FRAMEWORK OF A CLINICAL DECISION SUPPORT SYSTEM FOR

DIAGNOSIS OF TYPE 1 DIABETES

VICTOR JOB

1163-04395-08353

A RESEARCH SUBMITTED TO THE SCHOOL OF COMPUTING AND

INFORMATION TECHNOLOGY IN PARTIAL FULFILLMENT FOR

THE REQUIREMENT OF THE AWARD OF POSTGRADUATE

DIPLOMA IN COMPUTER SCIENCE OF

KAMPALA INTERNATIONAL

UNIVERSITY

MAY, 2019

ii

DECLARATION

I declare that this research is my original work and has not been submitted for

any other award of a degree and published at any institution of higher learning,

except where due acknowledgement has been made in the text.

_______________________________ ____________________________

Victor Job Date

iii

APPROVAL

I declare that this research has been done by the student under my supervision

and is ready for further cross-examination by other examiners.

______________________ _______________________

Dr. Chinecherem Umerzuruike Date

iv

DEDICATION

I dedicate this research to God Almighty for His grace to get educated.

v

ACKNOWLEDGEMENT

My deepest, heartfelt thanks, first of all goes to the Almighty God for granting

me his protection and knowledge in coming out with this important study.

Furthermore, I wish to recognize my supervisor, Dr. C. Umerzuruike with whom

I worked closely during the stages of writing this thesis. I am also thankful to all

my lecturers at School of Computing and Information technology of Kampala

International University for being helpful and imparting knowledge to me in

diverse ways during Face-to-Face and beyond.

I would like to express utmost gratitude to my beloved family members and to

for their enthusiastic support, constant inspiration and blessings during my study.

Last but not least, I am also indebted to Mr Andrew Paul Okidi for offering me

various assistances during my study and research work. May God richly bless

you, including those who played diverse roles relating to the production of this

work but whose names did not come up for mentioning.

vi

LIST OF ACRONYMS

AI Artificial Intelligence

CDSS Clinical Decision Support System

CSS Cascading Style Sheet

DBMS Database Management System

DM Diabetes mellitus

PHP Hypertext Pre-processor

SDLC Systems Development Life Cycle

SMBG Self-Monitoring of Blood Glucose

UCI University of California

WHO World Health Organization

vii

ABSTRACT

Diabetes is one of the most prevalent chronic mellitus which occurs when blood

glucose (sugar) increases more than normal. Diabetes management has been

playing an important role on preventing or delaying occurrence of the life

threatening diabetes complications. By providing a roaming operation

environment, internet healthcare technologies can bring efficiencies and

convenience to both diabetics and relative medics on diabetes management. The

remote monitoring of patients is considered an important tool in facilitating

improvements in diabetes care. By using a web-based CDSS patients will have

access into database to enter their medical readings and queries which are related

with their chronic illness and they will be able to receive the latest developments

and recommendations to take careful and follow the medical advice by

practitioners. The system will gather their medicinal data and these data will be

held on a database and the system will process these data to validate the progress

of the patients. It will allow the health care provider, to make an informed

healthcare decision based on the collected data. The readings of diabetes and

recommendations of treatments are sent to the patient and their general

practitioners. This system will have the potential to reduce healthcare costs and

increase patient program compliance.

viii

TABLE OF CONTENTS

DECLARATION ................................................................................................................... ii

APPROVAL ......................................................................................................................... iii

DEDICATION ....................................................................................................................... iv

ACKNOWLEDGEMENT ....................................................................................................... v

LIST OF ACRONYMS .......................................................................................................... vi

ABSTRACT .........................................................................................................................vii

CHAPTER ONE ..................................................................................................................... 1

INTRODUCTION .................................................................................................................. 1

1.1 BACKGROUND TO THE STUDY .................................................................................. 1

1.2 STATEMENT OF PROBLEM .......................................................................................... 3

1.3 AIM AND OBJECTIVES ................................................................................................. 4

1.4 METHODOLOGY ............................................................................................................ 5

1.5 SIGNIFICANCE OF THE STUDY ................................................................................... 7

1.6 SCOPE OF THE STUDY .................................................................................................. 7

CHAPTER TWO .................................................................................................................... 9

LITERATURE REVIEW ........................................................................................................ 9

2.1 INTRODUCTION ............................................................................................................. 9

2.2 DIABETES ..................................................................................................................... 11

2.3 CLINICAL DECISION SUPPORT SYSTEM (CDSS) .................................................... 13

2.3.1 BENEFITS OF CDSS .................................................................................................. 16

2.3.2 APPLICATIONS OF CDSS ......................................................................................... 19

2.3.3 REVIEW OF EXISTING LITERATURE RELATED TO CDSS ................................. 21

CHAPTER THREE ............................................................................................................... 26

ix

METHODOLOGY................................................................................................................ 26

3.0 INTRODUCTION ........................................................................................................... 26

3.1 RESEARCH DESIGN..................................................................................................... 26

3.2 MODEL DESIGN – ALGORITHM ................................................................................ 28

3.2.1 NAIVE BAYES ........................................................................................................... 28

3.3 RESEARCH TOOLS ...................................................................................................... 29

3.3.1 WAMP PACKAGE ...................................................................................................... 29

3.3.2 ADOBE DREAMWEAVER (CC2015) ........................................................................ 31

3.3.3 RAPID MINER VERSION 6.2..................................................................................... 32

3.4 SOFTWARE DEVELOPMENT LIFE CYCLE ............................................................... 32

3.4.1 PROTOTYPE MODEL ................................................................................................ 33

CHAPTER FOUR ................................................................................................................. 37

4.1 CLINICAL DECISION SUPPORT SYSTEM MODULES ............................................. 37

4.1.1 USE CASE DIAGRAM FOR THE SYSTEM .............................................................. 41

4.1.2 DATABASE DESIGN ................................................................................................. 46

CHAPTER FIVE .................................................................................................................. 51

SUMMARY AND RECOMMENDATION .......................................................................... 51

5.1 SUMMARY .................................................................................................................... 51

5.2 RECOMMENDATION ................................................................................................... 51

5.3 FUTURE RESEARCH .................................................................................................... 52

REFERENCES ..................................................................................................................... 53

1

CHAPTER ONE

INTRODUCTION

1.1 BACKGROUND TO THE STUDY

The world is fast evolving and in order to cope with the insatiable demand of the

human race for the kind of living that can be described as top-notch in which

people have all they need at their desk and call, there is the need to develop

intelligent decision making applications that will drive systems or devices to carry

out tasks that require human intelligence. This concept is known as Artificial

Intelligence (AI). In science and technology, the desire for improvement is a

constant subject which triggers advancements (Kuyoro et al., 2018). Technology

has changed civilization in many different ways. Humans have always been on a

path of progression through the help of technology, the twentieth and twenty-first

centuries have seen a number of advancements that revolutionized the way people

work, live and play (Dejan, 2014).

A Clinical Decision Support System (CDSS) is an active knowledge system, where

two or more items of patient data are used to generate case-specific

recommendation(s) (Greenes, 2014). This implies that a CDSS is a Decision

Support System (DSS) that uses knowledge management to achieve clinical advice

for patient care based on some number of items of patient data. This helps to ease

2

the job of healthcare practitioners, especially in areas where the number of patients

is overwhelming (Sakiko et al., 2018). Clinical decision support system (CDSS)

provides clinicians, staff, patients or other individuals with knowledge and person-

specific information, intelligently filtered or presented at appropriate times, to

enhance health and health care. A CDSS can also be seen as an application that

analyses data to help healthcare providers make clinical decisions (Rouse, 2014).

