0% found this document useful (0 votes)
17 views9 pages

1530-Article Text-6070-6388-10-20211125

This research article applies the Gk/Gd/1 queuing model to analyze patient flow in a hospital in Raipur, India, addressing the imbalance between patient demand and service supply. The study derives appointment probabilities, expected queue lengths, and waiting times, demonstrating that using multiple servers can significantly reduce patient waiting times. The findings aim to enhance healthcare service efficiency and are relevant for academic literature and healthcare practitioners.

Uploaded by

Daniela Moura
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
17 views9 pages

1530-Article Text-6070-6388-10-20211125

This research article applies the Gk/Gd/1 queuing model to analyze patient flow in a hospital in Raipur, India, addressing the imbalance between patient demand and service supply. The study derives appointment probabilities, expected queue lengths, and waiting times, demonstrating that using multiple servers can significantly reduce patient waiting times. The findings aim to enhance healthcare service efficiency and are relevant for academic literature and healthcare practitioners.

Uploaded by

Daniela Moura
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Communications in Mathematics and Applications

Vol. 12, No. 3, pp. 645–653, 2021


ISSN 0975-8607 (online); 0976-5905 (print)
Published by RGN Publications http://www.rgnpublications.com

DOI: 10.26713/cma.v12i3.1530

Research Article

Application of G k /G d /1 Queuing Model to Patient


Flow at Hospital
Manish Kumar Pandey * , D. K. Gangeshwer and Thaneshwar Lal Verma
Department of Applied Mathematics, Chhattisgarh Swami Vivekanand Technical University, Bhilai, India

Received: February 27, 2021 Accepted: May 29, 2021

Abstract. The health systems should have an ability to deliver efficient and smooth and safe services
to the patients. Now-a-day, in hospitals, to get timely appointments to doctors, is a very difficult task,
for most of patients long wait for appointments, that means demand and supply are imbalanced in a
queue. Queuing theory is the branch of operations research in applied mathematics and deals with
the phenomenon of waiting lines. Therefore, the present paper deals with the application of G k /G d /1
queuing model to patient flow at hospital namely Raipur, India. The arrival process is measured
by exponential distribution and the service process is measured by Poisson distribution. Finally,
appointment probabilities of waiting time of patients have been derived, and also expected queue
length, waiting time for the patients in the model have been shown. It has also been observed that
waiting time for patients can be reduced by using multiple servers instead of a single server queued
model. Lastly, a numerical illustration of the model has been provided. The proposed result would be
useful for academic literature, queuing scientists, and practitioners.
Keywords. Queuing theory; G k /G D /1 model; Appointment probability
Mathematics Subject Classification (2020). 68M20; 90B22

Copyright © 2021 Manish Kumar Pandey, D. K. Gangeshwer and Thaneshwar Lal Verma. This is an open access
article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution,
and reproduction in any medium, provided the original work is properly cited.

1. Introduction
Saima and Nisha [13] remarked that in 1930’s operations research is taken as a discipline
of science and branch of applied mathematics. The past of queuing theory was nearly 100
years. A Danish telephonic engineer A. K. Erlang developed queuing theory in early 1920’s.
During study in applications of automatic telephone switching, Erlang was concerned with

* Corresponding author: [email protected]


646 Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al.

the capacities and utilization of the equipment and lines. Queuing theory was continued in
applications to furnish for a large number of situations, at the end of World War-II. Queuing
theory is actually a study of waiting lines. The theory allows the calculation and derivation of a
number of representative measures which includes the estimated number of receiving service.
The probability of encounters the system in certain cases, e.g. having to wait for certain time to
be served or an available server, empty or full and average waiting time in the queue.
As a result of the rapid growth of the population, the people of Indian have started to
demand more efficient health care at resonable cost. Hospitals are one of the most important
place in the healthcare service. Quality and their services directly affect human life. Simulation
is one of the oldest analysis tools; simulation is the process of making a model of a system and
conducting experiments with model for the purpose of knowing the behavior of the operation of
the system. Simulation is a problem solving technique that helps to study problems that cannot
be analyzed using direct and formal analytical methods.
Use of simulation in hospitals there are many examples as cited in the several reviews of the
academic literature, such as those by Vassilacopoulos [16] determined the bed complement in
hospital inpatient departments to meet a pre determined demand for service and considered the
models: (i) emergency patients should be admitted without delay, (ii) occupancy should not fall
below a pre specified level, and (iii) the waiting list length should not exceed a predetermined
number. Lane et al. [9] described the collaborative process of building a simulation model in
order to understand patient waiting times in an accident and emergency department and they
focused about on some general themes that can be discovered running through the process.
Brailsford et al. [3] analyzed the relative frequency to use of a range in operational research
with modeling approaches in health care along the specific domains of application. The level
of implementation and provides new insights into the level of activity across many areas of
application, highlighting important relationships and pointing to key areas of omission and
neglect in the literature. Gunal and Pidd [5] reviewed the literature about discrete event
simulation for performance modeling in health care.
Shi et al. [15] analyzed a representative telephone response system of Veterans Affairs (VA)
hospitals, address the existing inefficiency issues such as long call waiting time, and improve
system resilience to changes use the methodology of discrete event simulation is adopted to
model the current system and the resource sharing schemes and find out the resource sharing
schemes dramatically improve system performance reflected by the decrease of call waiting
time and queue, as well as the extreme high utilization of agents in a key unit. Pinto et al. [11]
addressed the question of generic simulation models and their role in improving emergency care
around the world and report the construction of a reusable model for ambulance systems and
also described about the associated parameters, data sources, performance measures and report
on the collection of information, as well as the use of optimization to configure the service to
best effect. Jiang et al. [7] used the statistical method for statistical analysis of MRI inspection
time and calculated the waiting queue length, average queue length, and waiting time then

