ASARM

A System for CFS/ME Monitoring and Treatment

Philip McDermott

, Steve Pettifer

and Paul Abeles

School of Computer Science, The University of Manchester, Manchester, U.K.

Royal Manchester Children’s Hospital, Manchester, U.K.

Keywords:

Chronic fatigue syndrome, Myalgic encephalomyelitis, Actigraphy, CBT, Mobile computing.

Abstract:

CFS/ME (Chronic Fatigue Syndrome/Myalgic Encephalomyelitis) affects up to 2.5% of adults in the UK and

USA and between 1% and 2% of children in the UK. Sufferers report that they are low on energy, and ﬁnd

performing everyday activities difﬁcult. The illness is commonly treated using Cognitive Behavioural Therapy

(CBT), which aims to help patients learn how to build up their energy levels in a gradual way, and how best to

spend and preserve their energy. A crucial aspect of this treatment is for the health care professional to monitor

and record how the patient spends time on a day-to-day basis, then prescribe appropriate and precise baseline

levels for periods of rest, sleep and activity. These levels are then gradually adjusted as the patient’s condition

improves. Current methods typically rely on paper diaries, however, these offer little guidance to the patient

and are time consuming for the health care professional to analyse. The ASARM (Advanced Sleep, Activity

and Rest Monitoring) system combines an electronic diary with automated recording of actigraphy data, with

the aim of improving the process of assessing, monitoring, prescribing for, and then treating patients with the

condition.

1 INTRODUCTION

CFS/ME (Chronic Fatigue Syndrome/Myalgic En-

cephalomyelitis) is a condition characterised by a de-

bilitating fatigue that is not relieved by rest, and per-

sists over at least six months (Reeves et al., 2005;

Royal College of Paediatrics and Child Health, 2004;

Fukuda et al., 1994). It typically co-occurs with other

symptoms such as muscle pain, headache, sore throat

and memory or sleep problems. While the preva-

lence of the illness can be difﬁcult to deﬁne, since

this depends on the diagnostic criteria used, recent

studies suggest up to 2.5% of the population in the

UK and USA may be affected (Avellaneda Fernandez

et al., 2009; Reyes et al., 2003; Jason et al., 1999),

and between 1% and 2% of children in the UK. Ap-

proximately one in ten of these will be of a severity

that means the illness is disabling, leaving the patient

housebound (Farmer et al., 2004). The most recently

published UK guidelines (National Institute of Clin-

ical Excellence (NICE), 2007) recommend immedi-

ate assessment by a specialist, and (Hinds and Mc-

Cluskey, 1993) found that children and young people

with the condition had a signiﬁcantly better prognosis

than adults if diagnosed early.

Patients with the condition ﬁnd it difﬁcult to reg-

ulate their energy expenditure. They can frequently

become very low on energy, which then makes per-

forming everyday activities challenging. A critical

aspect of CFS/ME is that an activity can be costly to

the energy levels of the person with the condition by

being either physically, emotionally or cognitively de-

manding. This means that a thorough understanding

of patients’ lifestyles is vital to prescribing effective

treatment.

We present a technology-based solution for both

monitoring and then treating CFS/ME by collecting

and analysing data around patients’ everyday activ-

ities, making use of subjective metrics recorded by

the patients themselves and objective data describing

their movements.

2 CLINICAL NEED

Currently, the most prominent and successful treat-

ment of CFS/ME is the use of Cognitive Behavioural

Therapy (CBT) to help people with the condition

learn how to build up their energy levels in a gradual

way, and inform them how best to spend and preserve

160

McDermott P., Pettifer S. and Abeles P..

ASARM - A System for CFS/ME Monitoring and Treatment.

DOI: 10.5220/0003765701600166

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2012), pages 160-166

ISBN: 978-989-8425-91-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

the right amount of their energy (Stulemeijer et al.,

2005). This approach is applicable to both young

people and adults, and a recent randomised control

trial (the PACE Trial

) using CBT for adults with

CFS/ME has further reinforced its clinical effective-

ness for treating this condition (White et al., 2011).

