Content uploaded by V.aleria V.ictorovna Gribova
Author content
All content in this area was uploaded by V.aleria V.ictorovna Gribova on Feb 20, 2019
Content may be subject to copyright.
Medical diagnosis ontology for intelligent decision
support systems
Valeriya Gribova1, Dmitry Okun1, Мargaret Petryaeva1, Elena Shalfeeva1
1 Institute for Automation & Control Processes, FEBRUS, Russia
The amount of knowledge in medicine is growing in an avalanche. For making
diagnostic decisions a practical physician needs to take into account a number of
factors: symptoms and syndromes of the disease, its nosological forms, etiologies,
pathogenesis, clinical manifestations, taking into account the individual
characteristics of patients. Differentiation of similar diseases often requires labor-
intensive and time-intensive analysis or studies; precise diagnosis requires
specification of the diagnosis according to the accepted classifier, there is always
incompleteness of information about all the features of an individual patient. In doing
so, time for acceptance of the decision by doctor to make a decision is not increased.
As a result, the number of medical errors is increasing, which, according to the
estimates of literary sources, reaches up to 30% in some countries [1]. The quality of
medical care depends not only on the level of training (competence) of medical
personnel, but also on systems that provide support for decisions of the doctor [2].
For the period from the first works on the creation of intellectual systems (IS) for
medical diagnostics to the present time a huge spectrum of such systems has been
implemented, but they are practically not introduced into the daily practice of a
doctor.
Common causes of problems with implementation of IS are huge labour required
for development of knowledge bases that are close to the real medical practice. It is
due to the inability to assign this task to highly qualified experts without knowledge
engineers as intermediaries.
In addition, those IS, which task is to make a diagnosis instead of a doctor (or to
offer him a list of possible diagnoses organized by their probability), are not required
by experienced experts if they solve the problem no better than the expert himself.
Those IS that support decisions of the doctor in a difficult situation "at the level of
council of physicians" are rare and the decision support systems usually do not take
into account many of the signs and factors identified from the anamnesis and other
sections of the case record with personal information.
The solution of these problems is the creation of specialized ontology-oriented
shells for the development of differential diagnosis systems for diagnosing diseases at
various stages of their development. In such shells, the model for representation of
information (knowledge and data) should be expert-oriented and understandable by
experts, be corresponding to a wide range of diseases (rather than a specific medical
profile).
The aim of the paper is to describe the ontology for differential diagnosis of
diseases, taking into account the dynamics of their development for the development
of support systems for making diagnostic decisions in medicine.
2
1 Ontology requirements
One of the first ontologies of medical diagnostics close to the real notions of medicine
and matching a wide range of diseases was the "Ontology of medical diagnostics of
acute diseases" [3, 4]. It describes the clinical picture of diseases in the dynamics of
the pathological process (over time), as well as impacts of therapeutic measures and
other events on manifestations of diseases.
Using this ontology the knowledge bases of diseases of some body systems have
been developed: respiratory organs (bronchial asthma, pneumonia), digestive organs
(peptic ulcer, acute appendicitis, acute and chronic pancreatitis, acute and chronic
colitis), eyesight (conjunctivitis, keratitis, glaucoma), etc. [5], and also have been
implemented a some of software services.
The experience of more than ten years of using the ontology for the formation of
knowledge bases on the diagnosis of a number of diseases has made it possible to
accumulate and reveal a number of limitations: impossibility of describing the clinical
manifestations of disease for different groups of patients, impossibility of describing
alternative diagnostics and taking into account in the final diagnosis the forms,
variants of disease and its severity level, spectrum of modality values of
characteristics.
Modern systems should be able to diagnose and conduct differential diagnosis with
other diseases at different periods of disease development, analyzing disease
development in the period before visit to doctor and considering that patient can come
to reception at different times from the onset of disease, as in the first hours, so and at
a moment when semiology is quenching.
The ontology of medical diagnostics should have the following basic
characteristics.
1. The possibility of forming symptom-complex diseases taking into account the
categories of users using reference ranges instead of certain "norms" for laboratory
and instrumental indicators [6].
2. The possibility of forming alternative symptom complexes with different
approaches to identifying reliable signs of the disease in order to choose the most
sparing, fast or inexpensive diagnostic process.
3. Ability to clarify the diagnosis, taking into account the etiology, pathogenesis,
variant of the flow, etc. for differential diagnosis of diseases and the selection of
appropriate methods of treatment.
4. Uniform formalization of stages of chronic diseases and periods of development
of acute diseases.
5. Expansion of a number of values of modality, previously represented by
"obligation" and "possibility," the case of "specificity" (for a specific symptom).
6. Taking into account the values of characteristics and characteristics, modified by
the impact of events. The presence of such an element of cause-effect relationships
allows us to take into account external influences exerted on the patient's organism at
different stages of the disease.
