Received: 12 April 2019

DOI: 10.1002/mma.5681

RESEARCH ARTICLE

A study of the fractional-order mathematical model of

diabetes and its resulting complications

Hari M. Srivastava1,2 Ravi Shanker Dubey3 Monika Jain4

1
Department of Mathematics and
Statistics, University of Victoria, Victoria, Diabetes is a worldwide problem that affects one of every 11 persons nowadays.
British Columbia, Canada The IDF Diabetes Atlas (Eighth edition, 2017) states that approximately 415
2
Department of Medical Research, China million people in the world are living with the disease and that this number will
Medical University Hospital, China
Medical University, Taichung, Taiwan, rise to 629 million by the year 2045. It is a very serious problem of the world. A
China major part of the world population is affected by this disease and its resulting
3
Department of Mathematics, Amity complications. In this paper, we propose to investigate a fractional-order model
School of Engineering and Technology,
of diabetes and its resulting complications. The mathematical model's parame-
Amity University, Jaipur, India
4
Department of Mathematics, JECRC
ters define the population of diabetic patients and those who are diabetic with
University, Jaipur, India complications at a given time t. We have also discussed the existence, unique-
ness, and stability of the fractional-order model, which we consider here. We
Correspondence
Hari M. Srivastava, Department of make use of the homotopy decomposition method (HDM) in order to solve the
Mathematics and Statistics, University of problem.
Victoria, Victoria, British Columbia V8W
3P4, Canada. K E Y WO R D S
Email: harimsri@math.uvic.ca
diabetes and its resulting complications, existence, fractional calculus, fractional-order modeling,

Communicated by: W. Sprößig homotopy decomposition method (HDM), uniqueness and stability

MSC Classification: 92B05; 92C60; 26A33

1 INTRODUCTION, BAC KG ROUND, AND PRELIMINARIES

Diabetes mellitus (DM), which is also known as a silent killer in medical science, is a metabolic disorder that occurs when
the sugar level in blood is not appropriate. The glucose in human blood is the main source of energy and comes from
the food insulin, a hormone made by the pancreas, which helps glucose from food to get into our cells and to use it for
energy. Insulin also controls and adjusts the quantity of sugar in human body.1 When a human body is unable to make
the desired amount of insulin or is not able to use it well, then the glucose does not reach the cells and stays in the blood.
Frequent urination (polyurea), feeling more thirsty and hungry (polydipsia and polyphagia) are the symptoms that the
patients having high blood sugar typically experience.
Diabetes is known as a Raj Rog in ancient India, that is, a disease which affects those people having lots of wealth and
live a comfortable prosperous life by using servants for doing their works and chores. But, in recent years, it has become
a problem that affects the whole society. In ancient times, this problem was seen in the age group above 70 years, but, in
today's era, it is a problem of all age groups. And, in very short time period, it has become a problem of the whole world.
The issue is very serious and researchers must contribute their best to control this problem.
Generally, diabetes is considered to be of the following two types.
Type I diabetes: Type I diabetes requires a daily dose of insulin taken by the patient to regulate the amount of
glucose in blood. Nonaccess of insulin leads to certain complications or danger to the patient's life. The cause of

4570 © 2019 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/mma Math Meth Appl Sci. 2019;42:4570–4583.
SRIVASTAVA ET AL. 4571

TABLE 1 Affected people by diabetes country-wise

Name of the Country Total Population People Affected by Diabetes
People's Republic of China 1 411 316 334 109.6 million
India 1 339 180 127 69.2 million
USA 324 459 463 30.3 million
Brazil 209 288 278 14.2 million
Russian Federation 143 981 725 12.5 million

the type I diabetes is not known as yet and is currently not preventable. Symptoms include excessive and frequent
urination and thirst, constant hunger, weight loss, vision change, and fatigue.
Type II diabetes: In this type of diabetes, we consider those patients who do not depend on insulin. For type II dia-
betes, the doctor recommends the patient to follow a particular diet chart and exercise. By means of this restriction,
one can control one's type II diabetes. Yoga also plays a very efficient role in order to control the type II diabetes.
Symptoms of the type II diabetes may be similar to those of the type I diabetes but are often less marked or absent.
Diabetes is steadily increasing everywhere not only in adults but also in children most markedly in the world's
middle-income countries.2 There is a prediction on the basis of the increasing number of diabetes patients that, in
the world, the diabetes cases can jump to 55% in 2035.