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al. 647

finally conclude the problem of health allocations of resources could be solved by establishing
regional MRI examination center radiation regional around. Subsequently, queuing theory
for the simulation in health care have been studied by several authors, Bhattacharjee and
Ray [2] told about appointment systems for scheduling patients to a hospital facility play an
important role in controlling and synchronizing the arrival of patients with resource availability
thereby reducing the waiting time of patients and increasing the utilization of resources. Kuo
et al. [8] presented a case study which uses simulation to analyze patient flows in a hospital
emergency department in Hong Kong and analyzed the impact of the enhancements made
to the system after the relocation of the emergency department after that, they developed a
simulation model to capture all the key relevant processes of the department. Ben-Tovim et
al. [1] outlined the design, development and application of a hospital patient flow management
support tool – Hospital Event Simulation Model: Arrivals to Discharge (HEAD). Rodrigues et
al. [12] developed a discrete event simulation model that estimates Level 2 bed needs for a large
university hospital and they innovates by simulating the entirety of the hospital’s inpatient
flow and most importantly, the ICU’s daily stochastic flows based on a nursing workload scoring
metrics. Hu et al. [6] examined the contributions of queuing theory in modeling EDs and assess
the strengths and limitations of this application. Cocchi et al. [4] developed a methodology able
to improve non-clinical front office operation for the patients keeping the costs under control
and find out that DES is a valuable tool that can be used to save money and improve clinical
processes. Recently, Pandey and Gangeshwer [10] studied a queuing model with heterogeneous
servers for specific service in health on health sector.
This paper solves the imbalance problem of demand and supply for queue in the field of
hospitals which are managed by central government or state govt. In Section 2, we analyze
about materials and methods in which three main sector of hospital Raipur is focused which
are patient registration department in which we seen how to patient will register and take the
service then next service for patients is out patients department i.e. OPD where some special
area for patients treatment, after treatment of patient analysis of the pharmacy department. In
Section 3, we analyze our model G k /G d /1 about probability and departure. Finally, in Section 4
result and discussion about appointment system and new appointment probabilities of waiting
time of patients have been derived; also found the expected queue length and waiting time for
the patients in the model. This model provides safe, efficient and smooth services to the patients.

2. Materials and Methods


The material and methods of the proposed study are as below:
(i) Data are collected from a district hospital of Raipur India, for two days of a week. Three
main sector of district hospital Raipur is focused which are (a) Patient’s registration,
(b) Out patients department (OPD) and (c) Pharmacy Section. ‘Questionnaire’, ‘direct
observation’, and ‘interviews’ methods are used for data collection. Some assumptions are
used for data of queuing models.

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
648 Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al.

(ii) Suppose that the patient’s obey a Poisson probability distribution.


(iii) Patients are independent and exponentially distributed in the time of inter-arrival.
(iv) Service time is also exponentially distributed.
(v) Suppose that patients are following the rule first-come first-served (FIFO) basis and also
without getting service no patient will leave the queue.
(vi) The queue is endless for OPD, doctors were only servers.
(vii) Rate of serving was not dependent on the queue length.

3. Analysis of G k /G d /1 Queuing Model


In this model, single doctor is available but patients are N as single server queuing system
and patients required for service is a Poisson process with rate is λ N . This is free from bulk
patient in service and appointment backlog. Let finite queue length is K with finite waiting
room, therefore excessive patients will not make an appointment and seek treatment. Here,
service rule is FCFS order and doctor service times are with length T , because service times of
doctor are fixed.

