International
Journalof
ELSEVIER
International
Journal
of Bio-Medical
Computing
42 (1996)
35-41
IMedical
Computing
The core of computer based patient records in family practice:
episodes of care classified with ICPC
Henk Lamberts”,
Inge Hofmans-Okkes
Abstract
A central element in the definition of primary care is that primary care clinicians address the large majority of
personal health care needs of their patients. As a consequence. they should document data on these health care needs
reliably and continuously.
To establish whether this occurs, the episode of care is the most appropriate
unit of
assessment: a health problem from its first encounter with a health care provider until the completion of the last
encounter for it. An episode of care is distinguished from episodes of disease and of illness. The episode of care as
an epidemiological
concept for the calculation of rates has evolved into a central element of a computer based record.
Episode oriented data classified with the International
Classification of Primary Care (ICPC). and specified with
ICD-10 as a nomenclature are especially suitable as the core of a generic patient record in family practice. ICPC has
been available to the family medicine community for well over a decade as the main ordering principle of its domain.
The basic structure of an encounter (within the string of encounters which together form an episode of care)
distinguishes reasons for encounter, diagnoses and diagnostic and therapeutic interventions. In this article, a more
refined structure of encounters is proposed for a more precise documentation
of episodes of care in a computer based
patient record. The conversion structure between ICPC and ICD-10 allows both a high level of specificity in the
patient’s problem list and optimal communication
with specialists who contribute to the episodes of care for which
the documentation
is the primary care provider’s responsibility.
Keywwdss:
Episode of care: Computer
based record: International
Classification
of Primary
Care (ICPC):
ICD-10;
Classification
1. Introduction
* Corresponding
20 6918806.
0020-71Olj96/$15.00
PII SOO20-7101(96)01
author.
Tel.:
+ 31 20 566471 I; fax
0 1996 Elsevier
179-S
Science
Ireland
Ltd.
+ 31
All rights
In many countries
a major shift in the interest
for the position of primary care in the health care
system can be observed, for both economical and
quality of care reasons. More rational health care
reserved
Fig.
I. An episode
may lead to a reduction of cost and interventions,
and to better health care. This is reflected in the
new Institute of Medicine (IOM)-definition:
‘Primary care is the provision of integrated accessible
health care services by clinicians who are accountable for addressing a large majority of personal
health care needs, developing a sustained partnership with patients and practising in the context of
family and community’ [l]. Primary care clinicians
(in Europe practically always family physicians,
but in the United States also general internists
and paediatricians)
are responsible for patient
documentation,
the rules for which have been
spelled out in another important IOM-publication
121.
The ‘episode of care’ is the most appropriate
unit for assessing to what degree clinicians indeed
provide primary care for the large majority of
personal health care needs. The episode of care is
defined as ‘a health problem or disease from its
first presentation to a health care provider until
the completion of the last encounter for it’ (Fig. 1)
[3,4]. Reasons for encounter reflect the patient’s
perspective, while the diagnosis forms the medical
interpretation;
the therapeutic and diagnostic interventions relate to both perspectives [5]. These
three elements are essential in all encounters
of care.
within the string of encounters which together
form an episode of care, and their careful documentation allows the observation of changes in
their relations over time (transitions).
An episode of care refers to all care provided
for a patient with a discrete disease or health
problem, including the contributions
of specialists
and hospital admissions. On this basis the large
majority of personal health care needs, the degree
of integration, accessibility and accountability can
be documented in a generic computer based patient record.
The episode of care is clearly distinguished
from an episode of disease and an episode of
illness [6-81. An episode of disease is a disease
from its onset until its resolution or the patient’s
death, while an episode of illness is the period that
an individual suffers from symptoms
or complaints that he/she experiences as an illness. Not
every disease and certainly not every illness results
in an episode of care. Most episodes of care,
however, are part of an episode of disease or
illness. In this framework,
health maintenance
episodes can be considered a special form of
episodes of care.
