pa-

tients with a clinical diagnosis of stable COPD were

included. COPD is a patient group of mainly elderly

people, who are in many cases not as familiar with

modern technology as younger generations. The ex-

periment was designed to simulate daily use of the

system and daily life activities in a laboratory set-

ting. Each patient wore the system while performing

a number of daily life tasks in the laboratory setting

while the system measured activity levels and gave

feedback. Usability scores were assessed by the Sys-

tem Usability Scale (SUS). For the activity monitor-

ing and feedback a mean SUS score of 67.1 was found

(range: 47.5 - 90.0), which corresponds to the aver-

age scores found in the literature, but also indicated

that usability should be improved. A questionnaire,

speciﬁcally designed for the platform, was ﬁlled in by

the patients to assess the usability of speciﬁc parts of

the system. The questionnaire consisted of 27 state-

ments about the system with a 7-point Likert scale

ranging from ‘strongly disagree’ to ‘strongly agree’.

Example statements include: “The meaning of the

graph is not clear to me” or “I ﬁnd it easy to un-

lock the Smartphone”. The questionnaire results pro-

vided us with detailed information about every aspect

of the Smartphone application and highlighted further

points for improvement. These improvements were

implemented before clinical trials started and we ex-

pect that the usability score will improve following

the modiﬁcations to the application.

The web portal was evaluated in a separate trial.

An evaluation study was carried out into the experi-

ences of end-users during a four week testing period

with the C3PO portal. Data was collected through in-

depth semi-structured interviews with ﬁve COPD pa-

tients, two physical therapists, one nurse practitioner

and two lung specialists, and through participative ob-

servations. Five aspects of acceptability of innova-

tions were addressed (relative advantage, compatibil-

ity, ease of use, image, and voluntariness). Quantita-

tive data was used as a secondary source of data pro-

viding supportive information.

There was a perceived high potential acceptabil-

ity, especially to physical therapists. The electronic

diary was considered easy to use and possibly of high

value to both patients and professionals. The use of a

web based portal was accepted by patients but profes-

sionals experienced it as time consuming and hence

professionals’ acceptance was lower than patients’ ac-

ceptance. Quantitative data showed that patients used

the portal at variable hours throughout the day, for 3 to

4 minutes per session. In addition, several points for

enhancing the portal emerged. These improvements

were implemented and, to increase acceptance by the

professionals, the portal was improved in close col-

laboration with the healthcare professionals. Taking

into account the feedback from patients and health-

care professionals allowed us to build a remote mon-

itoring and treatment platform with a high usability

and acceptability for its different classes of end-users.

6 DISCUSSION AND WORK IN

PROGRESS

Over the years many small and larger trials with a

wide variety of patients have been performed, and

following each trial the feedback from the various

users was taken into account. The re-design and re-

implementation of the C3PO platform resulted in a

system that is stable and ﬂexible in operation, as well

as simple to use for patients and healthcare profes-

sionals alike. However, one of the major issues that

arose from the requirements and analysis of older ver-

sions of the platform has not yet been fully addressed.

This concerns automated coaching, the means to pro-

vide real time feedback to the patients on their daily

physical activity in order to encourage behavioural

change. In the current platform, feedback is still

simplistic: feedback messages are presented at ﬁxed

time intervals, related to the amount of activity per-

formed compared to a reference value. Because ef-

fective coaching is a crucial element in affecting be-

havioural change, current research is focussing on de-

veloping an intelligent coach that runs on the Smart-

phone and acts as a personal companion that helps

Development and Evaluation of a Sensor-based System for Remote Monitoring and Treatment of Chronic Diseases - The