The International Diabetes Federation (IDF) had recently given the information to the effect that the top five countries
in the world have the highest numbers of diabetes patients (see Table 1).
There are some other countries also that have the highest proportion of people living with diabetes such as Tokelau
(29.7%), Mauritius (24.3%), Nauru (23.8%), Cook Islands (21.1%), and Marshall Islands (21.1%) (see, for details, Whiting
et al2 ).
Complications: All types of diabetes can lead to various diseases and health problems such as, for example, heart
attack, kidney failure, lower limb amputation, blindness, and nerve damage. In fact, diabetes increases the overall
risk of dying prematurely. For a pregnant woman, uncontrolled diabetes increases the risk of fetal death and other
complications (see other studies 3-15 ).
Economic impact: According to a new research conducted at the University of Göttingen and at the Harvard T. H.
Chan School of Public Health, which was published online on April 26, 2017, in The Lancet Diabetes & Endocrinol-
ogy, the world faces a notable economic burden from diabetes, that is, about 1.8% of global gross domestic product
(GDP). The study also found that, not only the direct costs of the disease such as (for example) expenditures for
insulin, testing strips, and treatment but also the indirect costs representing production shortfalls due to illness
and premature death accounts for nearly 35% of the total economic burden.
Prevention of diabetes: Type II diabetes can be prevented by losing weight and following a regular healthy diet and
exercise chart, but type I diabetes cannot be prevented with our current knowledge till today. Resistance training
(twice a week) and regular aerobic exercise maintain blood sugar level and are very effective in controlling diabetes.
WHO and FAO provide guidance for diet to prevent type II diabetes (see WHO/FAO Expert Consultation10 ). WHO
also recommends different physical activities for different age groups (see WHO Report16 ), because physical activity
and diet are significantly more effective than medication.
Diabetes is diagnosed by the glucose tolerance test (GTT) in which glucose level is measured by blood sample,
while the patient is in a fasting state. Diagnosis of diabetes of a patient in nonfasting state can be done by the
measurement of glycated hemoglobin (HbA1c), which reflects the average blood glucose concentration. However,
this test is more expensive than blood glucose measurement.

2 MATHEMATICAL MODELING OF THE PRO BLEM

In this section, we describe the mathematical model of diabetes and its resulting complications, which was considered in
the year 2004 by Boutayeb et al.17 This model is capable of studying diabetic patients with and without complications. In
this model, C(t) and B(t) represent patients with and without complications, respectively, at time t. We assume that the
total number of diabetic patients at a time t is E(t), so that
E(t) = B(t) + C(t).
4572 SRIVASTAVA ET AL.

FIGURE 1 An overview of the flow chart [Colour figure can be viewed at wileyonlinelibrary.com]

In order to understand the mathematical model, we present the flow chart (see Figure 1). In this model, we use the
following parameters.
A(t): The incidence of diabetes mellitus at time t;
B(t): The number of persons having diabetics without complications at time t;
C(t): The number of persons having diabetics with complications at time t;
E(t): The size of the population of diabetics at time t;
𝛿: The probability of a person having diabetes and developing complications;
𝜀: The natural rate of mortality;
𝜆: The rate of complications recovered;
𝜇: The rate of mortality due to diabetic complications;
𝜈: The rate of diabetic patients having complication and becoming severely disabled.

In the flow chart (see Figure 1), we consider A(t) patients of diabetes and diagnosed for some time interval t and without
any resulting complications diagnosed. In this chart, 𝛿B(t) represent the patients who have developed complications, 𝜆C(t)
represent the patients whose complications are recovered, 𝜈C(t) represent those patients who become strictly disabled
and whose disabilities cannot be recovered, 𝜇C(t) represent those patients who die from their complications, and C(t)
represent the disabilities resulting from complications. We are thus led to the following ODEs:
d
[B(t)] = A(t) − (𝛿 + 𝜀)B(t) + 𝜆C(t) (1)
dt
and
d
[C(t)] = A(t) − (𝜆 + 𝜇 + 𝜈 + 𝜀)C(t) + 𝛿B(t). (2)
dt

Also, as indicated above, we have

E(t) = B(t) + C(t), (3)
where E(t) represent the size of those persons who are diabetic at time t.
By using the relation (3), we get
d
[C(t)] = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (t > 0) (4)
dt
and
d
[E(t)] = A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t) (t > 0), (5)
dt
SRIVASTAVA ET AL. 4573

together with
C(0) = C0 and E(0) = E0 . (6)

Here, in our investigation, we have followed the standard formulation of the model of Boutayeb et al17 represented by
following the system of differential equations in WHO/FAO Expert Consultation,10 where
𝜑 = 𝜆 + 𝜇 + 𝜈 + 𝜀.
The structure of this paper is as follows. In the next section (Section 3), we provide a brief overview of the fractional
calculus and its application. In Section 4, we introduce and study a fractional-order model of diabetes and its complication,
uniqueness of the coupled solutions, and existence of solution of coupled solution. In Section 5, we derive the approximate
solutions of the fractional-order model of diabetes and its complication. We also discuss the solution of the first-order
model of Boutayeb et al.17 In Section 6, we derive the numerical results and show the efficiency of fractional calculus in
our study. Finally, in our last section (Section 7), we present some concluding remarks and observations relevant to our
investigation.