Generally, the CBT approach to managing and

treating CFS/ME is to monitor and record, with the

patient’s help, how his or her time is divided between

activities of varying energy-expenditure, speciﬁcally

periods spent sleeping, resting and being active. Av-

eraged over a period of days, these amounts serve as

baselines, which are then rebalanced to more healthy

and sustainable proportions. An important aspect of

this treatment is the method by which the patient’s

time is monitored and recorded, since this is the in-

formation that setting appropriate and precise base-

line levels relies on. In current clinical practice (for

example, in (Royal College of Paediatrics and Child

Health, 2004)), the process is as follows. In phase

one of treatment, baseline levels are established by

analysing data recorded by the patient in a paper di-

ary over a set period of time (typically a fortnight). In

these diaries, the patient records the amount of time

spent doing daily activities (including periods sleep-

ing or resting). This data informs the clinician who

establishes, by averaging the ﬁgures to remove daily

ﬂuctuations, the exact amounts of sleep, rest and ac-

tivity that the patient is capable of achieving without

detrimental effects on his or her health. In phase two,

the patient aims to follow these set ‘budgets’, and

over the course of the treatment they are gradually

adapted—by the clinician—to achieve a more healthy

balance. As part of a more detailed assessment, the

patient supplements activity diary information with

subjective ratings: a record of reactions to, and ap-

praisals of, the activities being performed. These sub-

jective ratings may be, for example, the sense of mas-

tery achieved or the level of fatigue felt while carrying

out a task, or the feeling of refreshment following rest

or sleep. These ratings contribute in two ways: they

provide additional information for the clinician to aid

prescribing a patient’s daily schedule, and also help

the patient by adding a degree of reﬂection to his or

her daily routine.

The collection of data through paper diaries, and

the data itself, can be problematic in three ways. First,

the data can be highly subjective, reﬂecting patients

own opinions on how they feel. Second, the diary

information relies upon the patient’s diligence for its

accuracy and completeness. Third, the actual act of

data collection places additional strains on the patient,

in particular on memory, and especially in the early

http://www.pacetrial.org/

stages of care which are often characterised by feel-

ings of hopelessness and disempowerment.

In the following sections, we describe how each

of these problems may be addressed, and a system we

have built to put these improvements into practice.

3 METHOD

Our method for addressing the shortcomings of the

current clinical practice (described in the previous

section) is to apply cost-effective, robust, technolog-

ical solutions to key aspects of the existing method-

ology, and to integrate and automate processes where

possible to provide a complete integrated system for

monitoring and treating the condition. Existing work

by (Abeles et al., 2009) involving the application of

computer-assisted CBT to treat a not dissimilar con-

dition (adolescent depression) has proved successful

(also see (Robinson et al., 2011)).

To compliment the subjective diary data used in

current practice, more objective data in the form

of gross-motor movement measurements can be col-

lected alongside it. This can be captured through an

actigraphy device: a small device worn on the body to

measure and record the patient’s movement through-

out the day. A variety of small, lightweight and low-

cost devices exist for this purpose, and have been tri-

alled in limited ways as part of clinical research. For

example, the Philips Respironics Actiwatch

has been

used as a secondary outcome measure while monitor-

ing CFS/ME patients (Kop et al., 2005). Prior to our

work, however, actigraphy data has not been used di-

rectly in treatment settings.

Data gathered using such devices is invaluable in

understanding patients’ energy expenditure, but does

not replace the existing diary system, since not all ac-

tivities will necessarily afford movement on the part

of the patient. A fundamental aspect of the condi-

tion is that physical, emotional and cognitive effort af-

fect the overall energy expenditure associated with an

activity (Afari and Buchwald, 2003). A cognitively-

demanding activity without physical movement, for

example, would not be detected by a system rely-

ing solely on actigraphy. If diary information is still

to be collected, however, the issues of accuracy and

completeness must be addressed. Our system uses an

electronic diary that can record information, speed up

the process of entering, and also serve as a visual aid.

