ABSTRACT - Wayne State University

Babbaljeet Kaur ([email protected]), Balraj Inder Kaur Mahal ([email protected])
Shashank Kamthan ([email protected]), Wayne State University, Detroit, MI, USA
ABSTRACT
Diabetes is one of the major health problems concerning the world today. According to National Diabetes Statistics Report, 29.1 million people in United States of America are diabetic out of which 21 million
people are diagnosed while 8.1 million are still undiagnosed. Though many advancements and recent updates keep coming in field of diabetes, there is a critical need to make more people aware of the risk factors associated with it. We did a survey on general population for risk factors of diabetes and developed a system using fuzzy logic i.e., Sugeno Type fuzzy inference system (7x1) We have made an attempt to device the system which can serve as a combined tool for risk factor assessment in patients. Our system is to make general population assess their risk factors themselves for diabetes and to see their primary care
physician for diabetes screening. Our aim is to motivate high risk population for early Diabetes Mellitus Screening and to educate people about risk factors of diabetes
.
INTRODUCTION

METHOD AND SURVEY
Diabetes Mellitus is a disorder which is characterized by increased blood sugar (glucose) level resulting from defects in insulin secretion, insulin action, or both.
Type 1 DM

Type 2 DM

autoimmune condition with severe insulin deficiency
the body develops resistance to insulin. Accounts for
due to permanent destruction of insulin producing cells >90% cases of diabetes. Gestational Diabetes: any deof pancreas. Has a Genetic predisposition.
gree of glucose intolerance with onset or diagnosis during pregnancy.

We started with seven risk factors of diabetes with aim of developing a system using fuzzy logic which can be
used by patients themselves to assess their own risk factors of Diabetes Mellitus and reach their primary care
physician for diabetes screening.
We did a survey on general population regarding seven risk factors, from which we developed a system for diabetes risk factor stratification using fuzzy logic by defining the rules.
Diabetes Mellitus is present in 11.3% of persons over age of 20 years in the United States, and 26.9% of those
over the age of 65 years. Figure shows the increase in diabetics to 29.1 million from 25.8 million in 2010, according to National Diabetes Statistics Report, 2014
MATLAB APPROACH USING FUZZY LOGIC






Define the linguistic variables and terms (initialization)
Construct the membership functions (Initialization)
Construct the rule base (initialization)
Convert input data to fuzzy values using membership functions
Evaluate the rules in the rule base (inference)
Convert the output data to non fuzzy values
Diabetes is predicted to become the 7th leading cause of death
in world by the year 2030 according to WHO. 90% of cases are Type 2 DM which is preventable with lifestyle
modifications.

RISK FACTORS AS INPUTS
Family History of Diabetes Mellitus
Risk increases if parent or sibling has Type 2 DM or if one has Prediabetes which is a precursor to Type
2 DM (Blood Sugar more than normal but less than definition of Diabetes)
Obesity
About 50% of men and 70% of women who have diabetes are obese. Body Mass Index which is weight/
square of height, if more than 30 indicates obesity. .
Lack of Physical Activity / Exercise
Life style modification, including a balanced hypocaloric diet to achieve 7% weight loss in overweight
patients and regular exercise of > 150 minutes per week, is recommended for persons with prediabetes to
prevent progression to type 2 DM.
Age more than 45 years
Increasing age increase risk of diabetes.
Dyslipidemia/ Abnormal lipid level
Increased blood cholesterol >240 mg/dl increases risk of diabetes. mellitus and cardiovascular disease.
Total cholesterol should be <150 mg/dl.
Hypertension
Polyuria
chronic high blood pressure>140/90 mm hg increases risk to diabetes, heart attack, stroke. Blood pressure target in diabetics is <130/80 mm hg.
increased blood sugar spills in urine and kidneys secrete excess water to dilute excess sugar resulting in
increased frequency of urination.
1. Mamdani Model
2. Rule Editor
3. Rule Viewer
4. Surface Viewer
RESULTS & ANALYSIS



Correlation factor is 0.816 via Sugeno Inference System
System was developed through which a patient would assess his own risk factors, get his probability of getting
diabetes and should see his primary care for early diabetes screening.
CONCLUSION & FUTURE WORK





The system needs further and larger evaluation to establish its safety and efficacy.
To get input from different physicians to define rules for developing a system using fuzzy logic.
To do more surveys on populations with different age, race, ethnicity.
To test the system on larger number of population.
To make a standardized system which can be used by everyone for their risk factor stratification.