3 A N I N T RO DU CTO RY OVE RVIEW OF FRACTIONAL C ALC ULUS

In the study and analysis of various real-world problems, the subject of fractional calculus has played a significant role in
their mathematical modeling. It is capable of modeling and analyzing linear as well as complex phenomena of engineering
and applied sciences. In modern time, fractional calculus plays very essential role in such fields as (for example) biology,
geophysics, biomedical engineering, fluid dynamics, electricity, and mechanics (see other works18-26 ). It has thus become a
very efficient tool to describe significant models of the natural phenomena. Many types of problems can be defined by the
help of fractional-order integration and fractional-order differentiation. These include, but are not limited to, stochastic
processes, blood glucose insulin model, Brownian motion, fractional oxygen diffusion problem, control theory, viscoelas-
ticity, fractals and nonlocal phenomena, time fractional telegraph equation, space-time fractional Fokker-Plank equation,
fractional LC-circuit models, fractional Klein-Gordon equations, heat conduction, image and signal processing, control
theory, and controllability of fractional delay dynamical systems (see, for details, other studies27-35 ).
Many mathematical scientists and researchers have been working in this field, such as (for example) those who are
included in the list of our citations, by introducing various fractional calculus models (see other studies20-45 ).
One can find a great deal of literature on fractional calculus and its wide-spread applications. In our investigation in this
paper, we have chosen to use the familiar Riemann-Liouville operators of fractional integrals and fractional derivatives.
The mathematical method, which we have used to solve the problem considered herein, is the familiar homotopy
decomposition method (HDM), which is already demonstrated to be very efficient method to solve such types of prob-
lems. In recent years, many authors used this method for solving such problems. For instance, we recall the solution of
fractional-order oxygen diffusion equation by Alkahtani et al,33 the solution of fractal partial differential equations arising
in mathematical physics by Yang et al,38 the modeling of the HIV infection of CD4+ cells and attractor one-dimensional
Keller-Segel equations by Atangana and Alabaraoye,40 the boundary value problems of the 2D and 3D poisson and bihar-
monic equations by Atangana,42 the solution of the space-time fractional Fokker-Plank equation by Dubey et al,37 the
solution of the quadratic Riccati differential equation by Odibat and Momani,44 the homotopy perturbation approximate
solutions for Bergman's minimal blood glucose-insulin model by Dubey et al,45 and so on.
Definition 1. The Riemann-Liouville fractional derivative operator D𝛼x of order 𝛼 is defined, for a real-valued
function f(x), by
x
1 d𝓁
D𝛼x [𝑓 (x)] = (x − 𝜏)𝓁−𝛼−1 𝑓 (𝜏) d𝜏 (7)
Γ (𝓁 − 𝛼) dx𝓁 ∫0
(x > 0; 𝓁 − 1 < 𝛼 ≦ 𝓁; 𝓁 ∈ N ∶= {1, 2, 3, … }) .

Definition 2. A real-valued function f(x) (x > 0) is said to be in the space 𝜇 (𝜇 ∈ R) if there exists a real number
p (p > 𝜇) such that f(x) = xp g(x), where g(x) ∈ C[0, ∞). Furthermore, the function f(x) (x > 0) to be in the space 𝜇n if

𝑓 (n) (x) ∈ 𝜇 (n ∈ N). (8)

4574 SRIVASTAVA ET AL.

Definition 3. The Riemann-Liouville fractional integral operator Jx𝛼 of order 𝛼 (𝛼 ≧ 0) is defined, for a real-valued
function 𝑓 (x) ∈ 𝜇 (𝜇 ≧ −1), by
⎧ 1 x 𝛼−1
𝛼 ⎪ Γ(𝛼) ∫0 (x − 𝜏) 𝑓 (𝜏)d𝜏 (x > 0; 𝛼 > 0)
Jx [𝑓 (x)] = ⎨ (9)
⎪ 𝑓 (x) (x > 0; 𝛼 = 0).
⎩

By assuming that
𝑓 ∈ 𝜇 (𝜇 ≧ −1), min{𝛼, 𝛽} ≧ 0 and 𝛾 > −1,

it is easily observed that

Jx𝛽 Jx𝛼 [𝑓 (x)] = Jx𝛽+𝛼 [𝑓 (x)] , (10)

Jx𝛼 Jx𝛽 [𝑓 (x)] = Jx𝛽 Jx𝛼 [𝑓 (x)] (11)

and
[ ] Γ (𝜆 + 1)
Jx𝛼 x𝜆 = x𝛼+𝜆 (x > 0; 𝜆 > −1). (12)
Γ (𝛼 + 𝜆 + 1)

Various properties of the fractional-order operators Jx𝛼 as well as D𝛼x can be found in, for example, other studies18,21,23
(see also Alkahtani et al37 ).

4 FRACTIONA L- ORDER MODEL O F D IABETES AND ITS RESULTING

CO MPLICAT IONS

In this section, by applying Definition 3, we describe the following fractional-order mathematical model of diabetes and
its resulting complications:
D𝛼t [B(t)] = A(t) − (𝛿 + 𝜀)B(t) + 𝜆C(t) (13)
and
D𝛽t [C(t)] = A(t) − (𝜆 + 𝜇 + 𝜈 + 𝜀)C(t) + 𝛿B(t). (14)

Just as above, we also let

E(t) = B(t) + C(t), (15)

where E(t) represents the size of those persons who are diabetic at time t. Indeed, by using the relation (15), we get

D𝛼t [C(t)] = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (t > 0; 0 < 𝛼 ≦ 1) (16)

and
D𝛽t [E(t)] = A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t) (t > 0; 0 < 𝛽 ≦ 1), (17)
together with
C(0) = C0 and E(0) = E0 , (18)
where
𝜑 = 𝜆 + 𝜇 + 𝜈 + 𝜀.

4.1 Existence of the coupled solutions

We make use of the Fixed Point Theory in order to determine the existence of the coupled solution. We first transform
(16) into an integral equation as follows:
t
1
C(t) = C(0) + (t − s)𝛼−1 [−(𝛿 + 𝜑)C(s) + 𝛿E(s)] ds (19)
Γ(𝛼) ∫0
(t > 0; 0 < 𝛼 ≦ 1)
SRIVASTAVA ET AL. 4575

and
t
1
E(t) = E(0) + (t − s)𝛽−1 [A(s) − (𝜇 + 𝜈)C(s) − 𝜀E(s)] ds (20)
Γ(𝛽) ∫0
(t > 0; 0 < 𝛽 ≦ 1).