During the past decades, indicators for these
different types of episodes have been derived from
H. Lamherts.
I. Hqflnuns-Okkrs
/ Internatimal
patient or population data with a limited time
frame, e.g. 2 weeks. 1 month, 1 year. However,
the documentation
of episodes of care in a computer based patient record should not be limited
to a specific time window:
it should in principle
contain the documentation
of events during the
complete course over time of the episode.
Seven stages can be distinguished
within
episodes (Fig. 2). Before an episode of disease is
actually established, there may be a period in
which the patient does not yet experience symptoms or complaints, but in which the diagnosis
might be established, e.g. through screening (part
1, Fig. 2). The actual start of an episode of a
disease occurs in part 2, where patients experience
symptoms or complaints without, however, seeking care. The episode of care starts in part 3. Once
an episode of care has been established it will be
included in the typical epidemiological indicators
such as 1 year incidences (part 3) and prevalences
(parts 3 and 4). However, it may also occur that a
patient has an episode of a disease without seeking care for it during 1 year (tO&tl) but does so
later (tlLt2, part 5). Finally, patients with a disease may not seek further care (part 6) or never
do so at all (part 7). The prevalence of an episode
of care in a certain period (parts 3 and 4, or parts
3, 4 and 5) is, by definition, smaller than the
prevalence of a disease in the population (parts
l-7). For important diseases both indicators will
be rather similar, especially in a large time window (fractures,
strokes, metastatic malignancy,
blindness). More often, however. there will remain
considerable differences between these indicators.
This is the result of structural differences between
data from population studies (point prevalences
of disease), from health interviews
(illness plus
disease in a limited time window),
and from uti-
Fig. 2. Episode of care and episode of disease.
Journal
of’ Bionwdiml
Computing
42 (1996)
35-41
31
lization studies (care in 1 year). In the USA the
National Ambulatory
Medical Care Survey
(NAMCS), the National Health Interview Survey
(NHIS), and National Medical Expenditure Survey (NEMS) are the major sources for these data
[9- 111.In the UK, the Fourth Morbidity Survey
in General Practice plays an important role, and
in the Netherlands a long standing history of
routine registration of morbidity in family practice exists, which has proven to be instrumental
for national policy making [4,12,13].
The main goal of this contribution is to illustrate how the original description of encounter
based episodes of care, structured with the International Classification of Primary Care (ICPC):
has evolved into a more detailed encounter structure which can serve as the core for a computer
based patient record (Fig. 3) [3].
2. A new encounter structure as the core for a
computer based patient record
Five elements (a t/m e) are essential for the
encounter/episode structure in primary care:
(a) The patient’s reason for encounter has been
established as a reliable and practical source
of information, because it serves explicit patient orientation and it explains variation in
interventions within the same episode type to
an important degree [5]. Reasons for encounter in the form of symptoms and complaints
play a major role in estimating prior and
posterior probabilities for certain diagnosesin
standard sex/age groups. ICPC incorporates
over 200 symptoms and complaints, and the
use of these in the reason for encounter mode
greatly enhances the clinical relevance and
reliability of patient documentation. In the
Transition project of the Department of Family Medicine, University of Amsterdam, a
large database has been collected with well
over 90 000 patient years and 230 000 episodes
of care which serves as an epidemiological
expert system for the computer record [13].
(b) Reasons for encounter in the form of requests
for interventions such as a prescription, an
X-ray, a referral. advice. etc. play an even
38
H.
Lnmhevts,
1. Hofmms-Okkes
: International
Journal
Fig. 3. The old and the new structure
more important role in explaining variation in
the distribution
of interventions within a certain episode type. Patients actively influence the
content of primary care provided by family
physicians, and it is by now well documented
that, with few exceptions, providers are inclined to reward their patients’ requests. Evidently, a cautious description of this type of
reasons for encounter is important for a better
insight in patient’s compliance (or lack of it) on
the one hand, and physician’s
adherence to
clinical guidelines (or lack of it) on the other.
(c) A patient will not always express all symptoms
and complaints
as a reason for encounter.