Continuous Care & Coaching Platform

patients achieve their activity goals. It is based on

the principles that feedback should be tailored to in-

dividual users in terms of its timing, content and style

of presentation to the patient. The next evolution of

the system will be able to learn to predict the opti-

mum timing for providing feedback by analyzing pre-

viously given feedback messages and learning when

a patient is likely to respond well to a given message

by relating relevant context factors to patient compli-

ance (op den Akker et al., 2010). In related work by

Wieringa et al. (Wieringa et al., 2011) the Smart-

phone application also learns to adapt the feedback

message content to the patient. Feedback messages

are stored in a structured manner based on message

content and style. During operation, the system stores

the reaction of the patient to each feedback instance

in order to make informed decisions on which type of

messages are most likely to elicit a positive response

the next time feedback is needed. As for the mode and

style of presentation of feedback to the patient, this re-

mains a signiﬁcant challenge in Human Computer In-

teraction research. Ongoing and planned research will

evaluate the effects of the smart feedback coach on

patient’s physical activity patterns and their percep-

tion of the system in terms of usability and treatment

compliance. Currently, trials are running in which

COPD patients will use the smart feedback coach for

a period of three months. In the meantime, large scale

evaluations of the remote monitoring and treatment

platform are running in which the platform is used in

a daily clinical care setting. These larger scale eval-

uations will show us how the system performs under

stress and will enable us to receive valuable feedback

on the latest version of the platform which can help

us to keep improving the system.

Based on the results of the research described in

this paper, we are conﬁdent that C3PO has the poten-

tial to contribute to delivering more cost effective and

qualitative healthcare. We believe our approach will

be part of the solution by providing more cost effec-

tive rehabilitation and secondary prevention since it

supports self management of patients; thereby shift-

ing the focus of control towards the patient. Using

C3PO the patient can determine where, when and how

intensively he follows his treatment regimen. It also

enables the health care professional to treat more pa-

tients at the same time, decreasing the per-patient cost

of highly trained personnel. Last but not least, auto-

matic data collection is facilitated by the use of this

technology and the data aggregated will result in large

corpora of clinical data which has the potential to sup-

port evidence based medicine. On the one hand it en-

ables an efﬁcient comparison of efﬁcacy of different

interventions, and on the other, data mining can gen-

erate new clinical knowledge of general relevance as

well as determining optimal treatments for a speciﬁc

patient.

ACKNOWLEDGMENTS

The C3PO platform is a group effort from the Roess-

ingh Research and Development, Telemedicine Clus-

ter and the authors would like to thank all of our col-

leagues for their commitment to the platform devel-

opment.

REFERENCES

Benyon, D. and Macaulay, C. (2002). Scenarios and the

HCI-SE design problem. Interacting with Computers,

14(4):397–405.

Berg, M. (1999). Patient care information systems and

health care work: a sociotechnical approach. Interna-

tional Journal of Medical Informatics, 55(2):87–101.

Broens, T. H., Huis In’t Veld, R. M., Vollenbroek-Hutten,

M. M., Hermens, H. J., Van Halteren, A. T., and

Nieuwenhuis, L. J. (2007). Determinants of success-

ful telemedicine implementations: a literature study.

Journal of Telemedicine and Telecare, 13(6):303–309.

Burkow, T. M., Vognild, L. K., Krogstad, T., Borch, N. l.,

Ostengen, G., Bratvold, A., and Risberg, M. J. (2008).

An easy to use and affordable home-based personal

eHealth system for chronic disease management based

on free open source software. Studies In Health Tech-

nology And Informatics, 136:83–88.

Carroll, J. (2000). Five reasons for scenario-based design.

Interacting with Computers, 13(1):43–60.

Chen, G., Yan, B., Shin, M., Kotz, D., and Berke, E. (2009).

MPCS: Mobile-phone based patient compliance sys-

tem for chronic illness care. Mobile and Ubiquitous

Systems Networking Services MobiQuitous 2009 Mo-

biQuitous 09 6th Annual International, pages 1–7.

Czabke, A., Loeschke, J., and Lueth, T. C. (2011). Concept

and modular telemedicine platform for measuring of

vital signs, ADL and behavioral patterns of elderly in

home settings. Architecture, pages 3164–3167.