Computer based methods are increasingly used to improve the quality of medical

services. Though, the fields in which computers are being used have very high

complexity and uncertainty; the uses of intelligent systems such as fuzzy logic,

artificial neural network and genetic algorithm have been developed (Jimoh et al,

2014). Recent advances in the field of Artificial Intelligence have led to the

emergence of expert systems and computational tools; designed to capture and

make available the knowledge of experts in a field. Hence, this work focuses on

the design a web-based clinical decision support system for diagnosis of type 1

diabetes.

Diabetes mellitus is a chronic disease caused by inherited and/or acquired

deficiency in production of insulin by the pancreas, or by the ineffectiveness of the

insulin produced. Such a deficiency results in increased concentrations of glucose

in the blood, which in turn damage many of the body's systems, in particular the

3

blood vessels and nerves. According to World Health Organization (WHO) (2014),

Diabetes mellitus is a group of metabolic diseases characterized by elevated blood

glucose level (hyperglycemia) resulting from defects in insulin secretion, insulin

action or both. Insulin is a hormone manufactured by the beta cells of the pancreas,

which is required to utilize glucose from digested food as an energy source.

Chronic hyperglycemia is associated with micro vascular and macro vascular

complications that can lead to visual impairment, blindness, kidney disease, nerve

damage, amputations, heart disease, and stroke (Harris, 2007). In 2014, an

estimated 29.1 million people or 9.3% of the U.S. population had diabetes

insinuating that both direct and indirect health care expenses were estimated at

$118 billion (National Diabetes Statistics Report, 2018).

1.2 STATEMENT OF PROBLEM

The World Health Organization (WHO) estimates that Diabetes Mellitus affects

over 366 Million people worldwide, and many without efficacious diabetes care. A

recent revelation by the WHO indicates that Diabetes has tripled in the last two

decade globally with the highest prevalence rates found in developing countries

(WHO, 2017).

In Nigeria with a population of over 180 million people, there are less than 200

diabetologists, few nutritionist, few diabetes educators to look after the estimated 8

4

million people with diabetes (Diabetes Association of Nigeria, 2018). Furthermore,

according to the WHO (2017), there is one specialist that treats diabetes available

for 10,000 Nigerians. There is a lot of concern due to this shortage of specialists,

there is the need for CDSSs to diagnose and treat diabetes in health care center.

Existing CDSSs make use of the blood sugar level to diagnose patients and come

up with a valid result as regards whether such person(s) have diabetes mellitus or

not. However, very little research efforts have been carried out using both the

blood sugar level and the plasma insulin level of the individual in question to

diagnose whether he or she has type 1 diabetes mellitus or not, thus making this

aspect presently pose itself as a gray area. This research consequently proposes

knowledge based adaptive clinical decision support system to diagnose and

prescribe treatment for type 1 diabetes mellitus based on the blood sugar level and

plasma insulin level of the person(s) in question.

1.3 AIM AND OBJECTIVES

The aim of this project is to develop knowledge based adaptive clinical decision

support system to diagnose and treat type 1 diabetes mellitus. The proposed system

will employ pattern classification algorithms that will help to determine if an

individual has type 1 diabetes mellitus or not.

5

Hence, the specific objectives of this research are to:

i. Evaluate pattern classification algorithms that are relevant to this study in

order to select the most accurate algorithm that will provide intelligence for

the clinical decision support system.

ii. Develop a web-based clinical decision support system to diagnose and

prescribe treatment for type 1 diabetes based on blood sugar level and

plasma insulin level.

1.4 METHODOLOGY

Carry out an evaluation of pattern classification algorithms on appropriate datasets

using the Rapidminer version 6.2 software. The process of evaluation will involve

the following steps:

Acquisition of relevant datasets from the UCI (University of California,

Irvene) repository to be employed for the experiments.

Preparation of the acquired dataset to a format supported by the Rapidminer

software.

Induce the chosen dataset with various classification algorithms and hence

obtain the accuracy of each of the algorithms used in the experiments.

6

The Clinical Decision Support System (CDSS) has a user interface that will be

used to feed patient data into the system; the system also has a Medical knowledge

base that houses the following modules: The diagnostic module that contains

symptoms, diagnostic results and prescribed medication. The diagnostic results and

prescribed medication will be displayed through the user interface.

The medical expert knowledge module that houses the knowledge of the

medical expert as regards type 1 diabetes.

The electronic medical records that contains the details of all persons that

the system has been used to diagnose as regards their type 1 diabetes status.

The CDSS also includes a Classification algorithm module within which patient

data is transformed before the classification is done by the algorithm.

The developed system also has an inference engine that searches through the

knowledge base to match the respective patient's data with the appropriate

inference.

The web-based user interface will be developed using:

Windows 7 operating system, Apache web server, MySQL database and

PHPmyAdmin

Mozilla Firefox web browser.

7

Cascading Style Sheet (CSS)

Adobe Dreamweaver CC2015

Front and back end Javascript framework

1.5 SIGNIFICANCE OF THE STUDY

This research provides an architecture that serves as the platform upon which an

intelligent CDSS will be built and operationalized for the diagnosis and treatment

of type 1 diabetes mellitus. It makes it possible for General Practitioners (GPs)

who have little idea in the area of diabetes to have insights into the diagnosis and

treatment including patient-specific care plans for patients with type 1 diabetes

mellitus. This project finds its usefulness in remote areas where accessibility is far

and medical experts are needed urgently and it also have the potential to

significantly improve management of diabetes mellitus Type 1, especially in

remote areas or rural areas that are having limited access to good quality health

service.

1.6 SCOPE OF THE STUDY

The dissertation intends to make contributions in terms of generating patient-

specific Clinical Pathways (CPS) for diabetes that targets the knowledge needs and

clinical duties of General Practitioners (GPs), especially those working in Nigeria.

Although, the CPS expected to be generated when this work is fully implemented

8

will be based on existing Clinical Practice Guidelines (CPGs), they will be made to

incorporate expert medical knowledge of specialists that treat diabetes who are

domicile in Nigeria.

9

CHAPTER TWO

LITERATURE REVIEW

2.1 INTRODUCTION

Diabetes is the condition in which the body does not properly process food for use

as energy (Centers for Disease Control and Prevention, 2017). Most of the food we

eat is turned into glucose, or sugar, for our bodies to use for energy. The pancreas,

an organ that lies near the stomach, makes a hormone called insulin to help glucose

get into the cells of our bodies. According to the National Institute of Diabetes and

Digestive and Kidney Diseases (NIDDKD) (2018), the number of diabetics in the

world stands at 365 million people, representing around 8.5% of the global

population. Elhendi (2015) explains that diabetes is a common hormonal problem

that if untreated can lead to diabetes complications such as diabetic neuropathy,

kidney problems, heart problems, retinopathy and other disorders. At advanced

stages, diabetes can cause kidney failure, amputation, blindness and stroke.

However, complications can be prevented or significantly delayed by exercising

good control of diabetes, blood pressure and cholesterol.

According to International Diabetes Federation (IDF) (2018), it was estimated that

in 2017 there were 451 million (age 18–99 years) people with diabetes worldwide.

These figures are expected to increase to 693 million) by 2045. It was estimated

that almost half of all people (49.7%) living with diabetes are undiagnosed.

10

Moreover, there was an estimated 374 million people with impaired glucose

tolerance (IGT) and it was projected that almost 21.3 million live births to women

were affected by some form of hyperglycaemia in pregnancy. In 2017,

approximately 5 million deaths worldwide were attributable to diabetes in the 20–

99 years age range. The global healthcare expenditure on people with diabetes was

estimated to be USD 850 billion in 2017.

