CHAPTER 6

AN EFFECTIVE DFUP APPLICATION –

ALTERNATIVE TO THE MANUAL PROCESS

6.1 INTRODUCTION

This phase of research deals with the replacement of manual process by

Diabetes Foot Ulcer Prediction (DFUP) Application. It integrates DFUP model with
an application. There are many ways to integrate prediction model with an
application such as web application, mobile application and desktop application.
Figure 6.1 shows the login workflow for DFUP integration.

Figure 6.1 DFUP Model for Login workflow

110
6.2 WEB APPLICATION

The web application is an application program that is stored in the remote

server and can be accessed through web browsers such as Google chrome, Internet
explorer and Mozilla Firefox etc. To develop the web application minimum
technical knowledge is required. The web application can be used to access a web
server via the internet and intranet. Java 8 is used for design, coding and validation
purpose.

In addition, for security authentication OAuth validation is used to secure the

login and data validation. User Interface/Web page design used to HTML, Angular
JS, and JavaScript are used. The advantages of the web application are

 Development is cost effective

 Can be accessed anytime

 Handles and supports several devices

 More Security

CSS (Cascading Style sheet), HTML (Hyper Text Markup Language) has
been used for the user interface, design, validation, visualization, color coding and
other alignment purpose. Ajax (Asynchronous JavaScript XML) is architecture but
not a technology. The main features of Ajax are

 Interaction using DOM which helps to display dynamically

 Data can be exchanged using XML and XSLT

 User Fridley

 Ajax will help to make web application in faster

 Ajax will support data binding

 Cost effective for development

 Easy to create and develop the web application

 Performance is good

111
  DFUP applications are run on the system own web browser via a
simple URL(Uniform Resource Locator)

 Less time for web application development

 User Interface design is good

 Reports are dynamically generated

 No need to spent any amount or fees for third-party registration

The limitation of the web applications are

 Availability of the application is less compared to mobile

 To run and execute the application require, an internet connection

 Application are not centralized and each application are stored

separately

 Applications can be slow when compared to other native application

 Screen size is big when compare to native application

6.3 MOBILE APPLICATION

Mobile applications are available in online and mobile store as well. Mobile
application development is high when compared to web applications. However,
performance and usage is more effective when compared to web application and
desktop applications. Mobile applications are stable with the hardware of mobile
devices, camera and GPS System.

Zhao et al. (2015) developed thousands of medical health care applications

using Android and iOS platform with free of cost. More than 50,000 mobile
applications are available in the field of healthcare in the world. There are mobile
applications that contribute in the monitoring of the health conditions of the patients
after the treatment like Diabetes amputation, heart operation and removal of cancer
cells. The mobile applications are used for an early identification of Diabetes Type I
or Type II, Heart Rate, Pulse point, and heart disease through iOS and android
operating system. The mobile app “Count of heart beat” has 33,000 registered users.
112
Taigang et al. (2013) examined the efficiency of the existing mobile application,
which supports the usage of mobile apps in disease diagnosis.

To develop the mobile application some set of processes and procedures

should be followed. The operating systems which are available in the market for
mobile application development are

 iOS (Operating System)

 Android OS(Operating System)

 Windows Phone

The main advantages of the mobile applications are

 Code reusable with multiple platform

 Performance is more when compared to web application

 Support distributed environment

 Access from any mobile device

 Availability of the application

The limitations of the mobile applications are

 High cost for the development of the mobile application

 Application development time is more

 The development of the mobile application requires technology skills

 The development of Mobile Application is quite Complex

 Mobile devices having the less storage and size

6.4 METHODOLOGY – DFUP MOBILE APPLICATION

Application development Java, J2EE, HTML, CSS, OAutho2, Angular JS,

Node JS, Struts framework, Java Script, R & R-Studio, Ajax, Application Server,
SQL Database, Spring boot, OAuth 2 and single sign on model (SSL) can be used.

113
DFUP MODEL
The outcome of the DFUP model which is mentioned in chapter 5 is used to
integrate with the “DFUP mobile application”. Based on the analytics using data
mining tool R and R-studio with the help of data mining classification algorithm, the
main important parameters are decided for the type of medical treatment to be
performed on the patient. If all the main significant attributes persist as a condition
for the patient, they have to go for an amputation. If all the significant attributes are
satisfied for the patient, it leads to amputation. If anyone of the attribute is not
satisfied, then leads to multidisciplinary approach.

DFUP can be easily treated by a multidisciplinary approach like glycolic

control, good education about wound dressing and diet. Early surgical wound
debridement helps to heal the wound faster and prevent rehabilitation. It is wise to
advise the patient with DFU to do multidisciplinary approach which prevents
permanent disability.

Model View controller (MVC) framework has been used for integration.
MVC is called one type of design pattern which helps to develop the web
application and mobile application. Model is holding all the data related logic which
required for client and also the model will act and transferred the data between view
and controller. Model is the central component of the design pattern. Model will
manage the data, rules of the application and main logic of the applications. View
components can be used for representation purpose such as Chart, diagram and flow
diagram. View is known as the user interface related application logic which is
present in the view components. Controller is a main component in this application.
Controller acts between view and Model.

The MVC (Model View controller) framework workflows for DFUP

application model are

 Client or user sends the request to Controller

 Controller will validate the request, if the request is valid then do the
business process to get the right business rule. Based on the rules the
request is forwarded to the model.
114
  The rules are validated by the model. Based on the rules, the data has
been fetched from the database.