Let us consider the kernels K1 (t, C) and K2 (t, E) given by

K1 (t, C) = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (21)

and
K2 (t, E) = A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t), (22)
respectively.

4.2 A set of theorems

Here, in this subsection, we state and prove Theorems 1 and 2 below.
Theorem 1. Each of the kernels K1 (t, C) and K2 (t, E) satisfies the Lipschitz condition.

Proof. First of all, we derive the Lipschitz condition for the kernel K1 (t, C) given by (21). Let C and C1 be two functions.
Then we have

‖K1 (t, C) − K1 (t, C1 )‖

= ‖[−(𝛿 + 𝜑)C(t) + 𝛿E(t)] − [−(𝛿 + 𝜑)C1 (t) + 𝛿E(t)]‖ , (23)

which, on using Cauchy's inequality, yields

‖K1 (t, C) − K1 (t, C1 )‖ ≦ |−(𝛿 + 𝜑)| · ‖C(t) − C1 (t)‖ , (24)

that is,
‖K1 (t, C) − K1 (t, C1 )‖ ≦ Λ ‖C(t) − C1 (t)‖ , (25)
where
|−(𝛿 + 𝜑)| ≦ Λ (Λ > 0). (26)

Next, in the case of the kernel K2 (t, E) given by (22), we similarly find for two functions E and E1 that

‖K2 (t, E) − K2 (t, E1 )‖ = ‖[A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t)] − [A(t) − (𝜇 + 𝜈)C(t) − 𝜀E1 (t)]‖ , (27)

which, by using Cauchy's inequality, yields

‖K2 (t, E) − K2 (t, E1 )‖ ≦ |𝜀| · ‖E(t) − E1 (t)‖ , (28)

that is,
‖K2 (t, E) − K2 (t, E1 )‖ ≦ Ω ‖E(t) − E1 (t)‖ , (29)
where

|𝜀| ≦ Ω (Ω > 0). (30)

We now consider the following recursive formula:

t
Cn (t) = K1 (t, Cn−1 ) + (t − s)𝛼−1 K1 (s, Cn−2 ) ds. (31)
∫0
4576 SRIVASTAVA ET AL.

The difference between the consecutive terms is given by

Un (t) = Cn (t) − Cn−1 (t)

= K1 (t, Cn−1 ) − K1 (t, Cn−2 )
t
+ (t − s)𝛼−1 [K1 (s, Cn−1 ) − K1 (s, Cn−2 )] ds, (32)
∫0

By taking the norm on both sides of (32), we find that

‖Un (t)‖ = ‖Cn (t) − Cn−1 (t)‖

= ‖K1 (t, Cn−1 ) − K1 (t, Cn−2 )
t ‖
‖
+ (t − s)𝛼−1 [K1 (s, Cn−1 ) − K1 (s, Cn−2 )] ds‖ , (33)
∫0 ‖
‖
which readily yields

‖Un (t)‖ = ‖Cn (t) − Cn−1 (t)‖

≦ ‖K1 (t, Cn−1 ) − K1 (t, Cn−2 )‖
t
‖ ‖
+ ‖(t − s)𝛼−1 [K1 (s, Cn−1 ) − K1 (s, Cn−2 )]‖ ds. (34)
∫0 ‖ ‖

In a similar manner, we can find that

‖Vn (t)‖ = ‖En (t) − En−1 (t)‖

≦ ‖K2 (t, En−1 ) − K2 (t, En−2 )‖
t
‖ ‖
+ ‖(t − s)𝛽−1 [K2 (s, En−1 ) − K2 (s, En−2 )]‖ ds. (35)
∫0 ‖ ‖

Theorem 2. The fractional-order mathematical model of diabetes and its complications has a coupled solution under
the following conditions with respect to t = t0 :
Λ + [Λ (t0 )]𝛼 < 1

and
Ω + [Ω (t0 )]𝛽 < 1.

Proof. We first observe that the differences Un (t) and Vn (t) in (34) and (35) are bounded. We have also proven that both
kernels K1 (t, C) and K2 (t, E) satisfy the Lipschitz condition. Consequently, from the results obtained in Equations 34
and 35, and by using the recursive technique, we get

[ ]n
‖Un (t)‖ ≦ ‖C(0)‖ · Λ + [Λ (t0 )]𝛼 (36)

and
[ ]n
‖Vn (t)‖ ≦ ‖E(0)‖ · Ω + [Ω (t0 )]𝛽 . (37)

Therefore, the above solutions exist and are continuous. Nonetheless, in order to show that each of the above
solutions is a solution of Equation 16, we get

C(t) − C(0) = Cn (t) − Pn (t), (38)

SRIVASTAVA ET AL. 4577

so that we have
‖
‖
‖Pn (t)‖ = ‖[K1 (t, C) − K1 (t, Cn−1 )]
‖
‖
t ‖
‖
+ (t − s)𝛼−1 (K1 (s, C) − K1 (s, Cn−1 )) ds‖
∫0 ‖
‖
≦ ‖K1 (t, C) − K1 (t, Cn−1 )‖
‖ t ‖
‖ ‖
+‖ (t − s)𝛼−1 [K1 (s, C) − K1 (s, Cn−1 )] ‖ ds
‖∫0 ‖
‖ ‖
≦ Λ ‖C − Cn−1 ‖ + Λt𝛼 ‖C − Cn−1 ‖ . (39)

By recursively solving this last Equation (39), we find that

‖Pn (t)‖ ≦ (Λ + Λt𝛼 )n+1 Θ (‖C − Cn−1 ‖ < Θ; Θ > 0) , (40)

which, at t = t0 , yields
‖Pn (t)‖ ≦ (Λ + [Λ (t0 )]𝛼 )n+1 Θ. (41)

Furthermore, by proceeding to the limit when n → ∞, we find from Equation 41 that

lim ‖Pn (t)‖ = 0.
n→∞

We have thus verified the existence of the solution. The other assertion of Theorem 2 can be proved similarly.