Diabetes is a chronic, metabolic disease characterized by elevated levels of blood

glucose (or blood sugar), which leads over time to serious damage to the heart,

blood vessels, eyes, kidneys, and nerves. The most common is type 2 diabetes,

usually in adults, which occurs when the body becomes resistant to insulin or

doesn't make enough insulin. In the past three decades the prevalence of type 2

diabetes has risen dramatically in countries of all income levels. Type 1 diabetes,

once known as juvenile diabetes or insulin-dependent diabetes, is a chronic

condition in which the pancreas produces little or no insulin by itself. For people

living with diabetes, access to affordable treatment, including insulin, is critical to

their survival. There is a globally agreed target to halt the rise in diabetes and

obesity by 2025 (Ogurtsova et al., 2017).

It is known that the normal blood glucose level lies between (70-100) mg/100 ml

when fasting and around 140 mg/ 100 ml otherwise. For a diabetic person, the

11

blood glucose is around 126 mg/ 100 ml when fasting and 200 mg/ 100 ml

otherwise. The most common symptoms observed in diabetic patients are:

polyuria, weight loss, excessive thirst, continuous hunger, blurring and changes in

vision, and fatigue. The overall risk of dying among people with diabetes is at least

double the risk of their peers without diabetes (Cho et al., 2018).

2.2 DIABETES

The earliest description of diabetes was documented in the writings of Hindu

scholars as long as in 1500 BC. They had already described "a mysterious disease

causing thirst, enormous urine Output, and wasting away of the body with flies and

ants attracted to the urine of people." The term diabetes was probably coined by

Apollonius of Memphis around 250 BC, which literally meant "to go through" or

siphon as the disease drained more fluid than a person could consume.

Later on, the Latin word "mellitus" was added because it made the urine sweet.

Diabetes is the condition in which the body does not properly process food for use

as energy. Most of the food we eat is turned into glucose, or sugar, for our bodies

to use for energy. The pancreas, an organ that lies near the stomach, makes a

hormone called insulin to help glucose get into the cells of our bodies. When you

have diabetes, your body either doesn't make enough insulin or can't use its own

insulin as well as it should. This causes sugars to build up in your blood. Diabetes

12

can cause serious health complications including heart disease, blindness, kidney

failure, and lower-extremity amputations. Diabetes is the seventh leading cause of

death in the United States.

Elevated blood sugar is a common effect of uncontrolled diabetes, and over time

can damage the heart, blood vessels, eyes, kidneys, and nerves. Some health

complications from diabetes include:

Diabetic retinopathy which is a significant cause of blindness, and occurs as

a result of long term accumulated damage to the small blood vessels in the

retina. After 15 years of diabetes about 10% of patients develop severe

visual impairment.

Diabetic neuropathy is damage to the nerves as a result of diabetes, and

affects up to 50% of people with diabetes. Common symptoms are tingling,

pain, numbness, or weakness in the feet and hands combined with reduced

blood flow, neuropathy in then feet increases the chance of foot ulcers and

eventual limb amputation.

Diabetes is among the leading causes of kidney failure; 10-20% of people with

diabetes die of kidney failure and diabetes increases the risk of heart disease and

stroke; 50% of people with diabetes die of cardiovascular disease (primarily heart

disease and stroke).

13

Type 1 diabetes can affect people at any age, but usually develops in children or

young adults. People with type 1 diabetes need daily injections of insulin to control

their blood glucose levels. If people with type 1 diabetes do not have access to

insulin, they will die.

The risk factors for type 1 diabetes are still being researched. However, having a

family member with type 1 diabetes slightly increases the risk of developing the

disease. Environmental factors and exposure to some viral infections have also

been linked to the risk of developing type 1 diabetes.

At present, type 1 diabetes cannot be prevented. The environmental triggers that

are thought to generate the process that results in the destruction of the body’s

insulin-producing cells are still under investigation.

2.3 CLINICAL DECISION SUPPORT SYSTEM (CDSS)

Since the beginning of computers, physicians and other healthcare professionals

have expected the time when machines would aid them in the clinical decision-

making and other restorative procedures. CDSS provides clinicians, patients or

individuals with knowledge and person-specific or population information,

intelligently filtered or presented at appropriate times, to foster better health

processes, better individual patient care, and better population health (Farooq et al.,

2017).

14

In the late 1950s, the very first articles regarding this provision appeared and

within a few years, experimental prototypes were made available. In 1970, three

advisory systems have provided a useful overview of the origin of the work on

CDSS: MYCIN system by Shortliffe for the selection of antibiotic therapy, a

system by deDombal for the diagnosis of abdominal pain, and a system called

HELP for generating inpatient medical alerts (Kuperman, et al., 2013).

Indeed the study of Clinical decision support systems (CDSS) constitutes a

significant field of usage of information technology in healthcare. CDSS are

designed to assist clinicians and other healthcare professionals in diagnosis as well

as decision-making. CDSS uses healthcare data and a patient’s medical history to

make recommendations. By using a predefined set of rules, CDSS intelligently

filters knowledge from complex data and presents at an appropriate time. By

adopting CDSS, healthcare can become more accessible to large populations.

However, it also implies that at times, CDSS may be used by people having literal

medical knowledge (Ahn et al., 2014).

Thus, clinical decision support systems (CDSS) have since been adopted in the

medical fraternity to help clinicians, patients, and others to suggest patient-

appropriate evidence-based treatment options. Ontologies are essential tools for the

organization and representation of knowledge (Bau et al., 2014). Ontologies

contain the collection of patients, symptoms, diseases, diagnoses, treatments, and

15

drug information, thereby creating a healing strategy according to patient’s

requirements to reconfigure a clinical decision support system (Chen et al., 2012).

Some of the studies suggested using ontologies to build clinical guidelines and care

plans (Zhang et al., 2016; Sherimon & Krishnan (2016); Marco-Ruiz et al., 2016).

In most of the knowledge ontologies, there is a design by the experience of domain

experts. For example, Bau et al., (2014) used domain ontology and rule reasoning

to construct a CDSS for diabetic patients undergoing surgery. They had three main

classes in this ontology: disease, management, and patient. The disease class

consisted of diabetes and comorbidity information. The management class

consisted of anesthesia, capillary glucose tests, control of DM, medication, no

medication, and water restriction information. The patient class consisted of the

patient clinical profile. The system constructed a clinical decision support system

(CDSS) for undergoing surgery based on domain ontology and rules reasoning in

the setting of hospitalized diabetic patients.

On the other hand, a study by Rung-Ching et al., (2017) adopted the American

Diabetes Association (ADA) and the European Association for the Study of

Diabetes (EASD) published to propose an HbA1c target and antidiabetic

medication recommendation system for patients. Based on the antidiabetic

medication profiles which were presented by the American Association of Clinical

Endocrinologists (AACE) and American College of Endocrinology (ACE), the

16

study used TOPSIS to calculate the ranking of antidiabetic medications. The

endocrinologist set up ten virtual patients’ medical data to evaluate a decision

support system. The system indicated that the CDSS performed well and was

useful to 87%, and the recommendation system was suitable for outpatients. The

evaluation results of the antidiabetic medications showed that the system had 85%

satisfaction degree which could assist clinicians to manage T2DM while selecting

antidiabetic medications. In addition to aiding doctors’ clinical diagnosis, the

system not only could serve as a guide for specialty physicians but also could help

non specialty doctors and young doctors with their drug prescriptions.

