Development of a computerized tool for self-assessment of subjective symptoms that can be integrated in a Personal health record
More than 10.000 people die from cancer every year in Norway. During the last stages of the disease, the aim of the therapy is no longer to prolong the life of the patient but rather to focus on the patient's quality of life. For the last 5 years the palliative care research community in Trondheim and Oslo have developed PAT-C, a computerized tool for assessment of patients subjective symptoms. This tool has earlier been implemented on a PDA platform and recently as a standalone application for PC's. We now aim to further develop the PAT-C application to become an application that adapts the questions to the capabilities of the individual cancer patient and to integrate PAT-C as a component for self assessment and reporting in the patient's personal health record. We here present the outline of this project.
Self-reporting of subjective symptomsPalliative carePersonal health recordsHealth informatics-
1 Introduction
The cancer risk is steadily increasing, and more than 24 000 Norwegians were
diagnosed with cancer in 2004 [
1
]. Among these, about 50% will be cured, and the
rest will be in need of palliative treatment at some stage of their disease trajectory [
1
].
Cancer typically affects the older population, and thus most palliative patients will, in
addition to their cancer, also have other co-morbidities [
2
]. The treatment goal in the
palliative phase of cancer diseases is to relieve symptoms and improve quality of life
[
3, 4
]. Most patients in the palliative phase of their disease experience several
symptoms at the same time, and different symptoms may dominate at different stages
of the palliative phase [
5
]. In general, the health care system does not function
optimally with respect to detecting and treating symptoms in palliative care patients
[
6
].
Despite huge efforts during the last 10-15 years, pain still remains inadequately
treated in a huge proportion of the palliative care population and in spite of many and
increasing number of treatment options [
7
]. There are probably many factors that
contribute to this rather unsatisfactory situation, but the inadequate assessment of
subjective symptoms including pain is probably on of the most important [
8
]. This
may relate to: A. Lack of consensus on how to assess subjective symptoms, B.
Assessment of subjective symptoms has not been integrated into clinical practice, C.
The existing tools are inappropriate for clinical use, D. Health care professionals lack
the necessary skills to use the available tools for symptom assessment, E. The
assessment results are not available where the clinical decisions are taken. The
existing assessment tools also have deficiencies with respect to measurement
capabilities [
9
]. This has consequences for the quality and quantity of clinical research
on symptom therapy. For example, meta-analyses of pain treatment have been
difficult to conduct due to the use of different outcomes, i.e. different pain
questionnaires in the different studies [
10
].
For the last 5 years, the palliative care research communities in Trondheim and Oslo
have been developing PAT-C, a computerized tool for patient’s reporting of pain and
other subjective symptoms, The application has undergone an iterative development.
Currently, the first clinical study utilizing computer-based assessment is now running.
Data from 1000 Norwegian palliative care patients will be collected May-December
2006. During the development of the PAT-C-tool, it has become evident that much
work has yet to be done on the development of optimal software solutions fitting the
palliative population with its special characteristics.
This work was included in a grant proposal from the EU and will receive funding
from late 2006 (the European Palliative Care Research Collaborative (EPCRC)).
2. Further development of an interactive questionnaire for self-reportingThe aims for the planned activities are:
1. To create a computer version for assessment of pain localization
2. To define the optimal characteristics of a Computerized Assessment Tool
for patients receiving palliative care
3. To study the effect of a Computerized Assessment tool on clinicians
decision making process
In the next iteration of PAT-C both architectural considerations, the principles of
user-centred design and item-response theory will be taken into account, this to meet
•
•
•
Personal health records as an emerging tool for patients and The special needs of patients receiving palliative care The need to adjust the questions to each individual patient by use of item response theory (IRT) and possible of artificial intelligence techniques.
We here will describe and discuss general design considerations, and which plans we
have for the further development of PAT-C integrated in the Personal health record
but also as a standalone application.
2.1 Design considerations
During previous iterations of PAT-C, several challenges have been identified, and
solutions have been proposed and implemented. These are summarized in table 1.
In a pilot-study before the official data collection in PAT-C, we assessed the use of
Tablet computers to see how the patients reacted to the software developed and the
use of computers for answering questions regarding their pain and physical
functioning. The response from the 20 patients in the pilot study was as follows
• 9 had not used computers before, 11 had used computers
• 11 persons preferred using computers for answering, 5 persons had no
preference and 3 would prefer using paper and pen (1 missing)
• 19 persons had no problem reading the text (1 missing)
• 16 persons had no problem clicking the radiobuttons for answering. 3 had
problems (1 missing)
• Everybody understood the concept of radiobuttons for answering
• 18 persons had no problem understanding how to use the body map, 1 had
problems (1 missing)
These responses are promising regarding the validity of the PAT-C tool. There are
many considerations for securing the validity, and the user interface is a key
component for that in computer-based testing [
11
]
2.2 Creating a computer version for assessment of pain localization
Clinicians rely on information about pain intensity and localization in their
diagnostics. For clinical as well as scientific purposes information on pain localization
is of great importance in order to understand, classify and treat the pain. Most clinical
studies assess pain intensity but not pain localization. Different types and
localizations of pain are thereby collapsed into the principal outcome in these studies.