During physical exam and history taking, the
physician will ask about the wide spectrum of
symptomatology
included in ICPC symptoms
and complaints. Together with those expressed
as reasons for encounter they allow the calculation of more reliable probabilities, while the
difference between a symptom expressed by the
patient and those ‘found out’ by the physician
is retained. Again this stresses the importance
of explicit patient orientation
in computer
based patient records, leaving the physicians
clinical interpretation
intact.
(d) ICPC allows a straight forward
diagnostic
qf Biomrdicd
for descibing
Computing
42 (I 996)
3.5
41
encounters.
label for approximately
400 episodes with a
prevalence of at least one per thousand enrolled patients per year in primary care. The
available 670 diagnostic classes in ICPC allow the structure
of the episode oriented
database for retrieval purposes, for the use of
expert systems and drug formularies and for
the introduction
of clinical guidelines, but evidently it does not cater for the need for
diagnostic specificity on the individual level.
The complete conversion between ICPC and
ICD-10 is a major step forward
in this respect, because it allows to add the state-ofthe-art medical nomenclature,
for detailed
description
of clinical problems [I 4,151. But
even the three or four digit ICD-10 labels do
not always contain sufficient clinical detail
for an individual patient. This is illustrated
by data from a patient with a cholecystitis
treated by a laparoscopic
intervention,
followed by peritonitis (Table 1). The conversion structure between ICPC and ICD-10 for
a relatively homogeneous class as cholelithiasis/cholecystitis
is evident, but for a seldom
occurring disorder as peritonitis ICPC has to
use a ragbag (D99), while ICD-10 allows the
use of a more specific class (Tables 2 and 3).
Table 1
List of episodes
of care of a patient
(underlined
is the patient
specific
ICPC
[CD-IO
Text
D99
D98
x31
s79
R75
W78
S82
k65
k80
27 I. I
d23
jO1
232.1
d22
Peritonitis
after laparosc.ectomie
Cholelithiasis’pigeon
eggjectomie
Cervical smear/yearly/request
Fibroma
pendulum
left shoulder/lcm
Acute sinusitis maxiI/&
Pregnancy:lwins
Melanocytic
naevua/Rcheek:3mm
with
natient
specific additions
However, in the physician’s problem list of
this patient, more individual additions are
necessary. The ICPC-code serves to structure
the episode oriented database, while the
ICD-10 nomenclature allows more specificity, also for the purpose of communicating
with specialists and hospitals, while the final
text in the problem list illustrates the specificity which is needed on the level dealing with
a specific patient’s problem.
The availability of diagnostic criteria helps
Table 2
Conversion
from
ICPC
to ICD-IO
ICPC
[CD-IO
Text
D98
-k80
-k8l
-k82
-k83
-k87
Cholelithiasis
Cholecystitis
Other diseases of gallbladder
Other diseases of biliary tract
Dis.gallbl.:bil.tr./pancr.
in dis.C.E.
D99
-k38
-k52.R
-k55
-k56
-k59.R
-k62.R
-k63
-k65
-k66
-k67
-k85
-k86
-k87
-k90
-k92.R
-k93
Other diseases of appendix
Other noninf.gastroenter./colitis
Vascular disorders
of intestine
Paralytic
ileus and intest.obstr.w/o
her
Other functional
intestinal
disorders
Other diseases of anus and rectum
Other diseases of intestine
Peritonitis
Other disorders of peritoneum
Disorder
of peritoneum
in inf.dis.c.e.
Acute pancreatitis
Other diseases of pancreas
Disor.gallbl./bil.tr.lpancr.in
dis.c.e.
Intestinal
malabsorption
Other diseases of digestive system
Disorders
of oth.dig.org.
in disce.
additional
text)
First
encounter
23-03-95
12-03-95
25-02-89
08-12-93
22-04-93
06-l 2-92
3 l-08-92
Last encounter
24-05-95
19-03-95
12-07-95
09-O 1-94
12-05-93
15-07-93
06-09-92
the physician to decide about the certainty of
the diagnosis [15,16]. It is essential that for
each encounter the ‘status’ of the episode is
indicated: whether it is a new episode, or the
follow-up of an ‘old’ one. During follow-up
the diagnostic label of the episode can be
modified in case more diagnostic or other
information has become available.