2.3.1 BENEFITS OF CDSS

The key benefits of CDSS reported in the studies conducted by Haynes and

Wilczynski (2010), Kawamoto et al., (2010), Wright et al., (2011), Musen et al.,

(2014) are as follows:

1. Higher standards of patient safety: CDSSs have helped healthcare organizations

all over the world acquiring higher standards of patient safety by adopting

standardized clinical procedures governed by the clinical workflows encoded

through these systems. Thus reducing diagnostic and prescribing errors and drug

doubling issues.

17

2. Improving the quality of direct patient care: Their research also concluded that

with the advent of CDSS, quality of care has improved to considerable levels with

this extra support provided to clinicians (who are already struggling to cope with

current healthcare demands). This has made it possible for clinical experts to

allocate more time in providing direct patient care.

3. Standardization and conformance of care using clinical practice guidelines: The

standardization of clinical pathways and procedures set precedents and evaluation

benchmarks for healthcare trusts to achieve higher patient satisfaction levels set out

by different healthcare organizations in different regions. CDSS also promote the

utilization of clinical practice guidelines (CPGs) for the development of

knowledge-aware systems capable of performing effective clinical decision-

making to promote standardized care.

4. Collaborative decision-making: CDSS have helped healthcare stakeholders that

include clinicians, healthcare trusts and policy makers to develop safe and efficient

care models using a collaborative decision-making approach to benefit both patient

and a clinician. CDSS have also helped healthcare trusts to improve effectiveness

in the prescribing facility through cost-effective drug order dispensation (Wright et

al., 2011). CDSS are also playing an important role in the integration of EHRs,

which will help healthcare authorities to streamline information collection and

18

clinical diagnosis operations in order to promote efficient data gathering (Ivbijaro

et al., 2008). The audit trail is another important aspect of modern healthcare

systems which is achieved through the intelligent exploitation of clinical decision

support capabilities.

Many reviews have identified the benefits of the CDSSs, in particular,

Computerized Physician Order Entry systems (Zuccotti et al., 2014). The CDSS as

part of the Computerized Physician Order Entry has been found to alleviate

adverse drug events and medication errors (Jaspers et al., 2011; Wong et al., 2012).

CDSSs also have demonstrated to improve clinician performance, by way of

promoting the electronic prescription of drugs, adherence to guidelines and to an

extent the efficient use of time (Jaspers et al., 2014). CDSSs perform a key role in

providing primary care and preventative measures at outpatient clinics, e.g. by

alerting caregivers of the need for routine blood pressure checking, to recommend

cervical screening, and to offer influenza vaccination (Ahmadian et al., 2011).

To provide effective healthcare delivery to patients, CDSS is used both in primary

and secondary care units. In order to take maximum advantage from cardiovascular

CDSS, it is required to ensure clinical governance in the next-generation clinical

systems by considering a strong foundation in well-established clinical practice

guidelines and evidence based medicine (Farooq & Hussain, 2016).

19

The adoption of CDSSs in diagnosis and management of chronic diseases, such as

diabetes (O’Connor et al., 2011), cancer, dementia, heart disease, and hypertension

have played significant clinical roles in the main health care organizations in the

improvement of clinical outcomes of the organizations worldwide at primary and

secondary care. These CDSS also provide the foundation to system developer and

knowledge expert to collate and build domain expert knowledge for screening by

clinicians and clinical risk assessment (Wright et al., 2011).

2.3.2 APPLICATIONS OF CDSS

CDSSs are considered as an important part in the modern units of healthcare

organizations. They facilitate the patients, clinicians and healthcare stakeholders by

providing patient-centric information and expert clinical knowledge (Classen et al.,

2011). To improve the efficiency and quality of healthcare, the clinical decision-

making uses knowledge obtained from these smart clinical systems. The

Automated DSSs of Cardiovascular are available in primary health care units and

hospital in order to fulfil the ever-increasing clinical requirements of prognosis in

the domain of coronary and cardiovascular diseases. The computer-based decision

support strategies have already been implemented in various fields of

cardiovascular care (Kuperman et al., 2007). In the US and the UK, these

applications are considered as the fundamental components of the clinical

informatics infrastructures.

20

Ontology-driven DSS are being used widely in the clinical risk assessment of

chronic diseases. The ontology-driven clinical decision support (CDS) framework

for handling comorbidities in (Abidi et al., 2012) presented remarkable results in

the disease management and risk assessment of breast cancer patients, which was

deployed as a CDSS handling comorbidities in the healthcare setting for primary

care clinicians in the Canada. They utilized semantic web techniques to model the

clinical practice guidelines which were encoded in the form of a set of rules

(through a domain-specific ontology) utilized by CDSSs for generating patient-

specific recommendations.

Matt-MouleyBoumrane from the “University of Glasgow, ‘UK” implemented an

ontology-driven approach to the development of CDSS in the pre-operative risk

assessment domain. In (Bouamrane, et al., 2009), they reported their work by

combining a preventative care software system in the pre-operative risk assessment

domain with a decision support ontology developed with a logic based knowledge

representation formalism. In (Farooq et al., 2012), the authors demonstrated

utilization of ontology and machine learning inspired techniques for the

development of a hybrid CDS framework for cardiovascular preventative care.

Their proposed CDS framework could be utilized for automatically conducting

patient pre-visit interviews. Rather than replacing human experts, it would be used

to prepare the patients before visiting a hospital, deliver educational materials,

21

preorder appropriate tests, cardiac risk assessment scores, heart disease and cardiac

chest pain scores. It would make better use of both patient and clinician time.

The ontology-driven recommendation and clinical risk assessment system could be

used as a triage system in the cardiovascular preventative care which could help

clinicians prioritize patient appointments after reviewing snapshot of patient’s

medical history (collected through an ontology-driven intelligent context-aware

information collection using standardized clinical questionnaires) containing

patient demographics information, cardiac risk scores, cardiac chest pain and heart

disease risk scores, recommended lab tests and medication details. In (Farooq and

Hussain 2016), they also have validated the proposed novel ontology and machine

learning driven hybrid CDS framework in other application areas.

2.3.3 REVIEW OF EXISTING LITERATURE RELATED TO CDSS

In their study Zeki et al., (2012) designed an expert system for diagnosis all types

of diabetes. After data acquisition and designing a rule-based expert system, this

system was coded with VP_Expert Shell and tested in Shahid Hasheminezhad

Teaching Hospital affiliated to Tehran University of Medical Sciences and final

expert system has been presented. Findings of this research showed that in many

parts of medical science and health care the expert systems have been used

effectively.

22

The acquisitive knowledge was represented in the diagrams, charts and tables. The

related source code using of the expert system was given and after testing the

system, finally its validation was done. It has been concluded here that the expert

system can be used effectively in all areas of medical sciences. In particular, in

terms of the vast number of diabetics throughout the world, the expert system can

be highly helpful for the patients.

They asserted that since this expert system gathers its knowledge from several

medical specialists, the system has a broader scope and can be more helpful to the

patients in comparison to just one physician. The flaw in this work however is that

in coding the knowledge from the different medical specialists in the knowledge

base of the expert system, the knowledge acquisition bottleneck is likely to result

leading to some information in the knowledge base being code wrongly and this

will make reduce the accuracy of the inferences that this expert system will

generate.