4.3 Uniqueness of the coupled solutions

In this subsection, we show that the coupled solutions presented in the above subsection are unique.
In order to prove the uniqueness, we assume that there exists another set of coupled solutions for the system given by
(16) and (17), say C(t) and E(t). Then we have
t
C(t) − C1 (t) = [K1 (t, C) − K1 (t, C1 )] + (s − t)𝛼−1 [K1 (s, C) − K1 (s, C1 )] ds, (42)
∫0
which, upon taking the norm on both sides, yields
t
‖C(t) − C1 (t)‖ ≦ ‖K1 (t, C) − K1 (t, C1 )‖ + ‖(s − t)𝛼−1 [K1 (t, C) − K1 (t, C1 )]‖ ds. (43)
∫0 ‖ ‖

Since the solution is bounded, by using the Lipschitz condition, having the fact in mind that the solution is bounded,
we get
C(t) = C1 (t). (44)

Similarly, we can show that

E(t) = E1 (t). (45)

This shows the uniqueness of the coupled solutions of the system given by (16) and (17).

4.4 Stability analysis

In this subsection, we describe the stability of the following fractional-order mathematical model of the diabetes and its
resulting complications:
D𝛼t [C(t)] = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (t > 0; 0 < 𝛼 ≦ 1) (46)
4578 SRIVASTAVA ET AL.

and
D𝛽t [E(t)] = A − (𝜇 + 𝜈)C(t) − 𝜀E(t) (t > 0; 0 < 𝛽 ≦ 1), (47)

where the parameters 𝛿, 𝜑, 𝜇, 𝜈, and 𝜀 are positive real numbers.

Just as we considered above, in the fractional-order model described by the nonlinear difference equations (46) and
(47), C(t) denotes the number of patients of diabetics with complications and E(t) denotes the population of those patients
who have diabetes of Type I and Type II at time t.
Now, as far as the solution of the above system of the nonlinear difference equations (46) and (47), we are interested
only in positive solutions. So, we consider the Jacobian matrix Δ of Equations 46 and 47 at any point, which is given by
[ ]
− (𝛿 + 𝜑) 𝛿
Δ= . (48)
− (𝜇 + 𝜈) −𝜀

For the trace of the matrix Δ, we have

Trace{Δ} = − (𝛿 + 𝜑 + 𝜀) . (49)

Thus, clearly, the system of the nonlinear difference equations (46) and (47) is stable if

Δ < 0, (50)
that is, if
𝛿 + 𝜑 + 𝜀 > 0. (51)

Since the parameters 𝛿, 𝜑 and 𝜀 are positive constants, the inequality in (51) always holds true. Consequently, the system
of the nonlinear difference equations (46) and (47) is always stable.

5 APPROX IMATE SOLUTIONS FO R THE FRACTIONA L- ORDER

MATHEMATICAL MODEL

Here, in this section, we consider the following system of fractional-order differential equations:

D𝛼t [C(t)] = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (t > 0; 0 < 𝛼 ≦ 1) (52)

and
D𝛽t [E(t)] = A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t) (t > 0; 0 < 𝛽 ≦ 1), (53)
together with
C(0) = C0 and E(0) = E0 .

By using the HDM steps, we arrive at the following equations:

∑∞
p t
pn Cn (t) = C(0) + (t − 𝜏)𝛼−1
n=0
Γ(𝛼) ∫0
( )
∑ ∞
∑
∞
· −(𝛿 + 𝜑) p Cn (𝜏) + 𝛿 p En (𝜏) d𝜏
n n
(54)
n=0 n=0

and
∑
∞
p t
pn En (t) = E(0) + (t − 𝜏)𝛽−1
n=0
Γ(𝛽) ∫0
( )
∑
∞
∑
∞
· A(𝜏) − (𝜇 + 𝜈) pn Cn (𝜏) − 𝜀 pn En (𝜏) d𝜏. (55)
n=0 n=0
SRIVASTAVA ET AL. 4579

Upon comparing the coefficients of the same powers of p, we get the following integral equations:

p0 ∶ C0 (t) = C(0) = C0 (56)

and
p0 ∶ E0 (t) = E(0) = E0 . (57)
t
1
p1 ∶ C1 (t) = (t − 𝜏)𝛼−1 [−(𝛿 + 𝜑)C0 + 𝛿E0 ] d𝜏 (58)
Γ(𝛼) ∫0
(C1 (0) = 0)
and
t
1
p1 ∶ E1 (t) = (t − 𝜏)𝛽−1 [A(𝜏) − (𝜇 + 𝜈)C0 − 𝜀E0 ] d𝜏 (59)
Γ(𝛽) ∫0
(E1 (0) = 0).
t
1
p2 ∶ C2 (t) = (t − 𝜏)𝛼−1 [−(𝛿 + 𝜑)C1 (𝜏) + 𝛿E1 (𝜏)] d𝜏 (60)
Γ(𝛼) ∫0
(C2 (0) = 0)
and
t
1
p2 ∶ E2 (t) = (t − 𝜏)𝛽−1 [−(𝜇 + 𝜈)C1 (𝜏) − 𝜀E1 (𝜏)] d𝜏 (61)
Γ(𝛽) ∫0
(E2 (0) = 0).