3
7. Accounting for different variants of the dynamics of the values of characteristics.
Such elements of knowledge allow to take into account (on the basis of medical
experience) the diversity of the course of the same diseases in different patients [3, 7].
All these types of connections between the concepts of medical diagnostics make
it possible to form a knowledge base sufficient not only to search for hypotheses
about possible diagnoses, but also to reduce the set of hypotheses about the diagnosis.
The latter is solved by searching for a request for such additional information about
the patient's condition (using the results of observations of the situation and
knowledge of internal processes and the influence of external events on them), the
answer to which will refute previously assumed diagnoses (identifying among the
known measurable signs of differentiators or "partial" [8]). It is possible to diagnose
the combined and complicated pathologies.
2 Basic characteristics of the ontology
The form of the representation of ontologies in general and medical ones in particular
(hierarchical semantic networks) used by the team (authors) and the applied IACPaaS
tools [9] allow to effectively modernize existing ontologies. New versions
(modernized ontologies) are taken as a basis for designing new versions of shells for
IS developers.
Each disease is represented by alternative symptom-complexes (complexes of
complaints and objective studies, laboratory and instrumental studies), necessary
conditions for this disease and can contain details of the corresponding diagnosis in
form, variant, severity, stage, etc. Symptomocomplexes in different diseases can be
different amounts: depending on the course of the disease in different age groups of
patients. A necessary condition for the disease is that event, without which the disease
would not have happened. In complexes of complaints, objective studies, laboratory
and instrumental studies, many features are presented, the changes in the values of
which are symptoms of the disease. Possible causes of the disease are events or
etiological factors that led or contributed to the development of the disease. They are
described by modality and temporal characteristics (interval before the onset of the
disease, duration of the event, etc.). The detailed diagnosis is an additional set of
symptoms (symptom complex), which allows to make an appropriate refinement to
the main diagnosis, taking into account the etiology, pathogenesis, variant of the
course, stage, etc. for more detailed (deep) diagnosis or differential diagnosis of the
disease (see Fig. 1).
In other words, relations (proposals) are introduced into the knowledge model:
• variant of the symptom complex for some disease;
• variant of the process of changing the values of the characteristic, characteristic
of some symptom complex;
• special conditions necessary for the onset of a disease;
• a variant of reaction to the impact of the event;
• variant of reaction to the influence of a combination of factors
and some others.
4
Taking into account the dynamics of the values of the signs allows describing the
diseases taking into account one of the main complexities of the diagnostic process in
medicine - the need to determine a continuously developing process (disease). Each
disease develops for more or less time. From the point of view of the speed of the
development of diseases, the sharpest - up to 4 days, acute - about 5-14 days, subacute
- 15-40 days and chronic, lasting months and years are distinguished. In the
development of the disease, it is almost always possible to distinguish the following
stages: 1) the onset of the disease (sometimes called the latent period); 2) the stage of
the disease itself; 3) the outcome of the disease. Diagnosis, as a rule, is carried out at
the stage of "actual disease". At this stage, the following periods of development are
distinguished: 1) the period of the increase in manifestations of the disease; 2) peak
period (maximum severity of symptoms); 3) the period of extinction of manifestations
of the disease (the gradual disappearance of clinical symptoms).
The symptom of the disease can be simple or compound, its values are presented
by the period of the dynamics of the development of the trait or the disease as a
whole; the necessary conditions for the consideration of the characteristic can be set.
Each period of dynamics is characterized by an upper and lower limit of the duration
of the period, a unit of measurement of boundaries. A simple symptom is modality
(the entry of a symptom into a clinical picture); in each period the characteristic can
have more than one variant of values. Each variant of the values specifies the set of
possible values of the characteristic, the necessary conditions for the presence of the
characteristic, and a description of the change in the value of this characteristic under
the influence of certain events. A composite feature contains a description of the sets
of its own time-varying characteristics with the modality of the entry of characteristic
characteristics into the clinical picture. The characteristic may have a number of
intrinsic characteristics and a plurality of variants of the value of this characteristic.
5
Fig. 1. The screenshot of the fragment of the ontology of knowledge about diagnosis of
diseases on the IACPaaS platform.
Each variant of the characteristic values contains a lot of its possible values and a
description of the change in the value of this characteristic under the influence of
certain events. The term "value modified by the impact of the event" allows
describing the change in the sign in the dynamics if, after the onset of the disease's
development, before the patient has consulted the doctor, the patient himself took any
measures, or the meanings of the symptoms (complaints, objective state) change
under the influence of some events or manipulations undertaken by the doctor.
6
3 Application of the formed ontology
Based on the described ontology (using the "Database of Medical Terminology and
Observations" IACPaaS), IACPaaS information resource "Knowledge Base on
Diagnosis of Diseases" was formed, representing a variety of diseases.