Efﬁng, T., Kerstjens, H., van der Valk, P., Zielhuis, G., and

van der Palen, J. (2009). (Cost)-effectiveness of self-

treatment of exacerbations on the severity of exacer-

bations in patients with COPD: the COPE II study.

Thorax, 64(11):956–962.

Ekeland, A. G., Bowes, A., and Flottorp, S. (2010). Effec-

tiveness of telemedicine: a systematic review of re-

views. International Journal of Medical Informatics,

79(11):736–771.

Hermens, H. J. and Vollenbroek-Hutten, M. M. (2008).

Towards remote monitoring and remotely supervised

training. Journal of Electromyography and Kinesiol-

ogy, 18(6):908–19.

EHST/ICGREEN 2012

Huis in’t Veld, R. M., Widya, I. A., Bults, R. G., Sandsj

L., Hermens, H. J., and Vollenbroek-Hutten, M. M.

(2010). A scenario guideline for designing new tele-

treatments: a multidisciplinary approach. Journal of

Telemedicine and Telecare, 16(6):302–307.

Jones, V. M., Hermens, H. J., Leijdekkers, P., and Rienks,

R. (2009). Extending remote patient monitoring with

mobile real time clinical decision support. In Annual

Symposium of the IEEEEMBS Benelux Chapter.

Kumar, S., Kambhatla, K., Hu, F., Lifson, M., and Xiao,

Y. (2008). Ubiquitous Computing for Remote Cardiac

Patient Monitoring: A Survey. International journal

of telemedicine and applications, 2008(iv):459185.

Laplante, C. and Peng, W. (2011). A Systematic Review

of e-Health Interventions for Physical Activity: An

Analysis of Study Design, Intervention Characteris-

tics, and Outcomes. Telemedicine and ehealth: the

ofﬁcial journal of the American Telemedicine Associ-

ation, 17(7):509–523.

Lewis, B., Williams, D., Dunsiger, S., Sciamanna, C.,

Whiteley, J., Napolitano, M., Bock, B., Jakicic, J.,

Getz, M., and Marcus, B. (2008). User attitudes to-

wards physical activity websites in a randomized con-

trolled trial. Preventive Medicine, 47(5):508–513.

op den Akker, H., Jones, V. M., and Hermens, H. J. (2010).

Predicting Feedback Compliance in a Teletreatment

Application. In Proceedings of ISABEL 2010: the

3rd International Symposium on Applied Sciences in

Biomedical and Communication Technologies, Rome,

Italy.

Polisena, J., Tran, K., Cimon, K., Hutton, B., McGill,

S., Palmer, K., and Scott, R. E. (2010). Home

telehealth for chronic obstructive pulmonary disease:

a systematic review and meta-analysis. Journal of

Telemedicine and Telecare, 16(3):120–127.

ahl, O., Gamb

ack, B., Hansen, P., Turunen, M., and Haku-

linen, J. (2008). A Mobile Fitness Companion. Sci-

ence And Technology.

United Nations, Department of Economic and Social Af-

fairs, Population Division (2002). World Population

Ageing 1950-2050. United Nations Publications, New

York, USA.

Van Weering, M. G., Vollenbroek-Hutten, M. M., T

onis,

T. M., and Hermens, H. J. (2009). Daily physi-

cal activities in chronic lower back pain patients as-

sessed with accelerometry. European Journal of Pain,

13(6):649–654.

Wieringa, W., op den Akker, H., Jones, V. M., op den Akker,

R., and Hermens, H. J. (2011). Ontology-Based Gen-

eration of Dynamic Feedback on Physical Activity. In

Proceedings of the 13th Conference on Artiﬁcial In-

telligence in Medicine (AIME), pages 55–59, Bled,

Slovenia. Springer.

Development and Evaluation of a Sensor-based System for Remote Monitoring and Treatment of Chronic Diseases - The

Continuous Care & Coaching Platform