In their study Devarapalli et al., (2013) proposed a novel concept of designing and

building intelligent expert systems for the detection and diagnosis of Diabetes

Mellitus. The expert system classification was based on critical diabetic parameters

like Brain-Derived Neurotropic Factor (BDNF) levels, and Fasting Blood Glucose

(FBG). The proposed rule-based expert system constructs large-scale

23

knowledgebase based on the behavior of the BDNF related diabetic data. The

system gave an expert decision taking into consideration all the valid ranges of

diabetic parameters. The proposed expert system could work effectively even for

large sets of patient data.

In this study, Sahar (2013) described a web based intelligent decision support

system (IDSS) for type 2 diabetes patients, where more emphasis is put on patient

empowerment and self-management in treatment. The IDSS was based on a well-

documented decision support system, designed mainly for type 2 diabetes patients.

The scope of IDSS was broad from being mainly used by type 2 diabetes patients

as a tool for their empowerment, self-management, communication, and education.

The result of the study revealed and gave 100% effectiveness and correctness as 21

cases already diagnosed were subjected to the Expert System which also diagnosed

positive. A major weakness identified with this study among others was non-

suitability and non-workability of the Expert System in sub Saharan nations and

also the implementation cost which is enormous.

In this study, Sridar and Shanthi (2013) fused Artificial Neural Network (ANN)

with Association Rule Mining (ARM) for better accuracy. The study was focused

on designing and developing a Web based Artificial Neural Network with

Association Rule Mining for the diagnosis of diabetes. The data gathered

comprised of the following attributes: glucose level, Body Mass Index (BMI),

24

Systolic Blood Pressure, Diastolic Blood Pressure, cholesterol and fasting blood

sugar. The web based system combined ANN and ARM to give better accuracy in

diagnosing Diabetes Mellitus using the above mentioned input parameters..

Kuyoro et al., (2018), designed a web-based Clinical Decision Support System for

the management of early diabetes neuropathy. Four pattern classification

algorithms (K-nearest neighbor, Decision Tree, Decision Stump and Rule

Induction) were adopted in this work and were evaluated to determine the most

suitable algorithm for the clinical decision support system. Datasets were gathered

from reliable sources; two teaching hospitals in Nigeria, these were used for the

evaluation Benchmarks such as performance, accuracy level, precision, confusion

matrices and the models building’s speed were used in comparing the generated

models. The study showed that Naïve Bayes outperformed all other classifiers with

accuracy being 60.5%. k-nearest neighbor, Decision Tree, Decision Stump and

Rule induction perform well with the lowest accuracy for x- cross validation being

36.5%. Decision Tree falls behind in accuracy, while k-nearest neighbour and

Decision Stump maintain accuracy at equilibrium 41.0%.

Therefore, Naïve Bayes is adopted as optimal algorithm in the domain of this

study. The rules generated from the optimal algorithm (Naïve Bayes) forms the

back-end engine of the Clinical Decision Support System. The web-based clinical

25

decision support system was then designed. The automatic diagnosis of diabetes

neuropathy is an important real-world medical problem. Detection of diabetes

neuropathy in its early stages is a key for controlling and managing patients early

before the disabling effect present. This system can be used to assist medical

programs especially in geographically remote areas where expert human diagnosis

not possible with an advantage of minimal expenses and faster results.

26

CHAPTER THREE

METHODOLOGY

3.0 INTRODUCTION

Research methodology refers to the means necessary to systematically proffer step-

by-step description of how solutions to actualize stated objectives of a research

work will be carried out. It is sometimes perceived as trying to understand how

research is scientifically conducted. It outlines the various steps that are adopted by

the researcher in achieving the objectives of the research along with the logic

behind the adopted methods.

3.1 RESEARCH DESIGN

1. Carry out an evaluation of pattern classification algorithms on appropriate

datasets using the Rapidminer version 6.2 software. The process of evaluation will

involve the following steps:

Acquisition of relevant diabetes mellitus patient's datasets from National

Referral Hospital in Nigeria repository to be employed for the experiments.

Preparation of the acquired dataset to a format supported by the Rapidminer

software.

Induce the chosen dataset with various classification algorithms and hence

obtain the accuracy of each of the algorithms used in the experiments. The

performance of the various classifiers were determined using the Rapid

27

Miner software in which diabetes mellitus patient's data were induced with

the above named algorithms and the percentage of correctly classified

instances were used to evaluate the classifiers and thus determine their

accuracy level. The Rapid Miner open source tool was employed because it

is open source and contains a rich suite of classification algorithms with

their implementation.

2. The Clinical Decision Support System (CDSS) has a user interface that will be

used to feed patient data into the system; the system also has a knowledge base that

houses the following modules:

The diagnostic module that contains symptoms, diagnostic results and

prescribed medication. The diagnostic results and prescribed medication will

be displayed through the user interface.

The medical expert knowledge module that houses the knowledge of the

medical expert as regards type 2 diabetes.

The electronic medical records that contains the details of all persons that

the system has been used to diagnose as regards their type 2 diabetes status.

The CDSS also includes a Classification algorithm module within which patient

data is transformed before the classification is done by the algorithm. The

developed system also has an inference engine that searches through the

28

knowledge base to match the respective patient's data with the appropriate

inference.

The web-based user interface will be developed using:

Windows 7 operating system, Apache web server, MySQL database

Mozilla Firefox web browser.

PHP MYADMIM

Cascading Style Sheet (CSS)

Adobe Dreamweaver CC2015

Front and back end Javascript framework

3.2 MODEL DESIGN – ALGORITHM

Accurate and most efficient algorithm (Naïve Bayes) was chosen after comparative

analysis has been done on different data mining algorithms. The model was built

from the surprised algorithm trained with the data collected from the diabetes

mellitus system. If-then rules that can be used for diagnostic purposes were

generated from the optimal algorithm after various comparisons has been done.

3.2.1 NAIVE BAYES

The Naive Bayes classifier is a simple probabilistic classifier which is based on

Bayes theorem with strong and naive independence assumptions. It is one of the

most basic text classification techniques with various applications in email spam

29

detection, personal email sorting, document categorization, sexually explicit

content detection, language detection and sentiment detection. Despite the naive

design and oversimplified assumptions that this technique uses, Naive Bayes

performs well in many complex real-world problems.

Even though it is often outperformed by other techniques such as boosted trees,

random forests, Max Entropy, Support Vector Machines etc, Naive Bayes classifier

is very efficient since it is less computationally intensive (in both CPU and

memory) and it requires a small amount of training data. Moreover, the training

time with Naive Bayes is significantly smaller as opposed to alternative methods.

Naive Bayes classifier is superior in terms of CPU and memory consumption as

shown by Huang (2003), and in several cases its performance is very close to more

complicated and slower techniques.

3.3 RESEARCH TOOLS

3.3.1 WAMP PACKAGE

WAMP is the abbreviation of the package: Windows, Apache, MySQL, and one of

Perl, PHP or Python in Microsoft Windows operating system. Apache HTTP

Server is the web server software. PHP stands for the Hypertext Pre-processor. It is

a kind of HTML embedded language which executed on the server. MySQL server

is a small relational database management system. Initially, they are the entire

30

independent program for each other, however, they are always used together, and

have increasingly high compatibility degrees. So, these packages formed a

powerful web application.

PHP (PHP Hypertext Preprocessor) this served as the server side scripting

language and was used to connect to the database also alongside with the

HTML code.

HTML (Hyper Text Markup language): This is a markup language that was

used to create or design a website of the system.

CSS (Cascading style Sheet): This will be used to make all the designs the

website needs to have. Like the looks and feel of the website.

Jet Brains PI-IP Storm was used as the integrated development

environment in which all the web language codes were executed.