⋮
t
1
pn ∶ Cn (t) = (t − 𝜏)𝛼−1 [−(𝛿 + 𝜑)Cn−1 (𝜏) + 𝛿En−1 (𝜏)] d𝜏 (62)
Γ(𝛼) ∫0
(Cn (0) = 0)
and
t
1
pn ∶ En (t) = (t − 𝜏)𝛽−1 [−(𝜇 + 𝜈)Cn−1 (𝜏) − 𝜀En−1 (𝜏)] d𝜏 (63)
Γ(𝛽) ∫0
(En (0) = 0).

The components of the series solution are obtained directly as follows:

C(0) = C0 and E(0) = E0 ; (64)

𝛼 ′ t𝛼 𝛽 ′ t𝛽
C1 (t) = and E1 (t) = ; (65)
Γ(𝛼 + 1) Γ(𝛽 + 1)
𝛼 ′ t2𝛼 𝛽 ′ t𝛼+𝛽
C2 (t) = −𝜒 +𝛿 (66)
Γ(2𝛼 + 1) Γ(𝛼 + 𝛽 + 1)
and
𝛼 ′ t𝛼+𝛽 𝛽 ′ t2𝛽
E2 (t) = −𝜉 −𝜀 . (67)
Γ(𝛼 + 𝛽 + 1) Γ(2𝛽 + 1)
⋮
The notations used above are given by

𝛿 + 𝜑 = 𝜒, 𝜇 + 𝜈 = 𝜉, 𝛼 ′ = −𝜒C0 + 𝛿E0
and 𝛽 ′ = 1 − 𝜉C0 − 𝜀E0 . (68)

The remaining components can also be obtained in this manner.

Here, we consider only a few terms of the series solutions which are given as follows:
C(t) = C0 + C1 (t) + C2 (t) + C3 (t) + … (69)
and
E(t) = E0 + E1 (t) + E2 (t) + E3 (t) + … . (70)
4580 SRIVASTAVA ET AL.

For the first-order ODE model considered by Boutayeb et al,17 if we set 𝛼 = 1 and 𝛽 = 1 in Equations 16 and 17, we
arrive at the following equations:
d
[C(t)] = −(𝛿 + 𝜑)C(t) + 𝛿E(t) (t > 0) (71)
dt
and
d
[E(t)] = A(t) − (𝜇 + 𝜈)C(t) − 𝜀E(t) (t > 0), (72)
dt
together with
C(0) = C0 and E(0) = E0 . (73)

Thus, upon solving by applying the HDM steps as above, we deduce the following approximate solutions to the problem:
1
C(t) = C0 + 𝛼 ′ t + (𝛿𝛽 ′ − 𝜒𝛼 ′ )t2 + … (74)
2
and
1
E(t) = E0 + 𝛽 ′ t − (𝜉𝛼 ′ + 𝜀𝛽 ′ )t2 + … . (75)
2
Only the second-order term of the series derived by the above-detailed methods was used in evaluating the approx-
imate solutions. The accuracy of our approximate solutions can be improved by computing higher order terms of the
approximate solutions.

6 NUMERICAL AND COMPUTATIONAL RESULTS

In this section, we have graphically analyzed the numerical and computational results, which we have derived in this
paper. The obtained results show the importance of the usage of the fractional calculus in mathematical modeling.
We have considered the solution of the problem described by Equations 16 and 17 for different values of the parameters
involved therein. With the help of the time variable t, we choose to show the importance of fractional-order model. In order
to discuss the solution of Equations 16 and 17, we have considered the following data: (a) the number of incidences of the
diabetes mellitus given by A = 6 × 106 ; (b) the initial size of the population of the diabetes mellitus given by E0 = 6.11 × 107 ;
and (c) the initial size of the population of diabetics with complications given by C0 = 4.7 × 107 . We have also considered
the following specific values of the parameters:
𝜆 = 0.08, 𝜇 = 0.05, 𝜈 = 0.05, and 𝜀 = 0.02.
Figure 2 exhibits the behavior of the number C(t) of sufferers of diabetes with complications for different values of 𝛼 and
𝛽 with respect to time t. Moreover, it is clear from Figure 2 that the graph in the case when 𝛼 = 𝛽 = 1, in comparison with
the graph for fractional values of 𝛼 and 𝛽, decreases rapidly. However, as in the current scenario, it is not to be seen that

FIGURE 2 The graph of C(t) [Colour figure can be viewed at wileyonlinelibrary.com]

SRIVASTAVA ET AL. 4581

FIGURE 3 The graph of E(t) [Colour figure can be viewed at wileyonlinelibrary.com]

the number of persons having diabetes with complications decreases rapidly. Hence, from the results presented in this
paper, we can say the above-defined fractional-order model is better in comparison to the model considered by Boutayeb
et al.17 Also, in Figure 3, the graph of the population E(t) of those patients who have diabetes at time t for 𝛼 = 𝛽 = 1 shows
that the number of diabetic patients falls speedily in comparison with the graph in the case for fractional values of 𝛼 and
𝛽. Nevertheless, it is clear from researches and the WHO report that, in the present scenario, it is not the case. Hence,
from the results derived in this paper, it is clearly seen that the fractional-order model is better to describe the problem of
diabetes and its resulting complications.