 Model components will send the response to the view components

 User will be able to get the required information from DFUP Model

 DFUP application will have the proper approach for Diabetes patient

Load balancing deals with sharing network load. A load balancer is called as
distributes network. Load balancing helping in application traffic across a cluster of
servers to improve the performance and availability of the applications. Many load
balancing algorithms are available in the world. In this development the load
balancer algorithms such as Round Robin Load balancer, Least Connections Load
balancer and IP Hash Load balancer have been used.

Hardware Load balancer and Software Load balancer are the two types of
load balancer. In this research, the code of DFUP model has been stored in the GIT
Repository. GIT is an open source tool and the code can be accessed from anywhere
in the world. Figure 6.2 shows the DFUP Application solution architecture.

Figure 6.2 Architecture of DFUP Application

115
 For security and validation purpose, login and Authentication page has been
created. User can enter the login information from the application and it has been
validated. If it is correct then it allows the user to access the DFUP application. If
entered password or user id is wrong, then it won’t allow the user to access the
applications and valid response will be sent to the user.

6.5 RESULTS AND DISCUSSION

DFU can be applied through multidisciplinary approach like strict glycemic

control, good education about wound dressing and diet. Early surgical wound
debridement helps to heal the wound faster and prevent rehabilitation. Hence it
recommends the patient with DFU to do multidisciplinary approach which initially
prevents permanent disability.

The screens and pages of DFUP mobile application to predict the Diabetes
foot ulcer Type II are

 DFUP - Login Page

 DFUP – Diabetes Screening System

 DFUP – Diabetes Screening System calling Machine Learning Tool

 DFUP - Prediction Model

DFUP application login screen has two categories such as Doctor Login and
Patient Login. To login the DFUP application, user needs to create User ID and
password using the New Login registration button. After creating the new login
details for a particular patient, the patient can able to login. Figure 6.3 shows the
login page with new login registration facility.

Figure 6.3 DFUP New Patient Login Registration

116
 Using patient login, patient can enter their test report details such as Patient
id, Name, Age, Gender, Family History, Diabetes, BMI(Body Mass Index),
Diabetes, HbA1c,FS(Fasting Blood Sugar Test), PS(Post Blood Sugar Test),
Alcohol, Smoking, Single/multiple ulcer, Glycemic index and Etiology in the screen
“Star Diabetes Hospital - DFUP Application -Diabetes Screening System”. After
entering all the test data in the screen, click the “ADD DIABETES DATA” button.
Now the data will be stored in the DFUP database. Patient login credentials are
designed in such a way where the patient can view his / her data but cannot access
another patient’s data. The doctor can able to see the entire patient’s data though his /
her login. Figure 6.4 shows the sample entry in the DFUP application for the primary
attributes of clinical dataset.

Figure 6.4 DFUP Sample Data Set for Primary Attributes Validation

In doctor’s login, doctor has the access to view patient’s profile, test reports
which are entered and stored in the systems. Based on the test reports, doctor will
do the process and execute the operation.

117
 If the doctor wants to know whether the patient profile is Low Risk or High
Risk then click the “Liquid profile” button and the application will automatically do
the validation based on the test report and will throw the message like “Low Risk”
or “High Risk”.

Figure 6.5 shows the validation of the profile information based on the risk
values like Low Risk or High Risk.

Figure 6.5 DFUP Profile Validations

118
 The Doctor will proceed the further execution based on the outcome, by
clicking the “Validation” button in the DFUP application. The options available in
the DFUP - Calling Analytic Applications’ screen are

 Machine learning tools for prediction – Two machine learning tools are
integrated in the system, based on the request need to select tool like “R”
or “Weka”.

 DFUP Application – Developed application, can access two ways

o Mobile application
o Web application

 To call the DFUP model - The DFUP model has all the business logic,
which is covered in chapter 5. All the business prediction logic written
using machine learning tool.

After entering all the required values, click on DFUP PREDICTION button.
The request will be redirected to the business model which is developed as a part of
this research in chapter 5. It also validates the business rule.

Figure 6.6 shows the DFUP – Calling Analytic Application. Based on the
validation, the model will return to the approaches to be opted to continue further.
The options are Multidisciplinary and Amputation. When all the predefined business
rules are validated with the test reports and any one of the test reports are negative
for the patient, they will decide to go for prediction on Multidisciplinary. For this
execution, 448 patients will recommend and decide to go for prediction on
Multidisciplinary approach.

Figure 6.6 DFUP Model – Prediction

119
 When all the predefined rules are validated with the test reports and all the
main significant attributes persist as a condition for the patient, they will decide to
go for prediction on Amputation. For this execution, 8 patients will recommend and
decide to go for prediction on Amputation approach. Figure 6.7 shows the output of
DFUP model.

Figure 6.7 DFUP Model – Amputation - Multidisciplinary

SUMMARY
A prediction model has been developed using machine learning tool R, and
implemented as a web application and mobile application. These applications was
developed using Java, Angular JS, CSS, HTML and spring boot with GIT. Mobile
applications are the reason for a growth and modern revolution in health care
department which permits the healthcare specialists to quickly improve patients'
facilities. DFUP model has been built using the machine learning tool “R”.

Based on the experiment, among 456 Diabetes patients’ data, 8 patients have
to go for an amputation approach and 448 have to go for a Multidisciplinary
approaches.The “DFUP model” which has been designed as an alternative for
manual process has been implemented as web application and mobile applications.
The DFUP application has to be uploaded in play store. Hence doctors, physicians
and patients can easily download it.

120