The group of diseases "Diseases of the digestive system" consists of groups of
diseases - "Diseases of the esophagus", "Diseases of the stomach and duodenum",
"Bowel diseases", "Diseases of the liver", "Diseases of the gallbladder, biliary tract",
etc. Each group includes diseases with a common set of diagnostic features, which
contains those characteristics that are characteristic of this group. So for the group
"Diseases of the digestive organs" in the complex of diagnostic signs includes one
sign "Pain in the stomach". For the group "Diseases of the gallbladder and bile ducts,"
the diagnostic complex includes 5 signs: "Pain in the abdomen", "Nausea", "Skin
itching", "Body temperature rise", "Stress of the muscles of the anterior abdominal
wall". For the "Cholecystitis" group of diseases, more than 20 signs are already
diagnostic: all the same signs plus "Meteorism", "Vomiting" "Burp", "Nausea",
"Bitterness in the mouth", "Leukocytes", "ESR", "Wall thickness gallbladder on
ultrasound ", etc. The group of diseases "Cholecystitis" according to ICD 10
(International Classification of Diseases), includes the disease "Acute cholecystitis."
The description of the disease includes a description of the symptom complexes for
several age groups of patients (children 0-1 year, children 1-16 years, adults 17-59,
etc.) and the necessary condition for the onset of this disease.
On the basis of the formed ontology, versions of the ontology of the document
containing an explanation of the hypotheses about the diseases corresponding to the
analyzed medical history of the patient are proposed and tested, and recommendations
on the required observations to reduce the set of hypotheses.
4 Conclusion
The publication of ontologies of subject domains representing conceptualizations
close to those used in practice is of independent interest, since they can be used in the
development of a variety of software systems in the relevant areas without the need to
re-conduct a complex analysis of the subject area in each new project. The ontology
proposed in the present work enables the development of software components of an
ontology-oriented shell for the development of differential diagnosis systems for
diseases.
This ontology allows defining and forming knowledge bases for a medical
knowledge portal, for designing differential diagnosis systems for diseases of any
medical profile with the help of such a shell, or for developing systems for supporting
diagnostic solutions using another technology.
The ontology of knowledge about differential diagnostics of diseases is placed on
the platform and is already used by specialists to create knowledge bases in various
fields of medicine. Experts interested in the accumulation, improvement and
7
application of knowledge about diagnostics have the opportunity to join this process
(both for own bases and for collectively used resources).
The work is carried out with the partial financial support of the RFBR (projects
nos. 18-07-01079, 17-07-00956).
References
1. Galanova, G.: Medical errors – a problem not only the doctor. Health care Manager 8, 49–
52 (2014). [in Russian]
2. Artificial intelligence for Medicine – the review, http://www.livemd.ru/tags/
iskusstvennyj_intellekt, last accessed 2018/03/20. [in Russian]
3. Kleschev, A., Chernyakhovskaya, M., Moskalenko, Ph.: Model of ontology of subject
domain "Medical diagnostics". Part 1. Informal description and definition of basic terms.
NTI Series 2 Journal 12, 1–7 (2005). [in Russian]
4. Chernyakhovskaya, M., Moskalenko, Ph.: Formal description of a disease Chronic
pancreatitis. Informatics and control systems 4, 97–106 (2012). [in Russian]
5. Petryaeva, M., Okun, D.: Stomach ulcer in formal representation. In: XLVII of the
International scientific and practical Conference "A scientific discussion: medicine
questions", vol. 3, pp. 62–66. Internauk, Moscow (2016).
6. Kazakova, M., Lugovskaya, S., Dolgov, V.: Referensnye's values of indicators of the
general blood test of the adult working population. Clinical laboratory diagnostics 6
(2012), https://cyberleninka.ru/article/n/referensnye-znacheniya-pokazateley-obschego-
analiza-krovi-vzroslogo-rabotayuschego-naseleniya, last accessed 2018/03/20. [in Russian]
7. Zhmudyak, M., Povalikhin, A., Strebukov, A., Zhmudyak, A., Ustinov G.: The automated
system of medical diagnosis of diseases taking into account their dynamics. The
Polzunovsky Bulletin 1, 95–106 (2006). [in Russian]
8. Gribova, V., Kleschev, A., Shalfeeva, E.: Method of the solution of a problem of request of
additional information. Design ontology. 3(25), 310–322 (2017). [in Russian]
9. Gribova, V., Kleschev, A., Moskalenko, Ph., Timchenko, V., Fedorischev, L., Shalfeeva,
E.:The IACPaaS Cloud Platform: Features and Perspectives. In Second Russia and Pacific
Conference on Computer Technology and Applications (RPC), vol. 3, pp. 80–84. IEEE
(2017).