7 CO NCLUDING REMARKS AND O BSERVATIONS

In this paper, we have introduced and investigated a certain fractional-order mathematical model of diabetes and its
resulting complications. We have considered the possibility of using the fractional calculus in our investigation. By using
some fixed point theorems, we have proved the existence and uniqueness of the solution of a system of fractional-order
differential equations which mathematically model diabetes and its resulting complications. We have also discussed the
stability of the solution presented in this paper. In our investigation, we have used a relatively new analytical technique,
the HDM for solving the system of fractional-order differential equations of diabetes and its resulting complications. The
HDM is seen to be a powerful and efficient tool for solving a system of fractional-order partial differential equations
(FPDEs). Moreover, the calculations involved in HDM are not complicated. Just as in some of the other works, our
investigation has some limitations and can be improved further.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

ORCID

Hari M. Srivastava https://orcid.org/0000-0002-9277-8092

Ravi Shanker Dubey https://orcid.org/0000-0002-3019-4438
Monika Jain https://orcid.org/0000-0002-3432-5132

REFERENCES
1. WHO Report Commission. Definition, diagnosis and classification of diabetes mellitus and its complications: report of a WHO
consultation. Part 1: diagnosis and classification of diabetes mellitus. WHO/NCD/NCS/99.2, Geneva, World Health Organization; 1999.
http://www.who.int/iris/handle/10665/66040
2. WHO Report Commission. Global Report on Diabetes. Geneva: World Health Organization; 2016.
4582 SRIVASTAVA ET AL.

3. Whiting D., Unwin N., Roglic G. Diabetes: equity and social determinants. In: Blas E, Kuru A. S., eds. Equity, Social Determinants and
Public Health Programmes. Geneva: World Health Organization; 2010:77-94.
4. NCD Risk Factor Collaboration (NCD-RisC). Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies
with 4.4 million participants. Lancet. 2016;387:1513-1530. https://doi.org/10.1016/S0140-6736(16)00618-8
5. DIAMOND Project Group. Incidence and trends of childhood type 1 diabetes worldwide 1990–1999. Dia Med. 2006;23(8):857-866.
6. WHO Guideline. Sugars Intake in Adults and Children. Geneva: World Health Organization; 2015.
7. Morenga LT, Mallard S., Mann J. Dietary sugars and body weight: systematic review and meta analyses of randomized controlled trials
and cohort studies. British Med J. 2013;2013:1-25, Article ID 346. https://doi.org/10.1136/bmj.e7492
8. WHO Report. Global Status Report on Noncommunicable Diseases. Geneva: World Health Organization; 2014;2014.
9. WHO Report. Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013–2020. Geneva: World Health
Organization; 2013.
10. WHO/FAO Expert Consultation. Diet, Nutrition and the Prevention of Chronic Diseases, Report of a Joint WHO/FAO Expert Consultation.
Geneva: World Health Organization; 2003.
11. Johnsson IW, Haglund B, Ahlsson F, Gustafsson J. A high birth weight is associated with increased risk of type 2 diabetes and obesity.
Pediatr Obes. 2015;10(2):77-83.
12. Whincup PH, Kaye SJ, Owen CG, et al. Birth weight and risk of type 2 diabetes: a systematic review. J Am Med Assoc.
2008;300(24):2886-2897.
13. Harder T, Rodekamp E, Schellong K, Dudenhausen JW, Plagemann A. Birth weight and subsequent risk of type 2 diabetes: a meta-analysis.
Amer J Epidemiol. 2007;165:849-857.
14. Willi C, Bodenmann P, Ghali WA, Faris PD, Cornuz J. Active smoking and the risk of type 2 diabetes: a systematic review and meta-analysis.
J Amer Med Assoc. 2007;298(22):2654-2664.
15. Luo J, Rossouw J, Tong E, et al. Smoking and diabetes: does the increased risk ever go away? Amer J Epidemiol. 2013;178(6):937-945.
16. WHO Report. Global Recommendations on Physical Activity for Health. Geneva: World Health Organization; 2010.
17. Boutayeb A, Twizell EH, Achouayb K, Chetouani A. A mathematical model for the burden of diabetes and its complications. BioMedical
Engineering Online. 2004;3(20):Article ID PMC449725, 1-8.
18. Podlubny I. Fractional Differential Equations: An Introduction to Fractional Derivatives, Fractional Differential Equations, to Methods of
Their Solution and Some of Their Applications, Mathematics in Science and Engineering, vol. 198. New York, London, Sydney, Tokyo and
Toronto: Academic Press; 1999.
19. Caputo M. Linear models of dissipation whose Q is almost frequency independent—II. Geophys J Internat. 1967;13:529-539.
20. Chaurasia VBL, Dubey RS, Belgacem FBM. Fractional radial diffusion equation analytical solution via Hankel and Sumudu transforms.
Mathematics in Engineering Sci Aeros. 2012;3:1-10.
21. Miller KS, Ross B. An Introduction to Fractional Calculus and Fractional Differential Equations. New York, Chichester, Brisbane, Toronto
and Singapore: A Wiley-Interscience Publication, John Wiley and Sons; 1993.
22. Chaurasia VBL, Dubey RS. Analytical solution for the differential equation containing generalized fractional derivative operators and
Mittag-Leffler-type function. Internat Sch Res Not ISRN Appl Math. 2011;2011:1-9, Article ID 682381.
23. Kilbas AA, Srivastava HM, Trujillo JJ. Theory and Applications of Fractional Differential Equations, North-Holland Mathematical Studies,
vol. 204. Amsterdam, London and New York: Elsevier (North-Holland) Science Publishers; 2006.
24. Dubey RS, Goswami P, Belgacem FBM. Generalized time-fractional telegraph equation analytical solution by Sumudu and Fourier
transforms. J Fract Calc Appl. 2014;5:52-58.
25. Baleanu D, Diethelm K, Scalas E, Trujillo JJ. Fractional Calculus: Models and Numerical Methods, Series on Complexity Nonlinearity and
Chaos, vol. 3. Singapore, New Jersey, London and Hong Kong: World Scientific Publishing Company; 2012.
26. Srivastava HM, El-Sayed AMA, Gaafar FM. A class of nonlinear boundary value problems for an arbitrary fractional-order differential
equation with the Riemann-Stieltjes functional integral and infinite-point boundary conditions. Symmetry. 2018;10:Article ID 508, 1-13.
27. Atangana A, Alkahtani BST. Analysis of non-homogenous heat model with new trend of derivative with fractional order. Chaos Solitons
Fractals. 2016;89:566-571.
28. Ali MR, Hadhoud AR, Srivastava HM. Solution of fractional Volterra-Fredholm integro-differential equations under mixed boundary
conditions by using the HOBW method. Adv Difference Equ. 2019;2019:1-14, Article ID 115.
29. Yang X-J, Machado JAT, Cattani C, Gao F. On a fractal LC-electric circuit modeled by local fractional calculus. Commun Nonlinear Sci
Numer Simul. 2017;47:200-206.
30. Chaurasia VBL, Dubey RS. Analytical solution for the generalized time fractional telegraph equation. Fract Differ Calculus. 2013;3:21-29.
31. Debbouche A, Nieto JJ. Approximate controllability of fractional delay dynamic inclusions with nonlocal control conditions. Appl Math
Comput. 2014;243:161-175.
32. Srivastava HM, Günerhan H. Analytical and approximate solutions of fractional-order susceptible-infected-recovered epidemic model of
childhood disease. Math Methods Appl Sci. 2019;42:935-941.
33. Alkahtani BST, Alkahtani JO, Dubey RS, Goswami P. Solution of fractional oxygen diffusion problem having without singular kernel.
J Nonlin Sci Appl. 2016;11:1-9.
34. Alkahtani BST, Alkahtani JO, Dubey RS, Goswami P. The solution of modified fractional Bergman's minimal blood glucose-insulin model.
Entropy. 2017;19(5):1-11, Article ID 114.
SRIVASTAVA ET AL. 4583

