Diagnosis and Management of Diabetes Mellitus
through a Knowledge-Based System
Morium Akter1, Mohammad Shorif Uddin2 and Aminul Haque3
1
Dept. of Computer Science and Engineering, Jahangirnagar University, Savar, Dhaka, Bangladesh
2
Imaging Informatics Division, Bioinformatics Institute, A*STAR, Singapore
3
Bangladesh Institute of Health Sciences (BIHS) Hospital, Savar Center, Dhaka, Bangladesh
Abstract — The chronic disease diabetes mellitus is increas-
ing worldwide. It is caused by absolute or relative deficiency of
insulin, and is associated with a range of severe complications
including renal and cardiovascular disease as well as blindness.
Preventive care helps in controlling the severity of this disease;
however, preventive measures require correct educational
awareness and routine health checks. Medical doctors help in
effective diagnosis as well as treatment of diabetes, although it
is associated with high costs. With this view, the purpose of
the present research is to develop a low-cost automated knowl-
edge-based system that helps in self-diagnosis and manage-
ment of this chronic disease for patients as well as doctors. Our
knowledge-based system has an easy computer interface,
which performs the diagnostic tasks using rules acquired from
medical doctors on the basis of patient data. At present our
system consists of 26 rules and it has been implemented using
Prolog programming. Some real-life experimentations were
performed, which confirmed the effectiveness of the developed
system.
Keywords — Diabetes mellitus, insulin deficiency, disease
diagnosis, health care management, knowledge-based (expert)
system.
I. INTRODUCTION
In the year 2000, a study [1] among 191 World Health
Organization (WHO) member states confirms that 2.8%
people for all age groups had diabetes, and this is expected
to be 4.4% by the year 2030. This implies that the total
number of people with diabetes is projected to rise from 171
million in 2000 to 366 million in 2030. In Bangladesh, dia-
betes is reaching epidemic proportions; in some sectors of
our society more than 10% of people have diabetes [2].
Diabetes causes [3] severe life threatening complications,
such as hypoglycemic coma, blurred vision, loss of mem-
ory, severe impairment of renal function, insulin allergy,
acute neuropathy, etc. Diabetes management requires die-
tary control, physical exercise and insulin administration.
Medical doctors help in effective diagnosis as well as treat-
ment of diabetes. However, it obviously associates with
high costs. Despite remarkable medical advances, patient
self-management remains the cornerstone of diabetic treat-
Chwee Teck Lim, James C.H. Goh (Eds.): ICBME 2008, Proceedings 23, pp. 1000–1003, 2009
www.springerlink.com
ment. Knowledge-based intelligent system has been proven
effective in solving many real-world problems requiring
expert skills. Hence, to reduce the cost and to improve the
early detection as well as self-awareness of diabetes melli-
tus, automated expert system might be a promising solution.
Knowledge-based system for diagnosis can be used in vari-
ety of domains: plant disease diagnosis, credit evaluation
and authorization, financial evaluation, identification of
software and hardware problems and integrated circuit fail-
ures etc. [4]-[6].
The main objective of the present research is to develop a
low-cost automated knowledge-based system incorporating
the skills of medical experts to help the patients in self-
diagnosis and management of this chronic disease. Re-
cently, expert systems have been developed in laboratory
stage for diabetes awareness and management, and insulin
administration [7]-[11]. Compared to these systems, our
approach is more simple and pragmatic.
The paper is organized as follows. In Section II the
medical knowledge about diabetes is described briefly,
Section III presents system architecture, Section IV de-
scribes rule-based decision-making process as well as some
experimental results, and finally Section V draws the con-
clusions.
II. MEDICAL KNOWLEDGE OF DIABETES
Diabetes mellitus is a clinical syndrome characterized by
hyperglycemia, due to absolute or relative deficiency of
insulin. [12]
A. Classification of diabetes
1. Type-I (Insulin-Dependent Diabetes Mellitus—IDDM)
diabetes tends to occur in the young. Type-II (Non-
Insulin-Dependent Diabetes Mellitus—NIDDM) diabetes