Such a diary system can guide patients as they plan

and record their activities, helping to account for their

http://www.healthcare.philips.com/main/homehealth/sl

eep/actiwatch/default.wpd

ASARM - A System for CFS/ME Monitoring and Treatment

161

time and to classify it accordingly. Further, the di-

ary can address the third issue described, the cognitive

strain on the patient, by being more accessible than a

paper system, and by prompting the user where neces-

sary (for example, to record subjective feelings along-

side activities). During the ﬁrst phase of treatment,

where the patient is simply recording all activity, the

device can assist data entry, reducing the amount of

cognitive workload. In the second phase of treatment,

where the patient is following a prescription, the de-

vice can instruct the user as to what type of activity

should be carried out, and when, further assisting the

patient in his or her scheduling.

In addition to addressing these issues, an elec-

tronic system can improve communication between

patient and clinician. Both devices—the actigraphy

device and diary—communicate with a central server,

so that the data can be synchronised and viewed by the

clinician. The clinician can view both diary and actig-

raphy data daily via a web interface, unlike the tradi-

tional system whereby access to this data was limited

to scheduled appointments, typically occurring only

fortnightly. This allows actigraphy data to be used

as a primary treatment measure, to inform the clin-

ician at the point at which they are creating a pre-

scription for the patient. Not only can the clinician

monitor much more closely the progress of the pa-

tient, changes to the prescription can be made much

more rapidly: the clinician can revise and then ‘push’

a new prescription to the patient at any time. It is

these small but incremental updates to a prescription

that help the patient’s condition improve. Not hav-

ing to wait for more costly and less-frequent face-to-

face appointments for each adjustment can potentially

speed up the recovery of the patient.

4 RESULTS

The system developed in this work comprises: a cen-

tral server, used for data storage and communication;

a web front-end for the clinician; and a patient pack-

age. The patient package comprises: a wrist-worn ac-

celerometer device to measure sleep, activity and rest;

a docking station for the device; and an electronic di-

ary (as a smartphone application). An overview of the

architecture is shown in Figure 1.

4.1 Server and Web Application

The server combines a Pylons

web application and

a PostgreSQL

database. The database stores infor-

http://www.pylonsproject.org

http://www.postgresql.org

Figure 1: Architecture of the ASARM system, showing the

relationship between the clinician’s interface, the patient’s

interface and the central server.

mation about registered devices, activities, subjective

activity data, actigraphy data, prescriptions and al-

lowances. The web application (Figure 2) lets the

clinician view each registered ASARM device, its ac-

tivity information, and also create, view and amend

prescriptions for the patients associated with each de-

vice.

The primary role of the server is to provide Repre-

sentational State Transfer (Fielding, 2000) end-points

for the diary application to communicate with. New

data, and requests for existing data, are sent asyn-

chronously via HTTP from the mobile device to the

server, which in turn communicates with the database

and responds appropriately. The web application

communicates with the server in the same fashion.

4.2 Diary Application

The primary interaction with the ASARM system for

the patient is through the diary application, imple-

mented as an app for the iPhone (Figure 3). The di-

ary app’s main purpose is to provide a quick and easy

method for the patient to record day-to-day tasks. It

can be in one of two modes: data collection mode

or prescription mode. Data collection mode is used

to learn how the patient’s time is spent. Prescrip-

tion mode is used once a prescription has been cre-

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

162

Figure 2: The clinician’s web application, showing task timeline and actigraphy data, activities and calculated baselines.

ASARM - A System for CFS/ME Monitoring and Treatment

163

Figure 3: The diary app, in prescription mode.

Figure 4: The diary app, showing the task builder.

ated by the clinician and delivered to the patient: this

mode is designed to help patients follow their pre-

scription as closely as possible. Data collection gen-

erally lasts around two weeks, during which time the

patient uses the Task Builder to quickly describe the

activity he or she is engaged in. The Task Builder

presents high-level types of activity to the user (i.e.