JavaScript: This served as the client side scripting language. Used in this

research was a front end JavaScript framework (Angular IS) and backend

JavaScript framework (Node

Apache: This was the server used to execute the web code created with

HTML and PHP

phpAdmin: This was used to create my database and tables needed in the

website.

31

Localhost This is the host/address of the local machine used or in use. The

IP is 127.0.0.1

MySQL: This was the type of database used in the course of design.

3.3.2 ADOBE DREAMWEAVER (CC2015)

Adobe (CC2015) was a series of software suites of graphic design, video editing,

and web development applications made or acquired by Adobe Systems (Connor,

2018). Dreamweaver CS6 is the first web editor built for the multiplatform era,

with full support for HTML5 and CSS3. The Adobe CC2015 is the Integrated

Development Environment (IDE) for the design and programming of the system. It

is a proprietary web development tool developed by Adobe systems. It has

improved support for web technologies such as Cascading Style Sheet (CSS).

Cascading Style Sheet is a style sheet language used for describing the look and

formatting of a document written in a mark-up language. It defines how HTML

elements are to be displayed. CC2015 is designed primarily to enable the

separation of document content from document presentation, including elements

such as the layout, colors, and fonts. JavaScript and various server side scripting

language such as Hypertext pre-processor (PHP) and frameworks.

32

3.3.3 RAPID MINER VERSION 6.2

This is a software platform developed by the company of the same name that

provides an integrated environment for machine learning, data mining, text mining,

and predictive analytics. It is used for business and industrial applications as well

for research, education, training, rapid prototyping, and application development

and supports all steps of the data mining process including results validation,

visualization and optimization.

RapidMiner, was formerly known as YALE (Yet Another Learning Environment),

and it was developed in the year 2001 by Ralf Klinkenberg, Ingo Mierswa, and

Simon Fischer at the Artificial Intelligence Unit of the Technical University of

Dortmund. (Guido, 2010) Starting in 2006, its development was driven by Rapid-I,

a company founded by Ingo Mierswa and Ralf Klinkenberg in the same year. In

the year 2007, the name of the software was formally changed from YALE to

RapidMiner and the company Rapid-I GmbH was incorporated.

3.4 SOFTWARE DEVELOPMENT LIFE CYCLE

The systems development life cycle (SDLC), also referred to as the application

development lifecycle, is a term used in systems engineering, information systems

and software engineering to describe a process for planning, creating, testing, and

deploying an information system. It is composed of a number of clearly defined

33

and distinct work phases which are used by systems engineers and systems

developers to plan for, design, build, test, and deliver information systems. SDLC

provides a series of steps to be followed to design and develop a software product

efficiently. There are different types of software development life cycle which are:

Waterfall model

Iterative model

Spiral model

Agile model

Prototype model

RAD model (Rapid application development)

cocomo model: cost to cost model

V-model

Fish model

3.4.1 PROTOTYPE MODEL

This study will use the prototype model to develop the CDSS. The basic idea

in Prototype model is that instead of freezing the requirements before a design or

coding can proceed, a throwaway prototype is built to understand the requirements.

This prototype is developed based on the currently known requirements. Prototype

model is a software development model. By using this prototype, the client can get

34

an “actual feel” of the system, since the interactions with prototype can enable the

client to better understand the requirements of the desired system. Prototyping is

an attractive idea for complicated and large systems for which there is no manual

process or existing system to help determining the requirements.

The prototypes are usually not complete systems and many of the details are not

built in the prototype. The goal is to provide a system with overall functionality.

Diagram of Prototype model:

Advantages of Prototype model:

The advantage of prototypes is that they can be changed faster and

modifications cost less time and money. That’s why the development of a

prototype should take place at the beginning of the product development

process.

Users are actively involved in the development

Requirement

gathering

Quick Design Building Prototype

Engineer Product Refining Prototype Customer

Evaluation

Start

Stop

35

Since in this methodology a working model of the system is provided, the

users get a better understanding of the system being developed.

Errors can be detected much earlier.

Quicker user feedback is available leading to better solutions.

Missing functionality can be identified easily

Confusing or difficult functions can be identified

Requirements validation, Quick implementation of, incomplete, but

functional, application.

Disadvantages of Prototype model:

Leads to implementing and then repairing way of building systems.

Practically, this methodology may increase the complexity of the system as

scope of the system may expand beyond original plans.

Incomplete application may cause application not to be used as the

full system was designed

Incomplete or inadequate problem analysis.

When to use Prototype model:

Prototype model should be used when the desired system needs to have a lot

of interaction with the end users.

36

Typically, online systems, web interfaces have a very high amount of

interaction with end users, are best suited for Prototype model. It might take

a while for a system to be built that allows ease of use and needs minimal

training for the end user.

Prototyping ensures that the end users constantly work with the system and

provide a feedback which is incorporated in the prototype to result in a

useable system. They are excellent for designing good human computer

interface systems.

37

CHAPTER FOUR

4.1 CLINICAL DECISION SUPPORT SYSTEM MODULES

A web based Clinical Decision Support System (CDSS) will be developed with a

user interface that will be used to feed patient data into the system.

These tasks will be accessible by users through simple interfaces, such as:

1. Interface to login onto patient interface.

2. Interface for re-entering the correct user name and password.

3. Interface to edit patient details.

4. Interface to confirm that the patient details have been successfully updated.

5. Interface to add the medical readings of the patient such as, blood pressure,

glucose, and weight

6. Interface to confirm that the added reading has been successful.

38

The system will have a knowledge base that houses the following modules:

Figure 4.1: Architecture for the Design CDSS for Diabetes Type 1

The system will also have a knowledge base that houses the following modules:

i. Diagnostic Module: The diagnostic module that contains symptoms, diagnostic

results and prescribed medication / management techniques. The diagnostic results

and prescribed medication / management techniques will be displayed through the

user interface.

ii. Knowledge Base: The medical expert knowledge module that houses the

knowledge of the medical expert as regards type 1 diabetes neuropathy.

39

iii. The electronic medical records that contains the details of all persons that the

system has been used to diagnose as regards their type 1 diabetes neuropathy

status. As such, the medical record must contain sufficient information to identify

the patient to whom it relates, as well as information relevant to the patient's

treatment during current and future episodes of care.

iv. Classification Algorithm Module: Pattern classification algorithms refer to the

theory and algorithms of assigning abstract objects into distinct categories with

each category typically known in advance. The CDSS has a Classification

algorithm module within which patient data is transformed before the classification

is done by the algorithm. The result of classification given by the classification

algorithm gives the status of the person in question as regards diabetes neuropathy.

v. Naives Bayes Algorithm

Naives Bayes Algorithm is preferred in this study because it maintains accuracy

and avoids errors. Thus considerable care will be taken to ensure that the datasets

have correct values. To deal with the condition of zero probability values of some

of the parameters, the Laplace Correction will be used as demonstrated below:

40

Naives Bayes Algorithm architecture

Where; K-nearest neighbor (KNN), Decision Tree (DT), Decision Stump (DS), and

Rule Induction (RI)

vi. Inference Engine: The inference engine serves as the processing component

within the system that coordinates the activities within the system in such a way as

to be able to search thoroughly through the knowledge base to match the respective

patient data with the right diagnosis for such patients. It thereafter generates

reminders, alerts and therapeutic recommendations for each patient the system is

used to diagnose.

vii. Database Design: This is the design of the SQL database which consist of the

specifications of the various relations, fields that the table contains and their

corresponding data type as well as the field length specifications. In the MySQL

environment, the term database referred to a collection of tables (relations) and

41

other database objects such as indexes. A table consists of rows and columns, these

rows and columns the data for the table.