mellitus occurs more often in older people who are obese
and had sedentary lifestyles.
2. Gestational diabetes mellitus (GDM) is a glucose intoler-
ance detected in a pregnant woman who was not known
to have these abnormalities prior to conception. [3]
Diagnosis and Management of Diabetes Mellitus through a Knowledge-Based System
B. Diagnosis
Diagnosis [13]-[14] is a process, by which a doctor
searches for the cause (disease) that best explains the symp-
toms of a patient. Our knowledge-based system is mainly
used for performing diagnosis based on patient data. Patient
data can be demographic or clinical. Demographic data
relates the information such as patient’s age, sex, location,
income, etc. Clinical data is divided into physical signs and
laboratory results. Physical signs are those detected by a
physical examination of patient, like BMI (body-mass in-
dex), pulse rate and blood pressure. Laboratory results are
those detected via laboratory tests, like blood test, urine test,
etc. The diagnosis system is based on following patient
data. [12]
1. Test urine for glucose and ketones.
2. Measure random or fasting blood glucose:
x Fasting plasma glucose >= 7.0 mmol/l
x Random plasma glucose >= 11.0 mmol/l.
3. Oral glucose tolerance test:
x Fasting plasma glucose 6.1-6.9 mmol/l
x Random plasma glucose 7.0-11.0 mmol/l.
C. Method of Treatment
x Diet alone— 50% can be controlled adequately.
x Diet + oral hypoglycemic agent— 20-30% can be con-
trolled.
x Diet and insulin— 30% can be controlled.
III. SYSTEM ARCHITECTURE
The computer-based system consists of a user interface,
an inference engine, knowledge base, working memory and
case history file. The case history file contains the demo-
graphic and clinical data of the patient. The knowledge base
contains the heuristic knowledge encoded in some form
(e.g., rules) for diagnosing diabetes. The working memory
contains the (initial) patient data, partial conclusions, data
given by user and other information related to the case un-
der consideration. It provides facts and serves as a store for
inferences or conclusions.
Inference engine mimics the expert system’s reasoning
process. It woks on the facts in the working memory and
uses the domain knowledge contained in the knowledge
base to derive (or infer) new facts as well as conclusions.
[5] It achieves this by searching through the knowledge
base to find rules whose premises match the facts contained
in the working memory. If such a match is found, it adds the
conclusion of the matching rule to the working memory.
_______________________________________________________________
IFMBE Proceedings Vol. 23
1001
This process continues until the inference mechanism is
unable to match any rules with the facts in the working
memory. And finally, the user interface allows interaction
with the inference engine to diagnose diabetes.
The structure of our expert system is shown in Fig. 1.
User interface
Patient Working
database memory
User
Inference engine
Knowledge base
Fig. 1: Architecture of the proposed system.
IV. DIAGNOSTIC APPROACH AND EXPERIMENTAL RESULTS
Diagnostic approaches can be grouped into four classes:
model-based, rule-based, neural-network based and case-
based. The knowledge representation schemes of model-
based and rule-based diagnoses can be treated as symbol-
oriented, while neural network and case-based diagnoses are
instance-oriented. [4] In our system we use rule-based ap-
proach. A rule-based expert system is an expert system
based on a set of rules that are used to make decisions. [6]
To design an expert system a knowledge engineering proc-
ess is needed, in which the rules used by human experts are
accumulated and translated into an appropriate form for
computer processing.
In our expert system we divide the treatment of diabetes
into three categories.
1. Type-I diabetes— only insulin is used.
2. Gestational diabetes— only insulin is used, and
3. Type-II, where the oral hypoglycemic agent, diet, and
sometimes insulin are used. In this category, there are
special treatments:
x The drugs for obese and lean patients.
x The drugs, which should not used for the patient
who have renal diseases, lactic acidosis, liver dis-
ease, etc.
_________________________________________________________________
1002 Morium Akter, Mohammad Shorif Uddin and Aminul Haque