the patient), such as leisure, household, travel, etc.,

followed by additional types of information, such as

location, company, and so on, to quickly and accu-

rately describe an activity (see Figure 4). For exam-

ple, the user could create the activity ‘Doing home-

work at home with a friend’ in 5 touches, or the ac-

tivity ‘travelling to school on the bus, alone’ in 4

touches. This input mechanism is designed to min-

imise the interaction with the diary as much as pos-

sible: it is vital that the monitoring and treatment of

the patient do not increase the workload—either phys-

ically or cognitively—unnecessarily, and become a

signiﬁcant activity in itself. For instances where the

use of the task builder do not enable the user to de-

scribe the activity fully, manual entry using the de-

vice’s keyboard is possible. When the diary is in pre-

scription mode, instead of the Task Builder the user

chooses an action from a list of prescribed activities,

grouped by energy-expenditure. This encourages pa-

tients to follow the prescription set for them, and al-

lows the application to keep track of their time more

accurately. When an activity is not in the prescribed

list, the patient can use the Task Builder or manual

entry to add the new activity.

The diary application communicates with the

server using standard HTTP requests over the cellular

phone network (or a wireless broadband connection if

available). When a connection cannot be established,

the diary application queues requests to be sent when

a connection becomes available.

4.3 Actigraphy Device and Dock

The actigraphy device used is a 3-axis accelerome-

ter unit

. The device is small and lightweight, worn

on the wrist in a small pouch, and can record data

for up to 60 hours. It charges and transfers data via

USB. For this system the device is set to sample at

a rate of 10Hz in +/-2g range. When uploaded to

the ASARM server, these raw values are processed to

take the change in magnitude of the three-dimensional

vector each 1/10th of a second, and then the mean is

taken over each minute for display purposes.

A netbook is used as a ‘docking station’ for the

actigraphy device: when it is plugged into the docking

station (using a standard USB connection), all new

data on the device is backed up and uploaded to the

ASARM server. The netbook is conﬁgured to use

the iPhone as an internet ‘hotspot’ and so can auto-

matically upload data to the server without additional

conﬁguration at the patient’s home. This obviates the

need for the user to provide an internet connection for

the system to work. Unlike tele-medicine systems that

may need to optimise the use of a narrow-band phone

speciﬁcally, a Gulf Coast Data Concepts X6-2mini

http://www.gcdataconcepts.com/x6-2mini.html

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

164

line (for example, (Lai et al., 2005)), or make use of

an existing broadband connection (for example, (Lai

et al., 2007), high-throughput is not a requirement of

this system, and so ease of use can be given a higher

priority.

The netbook additionally acts as a power source

and charges the actigraphy devices while they are

docked. In operation, a pair of these devices is ac-

tually used: one is worn throughout the day, while

the other is worn at night. The night time device is

plugged into the docking station during the day, and at

bedtime (when the user of the diary selects ‘bedtime’

as the current activity), the two devices are swapped,

so that night time movement is recorded whilst the

day time device uploads data and recharges.

Following the initial period of activity monitor-

ing, using the data from the web application (the di-

ary and actigraphy data in Figure 2) the clinician cat-

egorises each activity into sleep, rest, or activity, or

even more precisely (deep or light rest; light, moder-

ate or vigorous activity). The rating of activity will

depend on both the actigraphy data and the descrip-

tion of the task, thereby allowing both physical and

cognitive aspects to be taken into account. For ex-

ample, watching a favourite television programme,

whilst being physically undemanding, is cognitively

much more demanding than spending the time just

sitting. When the clinician has completed the cat-

egorisation, the application automatically generates

starting baselines using the mean values of these cat-

egories over the monitoring period, as well as a set

of rated activities to be sent to the diary application

(Figure 2). This prescription—the baselines and set

of rated activities—is sent to the diary app, which

then changes mode from data collection to prescrip-

tion mode. From this point, the patient has a prede-

ﬁned set of rated activities and daily baseline amounts

to follow. The diary app interface changes accord-

ingly as described above, to show the patient his or

her activity choices, baseline amounts and ‘budgets’

of time for each category. The clinician continues to

monitor the patient’s activities, and can change and

update the prescription—either by altering the base-

lines or by adding or removing rated activities—-at

any time.

5 PRELIMINARY TESTS

The system as described has been ‘ﬁeld-tested’ over

a number of weeks to ensure each component works

as expected and reliably. This testing was successful,

and led to minor user interface improvements in the

diary application and docking station. The output of

the actigraphy device was compared with a commer-

cially available actigraphy device, and found to be of

comparable precision. It has not yet been used clini-

cally: this will be the next stage of the work.