(e,f) It is important to distinguish diagnostic and
therapeutic
interventions
occurring
during
the encounter (‘intermediate’)
from those
that will follow (‘resulting’).
These data are
important for the analysis of utilization, of
interdoctor variation. of compliance and of
the extent to which clinical guidelines are
followed. Also, this information
is helpful
for establishing the shift from prior probabilities during the first encounter in an
episode of care to posterior
probabilities
during follow-up.
‘Intermediate’
interventions did in fact occur, whereas resulting
interventions such as prescriptions:
referrals,
and additional examinations do not always
take place because, sometimes, a patient will
be ‘non-compliant’.
ICPC provides sufficient
specificity to allow the coding of intermediate interventions
in most countries; for
the classification
of resulting interventions,
however,
often a more detailed system is
necessary [17]. The ICPC drug coding system, for example, is based on an ATC-compatible conversion structure for the coding
of drugs within the structure of ICPC. So
far, no international
standards for coding
laboratory
tests etc. in primary
care are
available.
40
Table 3
Two examples
H. Lumherts,
of the eight
I. Hofinans-Okkrs
European
language-layer
/ International
of ICPC
Journal
with
of'
ICD-IO
Biomedical
Computing
Cholecystitis/cholelithiasis
Galdesten/galdebherebetamdelse
Cholecystitis/cholelithiasis
CholCcystite!cholelithiase
Cholezystitis/cholelithiasis
Colecistite/colelitiasi
Colecistite:‘colelitiase
Colecistitis:colelithiasis
D99
-k38-k52R-k55-k56-k59Rk62R-k63-k65-k66-k67k85-k86-k87-k90-k92R-k93
Other diseases digestive system
Anden sygdom mave/tarmsystemet
Andere ziekten spijsv*erteringsorganen
Autres maladies syst digestif
Sonstige Erkrankung
Verdauungstrakt
Altre malattie
sistema digerente
Outras doencas aparelho
digestivo
Otras enf ap digestivo
The domain of primary care has. during the
past decade, been described in detail. This has
lead to the availability of several international
information
tools: classification systems, conversion structures,
a glossary for primary care terminology
and conventions
for standard
data
presentation, all serving as building stones for a
computer based patient record [ 181.
The availability
of ICPC in 20 languages
greatly enhances the establishment of an international minimal basic data set for computer based
patient records in primary care. The core of such
a record, however, should preferably be language
independent because this enhances its international use.
The episode of care as a central concept has
been widely accepted [19]. It is by now also better understood that classification systems do not
primarily provide names for diseases, but, rather.
provide a structure to order the objects that are
generic to a certain domain on the basis of well
established and stable criteria. A nomenclature
changes considerably
over time and it needs
many language, culture and country specific inclusions: in essence it represents the professional
35 41
conversion
D98
-kSO-kSl-k82-k83-k87
3. Discussion
42 (1996)
jargon in medicine and its thesaurus (index) as a
storehouse of knowledge, much like an encyclopedia. ICPC is the preferred ordering principle
for the primary care domain, allowing access to
specific expert systems and clinical information.
ICD-10 and its index help to address global expert systems and databases. The proposed new
standardized
structure of encounters as part of
episodes of care in primary care as they develop
over time, indeed offers a core for a good computer based patient record.
References
[I]
Donaldson
M. Yordy
K, Vanselow
N (Eds): Dr$nirz~
PrimurJ
Cure. Apz lntrrim
Report. Washington
DC: National Academy
Press, 1994.
[2] Dick RS. Steen EB (Eds): The Computer-Based
Patient
Record. iln essrntial tecknology
for I~rcrlih c’are, Washington DC: National
Academy
Press, 1991.
[3] Lamberts
H and Wood M (Eds): lnternarionul
C/uss$cation c!f Primwy
Care (ICPC),
Oxford:
Oxford
University
Press, 1989.