The Brief Pain Inventory (BPI) is recommended by the European Association for
Palliative Care (EAPC) as the principal tool for assessment of pain in palliative care
[
10, 12
]. In addition to its 4 items on pain intensity and 7 items on the interference of
pain with functioning, the BPI also includes a body map. This is to be completed by
the respondent in order to assess the localization of their pain (Figure 1). The body
map is of value in the clinics but of very limited value in research due to the
difficulties in quantifying the drawings systematically. Such drawings on a paper
questionnaire also fit poorly with electronic medical records, which imply that the
information is not fully utilized in daily clinical practice. As far as we know, no
systematic attempts to computerize the body map have been performed.
2.2 Develop a model and architecture for integrating a self-assessment tool in thePersonal health records
We will describe a model that enables exchange of information between patients and
professional care providers. Both patients and providers will be able to take the
initiative to start registering symptoms. Both patient and provider will benefit from
having access to longitudinal data. The data will belong to the patient, and he will
decide who shall get access to the data.
2.3 Development of a questionnaire tool that uses item response theory to adapt the questions to the user.
Palliative patients have high level of symptoms, high age, co-morbidities and
decreased performance status [
3
]. This makes the completion of lengthy
questionnaires to a burden that can affect their quality of life. We will use techniques
from advanced psychometrics (item-response theory (IRT)) and artificial intelligence
(case-based reasoning (CBR)) to make the completion of the questionnaire less of a
burden for the patient. We will develop an application that uses these techniques to
minimize the amount of questions asked to the patients, as well as help the clinician in
the decision making process.
3. Conclusive remarks
In this paper we have outlined the work that will be undertaken to make PAT-C a
mature and advanced application for assessment of subjective symptoms in palliative
care.
1. Kreftregisteret:
Cancer in Norway 2004
. Kreftregisteret, Oslo (
2006
)
2.
Extermann
,
M.
:
Measurement and impact of comorbidity in older cancer patients
.
Critical Reviews in Oncology Hematology35
(
2000
)
181
-
200
3.
Kaasa
,
S.
,
Loge
,
J.H.
:
Quality-of-life assessment in palliative care
.
Lancet Oncology3
(
2002
)
175
-
182
4.
Norsk forening for palliativ medisin: Standard for palliasjon elektronisk ressurs
. [Den norske laegeforening], [Oslo] (
2004
)
5.
Jordhoy
,
M.S.
,
Fayers
,
P.
,
Loge
,
J.H.
,
Ahlner-Elmqvist
,
M.
,
Kaasa
,
S.
:
Quality of life in palliative cancer care: Results from a cluster randomized trial
.
Journal of Clinical Oncology19
(
2001
)
3884
-
3894
6.
Patrick
,
D.L.
,
Ferketich
,
S.L.
,
Frame
,
P.S.
,
Harris
,
J.J.
,
Hendricks
,
C.B.
,
Levin
,
B.
,
Link
,
M.P.
,
Lustig
,
C.
,
McLaughlin
,
J.
,
Reid
,
L.D.
,
Turrisi
,
A.T.
, 3rd,
Unutzer
,
J.
,
Vernon
,
S.W.
:
National Institutes of Health State-of-the-Science Conference Statement: Symptom management in cancer: pain, depression, and fatigue
,
July 15- 17
,
2002
.
JNatl Cancer Inst Monogr
(
2004
)
9
-
16
7.
Meuser
,
T.
,
Pietruck
,
C.
,
Radbruch
,
L.
,
Stute
,
P.
,
Lehmann
,
K.A.
,
Grond
,
S.
:
Symptoms during cancer pain treatment following WHO-guidelines: a longitudinal follow-up study of symptom prevalence, severity and etiology
.
Pain93
(
2001
)
247
-
257
8.
Vonroenn
,
J.H.
,
Cleeland
,
C.S.
,
Gonin
,
R.
,
Hatfield
,
A.K.
,
Pandya
,
K.J.
:
Physician Attitudes and Practice in Cancer Pain Management - a Survey from the EasternCooperative-Oncology-Group
.
Annals of Internal Medicine119
(
1993
)
121
-
126
9.
Kaasa
,
S.
,
Loge
,
J.H.
:
Quality of life in palliative care: principles and practice
.
Palliative Medicine17
(
2003
)
11
-
20
10.
Caraceni
,
A.
,
Cherny
,
N.
,
Fainsinger
,
R.
,
Kaasa
,
S.
,
Poulain
,
P.
,
Radbruch
,
L.
,
De Conno
,
F.
,
Network
,
E.R.
, :
Pain measurement tools and methods in clinical research in palliative care: Recommendations of an Expert Working Group of the European Association of Palliative Care
.
Journal of Pain and Symptom Management23
(
2002
)
239
-
255
11.
Booth
,
J.:The key to valid computer-based testing - the user interface
.
EUR REV APPL PSYCHOL41
(
1991
)
281
-
293
12.
Cleeland
,
C.
:
Pain assessment in cancer
. In: Osoba,
D
. (ed.):
Effect of cancer on quality of life
. CRC Press, Boca Raton (
1991
)
293
-
305