35. Yang A-M, Zhang Y-Z, Cattani C, et al. Application of local fractional series expansion method to solve Klein-Gordon equations on Cantor
sets. Abstr Appl Anal. 2014;2014:1-6, Article ID 372741.
36. Srivastava HM, Shah FA, Abass R. An application of the Gegenbauer wavelet method for the numerical solution of the fractional
Bagley-Torvik equation. Russian J Math Phys. 2019;26:77-93.
37. Dubey RS, Alkahtani BST, Atangana A. Analytical solution of space-time fractional Fokker-Plank equation by homotopy perturbation
Sumudu transform method. Mathematical Problems in Engineering. 2015;2015:1-7, Article ID 780929.
38. Yang X-J, Srivastava HM, Cattani C. Local fractional homotopy perturbation method for solving fractal partial differential equations arising
in mathematical physics. Romanian Rep Phys. 2015;67:752-761.
39. Srivastava HM, Saad KM. Some new models of the time-fractional gas dynamics equation. Adv Math Models Appl. 2018;3(1):5-17.
40. Atangana A, Alabaraoye E. Solving system of fractional partial differential equations arising in the model of HIV infection of CD4+ cells
and attractor one-dimensional Keller-Segel equations. Adv Difference Equ. 2013;2013:1-14, Article ID 94.
41. Baleanu D, Shiri B, Srivastava HM, Al Qurashi M. A Chebyshev spectral method based on operational matrix for fractional differential
equations involving non-singular Mittag-Leffler kernel. Adv Difference Equ. 2018;2018:1-23, Article ID 353.
42. Atangana A. Analytical solutions of boundary values problem of 2D and 3D poisson and biharmonic equations by homotopy decomposi-
tion method. Abstr Appl Anal. 2013;2013:1-9, Article ID 380484.
43. Singh H, Pandey RK, Srivastava HM. Solving non-linear fractional variational problems using Jacobi polynomials. Mathematics.
2019;7:1-24, Article ID 224.
44. Odibat Z, Momani S. Modified homotopy perturbation method. Application to quadratic Riccati differential equation of fractional order.
Chaos, Solitons Fractals. 2008;36(1):167-174.
45. Dubey RS, Belgacem FBM, Goswami P. Homotopy perturbation approximate solutions for Bergman's minimal blood glucose-insulin
model. Fractal Geom Nonlin Anal Med Biol. 2016;2(3):1-6.

How to cite this article: Srivastava HM, Shanker Dubey R, Jain M. A study of the fractional-order
mathematical model of diabetes and its resulting complications. Math Meth Appl Sci. 2019;42:4570–4583.
https://doi.org/10.1002/mma.5681