Fig. 2: Sample test result (case1). Fig. 5: Sample test result (case4).

At present, the system uses 26 rules and is implemented

using Prolog programming language. We did experimenta-
tions using data of 100 patients. Some self-explanatory
sample test results are shown in Figs. 2- 5.

V. CONCLUSIONS

A rule-based expert system for diagnosis of diabetes has

been introduced in this paper. The aim of our expert system
is to help the diabetes patients in low-cost treatment and
management. Some real life experimentations were per-
formed, which confirmed the effectiveness of the proposed
system. The system can be used in both home and hospital
environments.
Fig. 3: Sample test result (case2). There are some limitations in the developed knowledge-
based system. For example, the number of rules is not suffi-
cient for a general robust expert system. Moreover, wide
real-life experimentations were not performed yet. Rooms
are also available for improvement of the expert system on
the basis of patients’ feedbacks. Our future goal is to over-
come the above limitations to make it a robust one.

REFERENCES
1. S. Wild, G. Roglic, A. Green et al., “Global prevalence of diabetes -
estimates for the year 2000 and projections for 2030”, Diabetes Care,
vol. 27, pp. 1047-1053, (2004).
2. Diabetic Association of Bangladesh, “Diabetes Mellitus”, 2005
(ISBN: 984-32-2552-X).
3. Hajera Mahtab, Zafar A Latif, Md. Faruque Pathan, “Diabetes Melli-
tus - A Handbook For Professionals”, BIRDEM, 3rd Ed., 2004
(ISBN: 984-31-0100-6).
4. Jie Yang, Chenzhou Ye, Xiaoli Zhang, “An expert system for fault
Fig. 4: Sample test result (case3). diagnosis”, Robotics, vol.19, pp. 669-674, (2001).

_______________________________________________________________ _________________________________________________________________
IFMBE Proceedings Vol. 23
Diagnosis and Management of Diabetes Mellitus through a Knowledge-Based System
5. Karthik Balakrishnan and Vasant Honavar, “Intelligent diagnosis
systems”, available online at http://www.cs.iastate.edu/~honavar/
aigroup.html (accessed on December 17, 2007).
6. Costas Papaloukas, I. Dimiritrios, Aristidis Likas, S. Christos
Stroumbis, Lampros K. Michalis, “Use of novel rule-based expert
system in the detection of changes in the ST segment and the T wave
in long duration ECGs”, J. Electrocardiology, vol. 35, no.1, (2002).
7. V. Ambrosiadou and A. Boulton, “A knowledge-based system for
education on management of diabetes”, Proc. IMACS World Con-
gress on Scientific Computation, 1988, pp. 186-189.
8. Rudi Rudi, and Branko G. Celler, “Design and Implementation of
Expert-Telemedicine system for Diabetes Management at Home”,
Intl. Conf. on Biomedical and Pharmaceutical Engineering (ICBPE
2006), pp. 595-599.
9. Diabetes be aware: avbaible online at http://www.hpb.gov.sg/
diabetes/ (accessed on July 11, 2008).
10. Kyung-Soon Park, Nam-Jin Kim, Ju-Hyun Hong, Mi-Sook Park, Eun-
Jong Cha, Tae-soo Lee, “PDA based Point-of-care Personal Diabetes
Management System”, Proceedings of the IEEE 27th Annual Confer-
ence on Engineering in Medicine and Biology (Shanghai, China, Sep-
tember 2005), pp. 3749-3752.
_______________________________________________________________
IFMBE Proceedings Vol. 23
1003
11. V. Ambrosiadou, Diabetes, “An expert system for education in diabe-
tes management”, in Expert Systems Applications, Sigma Press, UK,
pp. 227-238, (1989).
12. Davidson, “Principles and Practice of MEDICINE”, Churchill Liv-
ingstone, 19th Ed, pp. 644, (2002).
13. G. Gogou, N. Maglaveras, V. Ambrosiadou, D. Goulis, C. Pappas, “A
neural network approach in diabetes management by insulin admini-
stration”, J. Med. Syst. Vol. 25, no. 2, pp. 119-31, (2001).
14. Bert Kappan, Wim Wiegerinck, Ender Akay, “Promedas: A clinical
diagnostic decision support system”, Available online at
http://download.intel.com/research/share/UAI03_workshop/kappen/
kappen_uai2003.pdf (accessed on June 8, 2007).
Corresponding author:
Mohammad Shorif Uddin
Bioinformatics Institute
30, Biopolis Street, #07-01 Matrix
Singapore 138671
E-mail: msuddin@gmail.com
_________________________________________________________________