4.1.1 USE CASE DIAGRAM FOR THE SYSTEM

A use case diagram can be defined as a representation of a user's interaction with

the system and depicting the specifications of a use case. A use case diagram can

depict the different kind of users of a particular system and will often be convoyed

by other types of diagrams as well. Below is the use case diagram for the web-

based Clinical Decision Support System for the diagnosis and management of

Type 1 diabetes showing all the stakeholders of the system.

42

Figure 4.2: Use Case Diagram for the CDSS for Type 1 Diabetes

43

Figure 4.3: Use Cases Model

44

Figure 4.4: Step by step decision support for diabetes type I

45

Table 4.1: Use Case Specification (Add patient profile)

Use case name: Add

Patient Profile

Brief Decryption The patient tries to add a profile to the system database

Primary actors Patient

Secondary actors Nurse, Physician

Precondition Login to the system

Description Step Action

1 When the patient selects Add Patient Profile

from the Menu, the use case will start.

2 Add Patient form will be loaded to the system

3 The patient will then enter details on the

necessary fields.

4 The patient will press on the submit button

5 The system will validate the inputted data

6 The system will add the patient profile to the

database

7 The system will inform the patient that the

details were successfully added to the system

46

Priority Must

Performance Response from system

Channels to actors Online

Post conditions Successfully added the patient profile to the medical

database

Alternative flow Entering invalid data

4.1.2 DATABASE DESIGN

The database design methodology will be used to translate the logical database

design (the entities, attributes, relationships, and constraints) into a physical

database which can be implemented using the target database management system

(DBMS). The physical design will specify the system usability such as screen

layout and patient menu structure. In order to validate the derived relations and to

ensure that they are correctly structured and implemented, a technique of

normalization will be applied. This technique begins by examining the

relationships (called functional dependencies) between attributes. This phase will

produce the workable table which will create SQL statements to be used in the

database through implementation stage.

47

Table 4.2: Patient_Details

Name Type Length Scale Allow Null

Id Int 0 0 N

UserName Varchar 100 0 N

Password Varchar 20 0 N

Firstname Varchar 20 0 N

Lastname Varchar 30 0 Y

Title Varchar 30 0 Y

Gender Varchar 10 0 Y

Birthdate Varchar 15 0 Y

Email Varchar 40 0 Y

Contactaddres

s

Varchar 100 0 N

marital_status Varchar 30 0 Y

Phone Varchar 20 0 N

Nextofkin Varchar 40 0 Y

Noxtofkinpho

ne

Varchar 20 0 Y

date_registere Varchar 20 0 Y

48

d

Thumbnail Varchar 200 0 Y

Nationality Varchar 20 0 Y

Table 4.3: User_Account

Name Type Length Scale Allow Null

UserName Varchar 50 0 N

Password Varchar 20 0 N

Status Varchar 30 0 N

Locked Varchar 30 0 Y

Privilege Varchar 30 0 Y

Table 4.4: Treatment_Notes

Name Type Length Scale Allow Null

NoteId Int 0 0 N

Note Varchar 20 0 N

DateAdded Date 10 0 N

EmployeeId Varchar 30 0 N

Table 4.5: Sensor_Type

49

Name Type Length Scale Allow Null

SensorId Varchar 10 0 N

sensorType Varchar 20 0 N

Description Varchar 10 0 N

Message Varchar 30 0 N

50

Table 4.6: Sensor_Reading

Name Type Length Scale Allow Null

ReadingId Int 0 0 N

ReadingTime Date/Time 20 0 N

Table 4.7: Sensor_ReadingRange

Name Type Length Scale Allow Null

RangeId Int 0 0 N

YellowUpper Double 20 0 N

YellowFloor Double 10 0 N

GreenUpper Double 10 0 N

GreenFloor Double 10 0 N

51

CHAPTER FIVE

SUMMARY AND RECOMMENDATION

5.1 SUMMARY

This study was about a web-based Clinical decision support system for diagnosis

and prescribe treatment for type 1 diabetes based on blood sugar level and insulin

level. The developed system consists of a user interface through which the system

receives data as input and makes result of the processing available to the user. The

system also has a medical knowledge base that consist the diagnostic module, the

medical knowledge module and the electronic medical records. This system has a

classification of algorithm section and an interface engine that searches through the

knowledge base to match each set of input parameters with the right rule as coded

in the knowledge base.

5.2 RECOMMENDATION

The study recommends that the web-based clinical decision support system should

be developed using windows 7 operating system, Apache web server, MySQL

database and PHP, myAdmin, Google Chrome web brower. Cascading Style Sheet

(CSS), Adobe Dreamweaver CC2015 and Front and Back end JavaScript

framework.

52

5.3 FUTURE RESEARCH

For further studies, researchers can improve on the proposed clinical decision

support system by employing more than one efficient algorithm to develop a

hybrid system.

53

REFERENCES

Abidi, S., Cox, J., Abidi, S., & Shepherd, M. (2012). Using OWL ontologies for

clinical guidelines based comorbid decision support. System Science

(HICSS), 2012 45th Hawaii International Conference on, IEEE.

Ahmadian, L., van Engen-Verheul, M., Bakhshi-Raiez, F., Peek, N., Cornet, R., &

de Keizer, N.F. (2011). The role of standardized data and terminological

systems in computerized clinical decision support systems: literature review

and survey. International journal of medical informatics, 80(2): 81-93.

Ahn, J. T., Park, G.H., Son, J., Lim, C.S., Kang, J., Cha, J., & Park, K. (2014).

Development of test toolkit of hard review to evaluate a random clinical

decision support system for the management of chronic adult diseases.

Wireless Personal Communications, 79(4): 2469-2484.

Bau, C.T., Chen, R.C., & Huang, C.Y. (2014). Construction of a clinical decision

support system for undergoing surgery based on domain ontology and rules

reasoning, Telemedicine and e-Health, 20(5), 460–472.

Bouamrane, M.M., Rector, A., & Hurrell, M. (2009). A hybrid architecture for a

preoperative decision support system using a rule engine and a reasoner on a

clinical ontology. Web Reasoning and Rule Systems, Springer: 242-253.

54

Centers for Disease Control and Prevention. National diabetes statistics report

(2017). Centers for Disease Control and Prevention

website. www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-

statistics-report.pdf

Chen, J.Q & Lee, S.M. (2002). An exploratory cognitive DSS for strategy decision

making Elsevier Science B.V., 8(6), 13-32.

Chen, R.C., Huang, Y.H., Bau, C.T., & Chen, S.M. (2012). A recommendation

system based on domain ontology and SWRL for anti-diabetic drugs

selection, Expert Systems with Applications, 39(4), 3995–4006.

Cho, N. H., Shaw, J. E., Karuranga, S., Huang, Y., da Rocha Fernandes, J. D.,

Ohlrogge, A. W., & Malanda, B. (2018). IDF Diabetes Atlas: Global

estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes

research and clinical practice, 138, 271-281.

Classen, D. C., Phansalkar, S., &Bates, D.W. (2011). Critical drug-drug

interactions for use in electronic health records systems with computerized

physician order entry: review of leading approaches. Journal of patient

safety, 7(2): 61-65.

55

Dejan, D. U. B. (2014). Clinical Decision Support Systems. Retrieved from

intechopen: http://cdn.intechopen.com/pdfs-wm/16881.pdf.

Elhendi, M. Y. (2015). Assessment of Type 2 Diabetes management practice: A

study in public hospitals outpatient clinics, Khartoum and Gezira,

Sudan (Master's thesis).