6 CONCLUSIONS

The ASARM system described herein was developed

to address issues identiﬁed in the current best prac-

tice methods to treat CFS/ME. The current method

is to use paper diaries to collect data on the patient’s

day-to-day activities, then prescribe (and periodically

update) baseline targets for levels of energy expendi-

ture. Using this method, it is difﬁcult for the clini-

cian to accurately analyse and categorise the patient’s

activities, since there is a lack of objective move-

ment information to compliment the diary informa-

tion. Additionally, the paper diary system is reliant

on the patient’s diligence, and can increase cognitive

strain. The ASARM system combines the collection

of objective movement data with diary information

and subjective ratings, presenting them together to the

clinician so that a more informed judgement can be

made on the patient’s condition. This data is available

remotely, and is updated much more frequently then

in current practice. These features allows the clinician

to monitor the patient’s progress much more closely

without disrupting the patient’s schedule with formal

appointments.

The ASARM system aids the treatment process in

additional ways. The electronic diary provides ﬁner-

grained timetabling than the existing paper system,

increasing the accuracy of recorded periods of rest,

sleep and activity. The diary app also automatically

prompts patients to answer questions about their cur-

rent state of wellbeing at the appropriate time, rather

than the patient recording this information at a later

stage (with the paper diary system this is often only

at the end of each day). The task builder in the di-

ary application reduces the physical and cognitive de-

mands of ﬁlling in a paper diary—and reduces mem-

ory load—by speeding up diary entry and by auto-

matically recording start and end times of activities.

Finally, the application allows a regular bedtime to be

prescribed and suggested to the patient, and can be

used to inform the patient of sleep periods. This can

be useful where the patient’s regular sleeping pattern

is inverted, a regular feature of CFS/ME.

7 FUTURE WORK

The next clinically relevant stage in the ASARM pro-

ASARM - A System for CFS/ME Monitoring and Treatment

165

ject is to use the system in a formal pilot trial, with a

select group of CFS/ME patients. The planned trial

would be performed in two stages as follows. In

stage one, an initial sample of ﬁve patients with the

condition will be offered traditional CBT, but will

use the ASARM system alongside the CBT and the

completion of paper diaries. Qualitative responses

from the use of the devices will also be captured.

This stage will lead to the development of a clini-

cal ASARM protocol which will subsequently be fol-

lowed for stage two, with a further 25 patients, as

well as allowing iterative reﬁnement of the ASARM

system and software. All participants will use the

ASARM devices at the various stages of treatment:

assessment (establishment of baseline activity level),

early treatment phase and late treatment phase. Rele-

vant clinical outcome measures will also be collected,

pretreatment and post-treatment, as well as informa-

tion on the patient experience of the ASARM devices.

ACKNOWLEDGEMENTS

This work was funded by The Manchester Centre

for Integrating Medicine & Innovative Technology

(http://www.mimit.org.uk/).

REFERENCES

Abeles, P., Verduyn, C., Robinson, A., Smith, P., Yule, W.,

and Proudfoot, J. (2009). Computerized CBT for ado-

lescent depression (”stressbusters”) and its initial eval-

uation through an extended case series. Behavioural

and Cognitive Psychotherapy, 37(2):151–165.

Afari, N. and Buchwald, D. (2003). Chronic fatigue syn-

drome: a review. American Journal of Psychiatry,

60:221–226.

Avellaneda Fernandez, A., Perez Martin, A., Izquierdo Mar-

tinez, M., Arruti Bustillo, M., Barbado Hernandez,

F. J., de la Cruz Labrado, J., Diaz-Delgado Penas, R.,

Gutierrez Rivas, E., Palacin Delgado, C., Rivera Re-

dondo, J., and Ramon Gimenez, J. R. (2009). Chronic

Fatigue Syndrome: aetiology, diagnosis and treat-

ment. BMC Psychiatry, 9 Suppl 1:S1.

Farmer, A., Fowler, T., Scourﬁeld, J., and Thapar, A.