[4] Lamberts
H, Wood
M and Hofmans-Okkes
IM, Eds.:
Internutiotzal
Cluss&ztion
of Putnary
Cure in the Europecl~ Cornmunir~.
Oxford:
Oxford
University
Press,
1993, pp. 45 61.
H. tamhert.~
1. Hofrnans-Okkes
/ International
[5] Hofmans-Okkes
IM: An international
study into the concept and validity
of the ‘reason for encounter’.
In The
Internutianal
Clas.s~yjcation qf’Pr~mar~~ Care in the European
Community
(Eds: H Lamberts.
M Wood and IM HofmansOkkes),
Oxford:
Oxford
University
Press, 1993.
[6] Greenlick
MR: Episodes ofCare Classification.
In: Pr~ncrv.~
Care Research: Theor)’ and Methods (Ed: ML Grady).
US
Department
of Health and Human Services: Public Health
Serbice, 1991.
[7] Hornbrook
MC, Hurtado
RV and Johnson RE: Healthcare
episodes. Definition,
measurement
and use, Med Cure Ret,
42 (1985) 163-218.
[S] Solon JA, Feeney JJ. Jones SH, Rigg RD and Sheps CG:
Delineating
episodes of medical care, Am J Public Health,
57 (1967) 401-408.
[9] Schappert
SM: National
Ambulatory
Care Survey,
1991
Summary.
National
Center
for Health
Statistics,
Vital
Health Stat, I3 (1994) I 16.
[IO] Adams
PF and Benson V: Current
estimates
from the
National
Health Interview
Survey. 1991. National
Center
for Health Statistics,
Vital Health Stat. IO (1992) 184.
[I I] Cornelius
L, Beauregard
K and Cohen J: L~LIU/
.s~urces
c!f
Medical Care and their Characteristics:
jliutiona(
.Medicul
E,~penditurr
Suruey Rcwurch
Findings II, Rockville.
MD.
Public Health Service. 1991, Agency for Health Care Policy
and Research publication.
91-0042.
[I21 McCormick
A, Fleming
D and Charlton
J: Morbidit~~
Journal
[I31
[ 141
[I 51
[I61
[17]
[I81
[I91
of Biomedical
Computing
42 (1996)
35 -41
41
Statistics ,fionl General Practiw.
Fourth
National
Study
1991~1992.London.HMS0,0PCS.seriesMB5no.3,1995.
Lamberts
H, Oskam
SK. Hofmans-Okkes
IM, et al.:
Episodeoriented
data from the Transition
project
on
diskette,
Huisarts
Wet, 37 (1994) 421 426.
WHO: IntentationalStcrtisfi~al
Cl&&ation
ofDisrcisesand
Related
He&h
Problems
(ICD- IO). Tenth
Revision,
Geneva, WHO,
1992.
WONCA
Classification
Committee:
Internrrtionul
Class~Jfcation qf Health Problems in Primary
Care (ICHPPC-2Defined), Oxford:
Oxford
University
Press. 1983.
Hofmans-Okkes
IM and Lamberts
H: The classification
of
psychological
and social problems in general practice. The
impact of the patient’s perspective,
Huisctrts M/et, 36 Suppl
(1993) 14-20.
Shepherd
S and Hofmans-Okkes
IM: The multilanguage
layer of ICPC. In Interncltional
Classific,ation
of’ Primary
Care in the European Community,
(Eds: Lamberts
H, Wood
M and Hofmans-Okkes
IM), Oxford:
Oxford
University
Press. 1993, pp. l73- 219.
Bentzen N, Ed.: WONCA
Classification
Committee:
An
international
glossary
for general/family
practice,
Fczm
Pratt. I2 (1995) 341 369.
Donaldson
MS, Yordy KD, Lohr KN, Vanselow NE, Eds.
Report ofthe Institute of Medicine. Primary Crrre:,4nwric~a’s
Health ma Nr~c Era, National Academy
Press, Washington
DC. 1996.