Farooq, K., & Hussain, A. (2016). A novel ontology and machine learning driven

hybrid cardiovascular clinical prognosis as a complex adaptive clinical

system. Complex Adaptive Systems Modeling, 4(1): 1-21.

Farooq, K., Khan, B.S., Niazi, M.A., Leslie, S.L., & Hussain, H. (2017). Clinical

Decision Support Systems: A Visual Survey. Advances in Brain Inspired

Cognitive Systems, Springer: 1-27.

Greenes, R.A. (2014). Clinical decision support: the road to broad adoption. 2nd

ed. Amsterdam Boston: Academic Press. Retrieved from:

https://asu.pure.elsevier.com/en/publications/definition-scope-and-

challenges.

Haynes, R. B., & Wilczynski, N.L. (2010). Effects of computerized clinical

decision support systems on practitioner performance and patient outcomes:

Methods of a decision-maker-researcher partnership systematic review.

Implement Sci, 5(1), 12.

56

International Diabetes Federation (IDF) (2018). Diabetes Atlas: Global estimates

of diabetes prevalence for 2017 and projections for 2045. Retrieved April 4,

2019 from https://www.diabetesresearchclinicalpractice.com/article/S0168-

8227(18)30203-1/fulltext

Ivbijaro, G., Kolkiewicz, L., McGee, L., & Gikunoo, M. (2008). Addressing long-

term physical healthcare needs in a forensic mental health inpatient

population using the UK primary care Quality and Outcomes Framework

(QOF): an audit. Mental health in family medicine, 5(1): 51.

Jaspers, M. W., Smeulers, M., Vermeulen, H., & Peute, L.W. (2011). Effects of

clinical decision-support systems on practitioner performance and patient

outcomes: a synthesis of high-quality systematic review findings. Journal of

the American Medical Informatics Association, 18(3): 327-334.

Jimoh R.G, A. J. (2014). Simulation of Medical Diagnosis System for Malaria

Using Fuzzy Logic. Cisdiar, 20-25. Retrieved from

http://cisdijournal.net/uploads/V5N2P6_CISDIAR_JOURNAL.pdf.

Kawamoto, K., Del Fiol, G., Orton, C., & Lobach, D.F. (2010). System-agnostic

clinical decision support services: benefits and challenges for scalable

decision support. The open medical informatics journal, 4: 245.

57

Kuperman, G. J., Bobb, A., Payne, T.H., Avery, A.J., Gandhi, T.K., Burns, G.,

Classen, D.C., &Bates, W. (2007). Medication-related clinical decision

support in computerized provider order entry systems: a review. Journal of

the American Medical Informatics Association, 14(1): 29-40.

Kuperman, G. J., Gardner, R.M., & Pryor, T.A. (2013). HELP: a dynamic hospital

information system, Springer Science & Business Media.

Kuyoro, S., Osisanwo, F., & Franklyn, A. (2014). A Web-Based Clinical Decision

Support System for the Management of Diabetes Neuropathy Using Naïve

Bayes Algorithm. European Journal of Computer Science and Information

Technology, 6(2), 30-41.

Kuyoro, S., Osisanwo, F., & Franklyn, A. (2018). A Web-Based Clinical Decision

Support System for the Management of Diabetes Neuropathy Using Naïve

Bayes Algorithm. European Journal of Computer Science and Information

Technology, 6(2), 31-41.

Marco-Ruiz, L., & Pedrinaci, C., Maldonado, J.A., Panziera, L., Chen, R., &

Bellika, J.G. (2016). Publication, discovery and interoperability of clinical

decision support systems: a linked data approach, Journal of Biomedical

Informatics, 62, 243–264.

58

Moja, L; Kwag, KH; Lytras, T; Bertizzolo, L; Brandt, L; Pecoraro, V; Rigon, G;

Vaona, A; Ruggiero, F; Mangia, M; Iorio, A; Kunnamo, I; Bonovas, S

(2014). "Effectiveness of computerized decision support systems linked to

electronic health records: a systematic review and meta-analysis". American

Journal of Public Health. 104 (12): e12–22. doi:10.2105/ajph.2014.302164.

Musen, M. A., Middleton, B., & Greenes, R.A. (2014). Clinical decision-support

systems. Biomedical informatics, Springer: 643-674.

NCD Risk Factor Collaboration (NCD-RisC) (2016). Worldwide trends in diabetes

since 1980: a pooled analysis of 751 population based studies with 4*4

million participants. Lancet 1980; 2016. https://doi.org/10.1016/S0140-

6736(16)00618-8

O’Connor, P. J., Sperl-Hillen, J.M., Rush, W.A., Johnson, P.E., Amundson, G.H.,

Asche, S.E., Ekstrom, H.L., & Gilmer, T.P. (2011). Impact of electronic

health record clinical decision support on diabetes care: a randomized trial.

The Annals of Family Medicine, 9(1): 12-21.

Ogurtsova, K., da Rocha Fernandes, J. D., Huang, Y., Linnenkamp, U.,

Guariguata, L., Cho, N. H., ... & Makaroff, L. E. (2017). IDF Diabetes Atlas:

Global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes

research and clinical practice, 128, 40-50.

59

Rouse, M. (2014). Clinical decision support system (CDSS) definition. Retrieved

from techtarget: http://searchhealthit.techtarget.com/definition/clinical-

decision-supportsystem-CDSS.

Rung-Ching, C., Hui, Q. J., Chung-Yi, H., & Cho-Tsan, B. (2017). Clinical

Decision Support System for Diabetes Based on Ontology Reasoning and

TOPSIS Analysis. Journal of Healthcare Engineering,

https://doi.org/10.1155/2017/4307508

Sakiko, O., Kaoru, M., Aizan, H., Yoshihito, N., & Hiroshi, T. (2018).

Improvements in Diabetic Patients’ Outcomes in a Clinical Decision Support

System. EJBI, 14(1), 30-36.

Sherimon, P.C., & Krishnan, R. (2016). OntoDiabetic: an ontology based clinical

decision support system for diabetic patients, Arabian Journal for Science

and Engineering, 41(3), 1145–1160.

Wong, A., Bright, T. J., Dhurjati, R., Bristow, E., Bastian, L., Coeytaux, R.R.,

Samsa, G., Hasselblad, V., Williams, J.W., & Musty, M.D. (2012). Effect of

clinical decision-support systems: a systematic review. Annals of internal

medicine, 157(1): 29-43.

Wright, A., Sittig, D.F., Ash, J.S., Bates, D.W., Feblowitz, J., Fraser, G., Maviglia,

S.M., McMullen, C., Nichol, W.P., Pang, J.E. (2011). Governance for

60

clinical decision support: case studies and recommended practices from

leading institutions. Journal of the American Medical Informatics

Association, 18(2), 187-194.

Wright, A., Sittig, D.F., Ash, J.S., Bates, D.W., Feblowitz, J., Fraser, G., Maviglia,

S.M., McMullen, C., Nichol, W.P., & Pang, J.E. (2011). Governance for

clinical decision support: case studies and recommended practices from

leading institutions. Journal of the American Medical Informatics

Association, 18(2): 187-194.

Zhang, Y, F., Tian, Y., Zhou, T.S., Araki, K., & Li, J.S. (2016). Integrating HL7

RIM and ontology for unified knowledge and data representation in clinical

decision support systems. Computer Methods and Programs in Biomedicine,

123, 94–108.

Zuccotti, G., Maloney, F., Feblowitz, J., Samal, L., Sato, L., & Wright, A. (2014).

Reducing risk with clinical decision support. Appl Clin Inform, 5(3): 746-

756.