(2004). Prevalence of chronic disabling fatigue in

children and adolescents. British Journal of Psychi-

atry, pages 477–481.

Fielding, R. T. (2000). Architectural Styles and the Design

of Network-based Software Architectures. PhD thesis,

University of California, Irvine.

Fukuda, K., Straus, S. E., Hickie, I., Sharpe, M. C., Dob-

bins, J. G., and Komaroff, A. (1994). The Chronic

Fatigue Syndrome; A Comprehensive Approach to its

Deﬁnition and Study. Annals of Internal Medicine,

121:953–959.

Hinds, G. M. E. and McCluskey, D. R. (1993). A restrospec-

tive study of chronic fatigue syndrome. Proceedings

of the Royal College of Physicians, pages 10–14.

Jason, L., Richman, J. A., Rademaker, A., Jordan, K. M.,

Plioplys, A. V., Taylor, R. R., McCready, W., Huang,

C., and Plioplys, S. (1999). A community-based study

of Chronic Fatigue Syndrome. Archive of Internal

Medicine, 159:2129–2137.

Kop, W., Lyden, A., Berlin, A., Ambrose, K., Olsen, C.,

Gracely, R., Williams, D., and D.J., C. (2005). Ambu-

latory monitoring of physical activity and symptoms

in ﬁbromyalgia and chronic fatigue syndrome. Arthri-

tis and Rheumatism, 52(1):296–303.

Lai, A. M., Nieh, J., and Starren, J. B. (2007). REPETE2:

A next generation home telemedicine architecture.

AMIA Annual Symposium Proceedings, 1020.

Lai, A. M., Starren, J. B., and Shea, S. (2005). Architec-

ture for remote training of home telemedicine patients.

AMIA Annual Symposium Proceedings, 1015.

National Institute of Clinical Excellence (NICE)

(2007). Chronic Fatigue Syndrome/Myalgic En-

cephalomyelitis (or Encephalopathy): diagnosis and

management of CFS/ME in adults and children.

http://www.nice.org.uk/CG53.

Reeves, W. C., Wagner, D., Nisenbaum, R., Jones, J. F.,

Gurbaxani, B., Solomon, L., Papanicolaou, D. A.,

Unger, E. R., Vernon, S. D., and Heim, C. (2005).

Chronic fatigue syndrome – a clinically empirical ap-

proach to its deﬁnition and study. BMC Medicine,

3(19).

Reyes, M., Nisenbaum, R., Hoaglin, D. C., Unger, E. R.,

Emmons, C., Randall, B., Stewart, J. A., Abbey, S.,

Jones, J. F., Gantz, N., Minden, S., and Reeves, W. C.

(2003). Prevalence and incidence of Chronic Fatigue

Syndrome in wichita, kansas. Archive of Internal

Medicine, 163:1530–1536.

Robinson, A., Yule, W., Verduyn, C., Smith, P., and Abeles,

J. P. P. (2011). Stressbusters: The development of a

computerised CBT programme for adolescent depres-

sion. In Yule, W. and Udwin, O., editors, Increasing

Access to CAMHS, volume 30, pages 45–50. ACAMH

Occasional Papers, Association For Child and Adoles-

cent Mental Health.

Royal College of Paediatrics and Child Health (2004).

Evidence based guidelines for the manage-

ment of CFS/ME in children and young people.

http://www.ayme.org.uk/article.php?sid=10&id=137.

Stulemeijer, M., de Jong, L. W. A. M., Fiselier, T. J. W.,

Hoogveld, S. W. B., and Bleijenberg, G. (2005). Cog-

nitive behaviour therapy for adolescents with chronic

fatigue syndrome: randomised controlled trial. Brit-

sish Medical Journal, pages 14–17.

White, P., Goldsmith, K. A., Johnson, A., and et al. (2011).

Comparison of adaptive pacing therapy, cognitive be-

haviour therapy, graded exercise therapy, and special-

ist medical care for chronic fatigue syndrome (pace):

a randomised trial. Lancet, pages 611–690.

BIODEVICES 2012 - International Conference on Biomedical Electronics and Devices

166