Figure 1

The results of this study stand up on the three characteristics of the rate of invalidations:
(a) There exists a no availability rate even for same day appointments.
(b) The rate of no-availability monotonically increase with the excessive patients until it
reaches a maximum.
(c) The rate of no-availability fixed when it reaches the high value.
Now, no-availability function is as Sharma [14]
γ( k) = ymax − ( ymax − γ0 ) e−k/c
µ ¶2 µ ¶3
k 1 k 1 k 1
½ µ ¶ ¾
= ymax − ( ymax − γ0 ) 1 − + − +...
c 1! c 2! c 3!
µ ¶2 µ ¶3
k 1 k 1 k 1
µ ¶
= γ0 + (γmax − γ0 ) − (γmax − γ0 ) + (γmax − γ0 )
c 1! c 2! c 3!
µ ¶3 ½µ ¶2 µ ¶4
k 1 k 1 k 1 k 1
½µ ¶ ¾ ¾
= γ0 + (γmax − γ0 ) + + . . . − (γmax − γ0 ) + +...
c 1! c 3! c 2! c 4!
k k
µ ¶ ½ µ ¶ ¾
= γ0 + (γmax − γ0 ) sinh − (γmax − γ0 ) cosh −1
c c

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al. 649

k k k k
µ ¶ µ ¶ µ ¶ µ ¶
= γ0 + γmax sinh − γ0 sinh − γmax cosh + γmax + γ0 cosh − γ0
c c c c
k k
µ ¶ µ ¶
= γmax + (γmax − γ0 ) sinh − (γmax − γ0 ) cosh , (3.1)
c c
where γ0 ≥ 0 indicate the minimum execute no-availability rate and γ0 < γmax = 1 maximum
execute no-availability rate; c is a no-availability parameter; k is the value of patient’s
appointment at the time when a patient joins the queue.

3.1 Probability of a Patient Finishes His/her Service


If a probability of a patient finishes his/her service in the time interval between t and t + ∆ t also
drop behind k patients for new appointment, for this we use D ( k, t, t + ∆ t), where 0 = k = K − 1.
Similarly, D ( t, t + ∆ t) = D (0, t, t + ∆ t) + D (1, t, t + ∆ t) + D (2, t, t + ∆ t) . . . D (K − 1, t, t + ∆ t).

3.2 Departure Rate


Now, the corresponding departure rate for patients be
d ( k, t) = lim D ( k, t, t + ∆ t), where 0 = k = K − 1 ,
∆ t→0
d ( t) = lim D ( t, t + ∆ t) .
∆ t→0
Let P ( k, t) be the probability of new appointment with includes k patients at the time t . Now,
assume the time intervals ϕn = { t : ( n − 1)T ≤ t ≤ nT }, n ∈ N . Numerically, if time t = 0 then
appointment system is empty. If time t − 1 then the time period ϕ1 = { t : 0 ≤ t < T } that means
no patients departures for the completeness we report here some result.

Theorem 3.1. Let p( k, t), k = 0, 1, 2, 3, . . . , K be the probability that the appointment backlog
include k patient at time t with time interval.

Proof. p(0, 0) = 1 that means no patients appointment when t = 0,


p( k, 0) = 0, k = 1, 2, 3, . . . , K when t = 0,
and d ( k, 0) = 0, k = 01, 2, 3, . . . , K − 1 when t = 0.
So probability distribution p( k, t) obey the following distribution:
d p(0, t)
= −λ N p(0, t).
dt
Thus, we get
d
dt p(0, t)
= −λ N .
p(0, t)
After integrating with respect to t , we get
log p(0, t) = −λ N t + C 1 ,
where C 1 is constant and put t = 0 in above we get C 1 = 0, so we get
log p(0, t) = −λ N t .

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
650 Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al.

Finally, we get
p(0, t) = e−λ N t . (3.2)
and
d
p( k, t) = −λ N p( k, t) + λ N p(K − 1, t)
dt
where k = 1, 2, 3, . . . , K − 1
d
p( k, t) + λ N p( k, t) = λ N p(K − 1, t).
dt
Integrating with respect to t , we get
Z Z
p( k, t) + λ N p( k, t) dt = λ N p(K − 1, t) dt ,
·Z Z ¸
p( k, t) = λ N p(K − 1, t) dt − p( k, t) dt (3.3)

and for k = K
d
p( k, t) = −λ N p(K − 1, t). (3.4)
dt

Now, the following result explain the evolution of the appointment system.

3.3 New Appointment Probabilities


The following result explains the evolution of the appointment system. Suppose ρ = λ NT , where
λ N is doctor’s service rate and T is service time.
ρk
α( k) = e−ρ
k!
k ¶ k
ρ ρ2 ρ3 kρ ρ
µ
= 1− + − + · · · + (−1) +...
1! 2! 3! k! k!
K +1 K +2 K +3 K K
ρ ρ ρ ρ kρ ρ
µ ¶
= − + − + · · · + (−1) +...
k! 1! k! 2! k! 3! k! k! k!
∞ ρ k+ m
(−1)m
X
= k=0
m=0 r ! k!
∞ (λ N )k+m
(−1)m
X
= k=0
m=0 r ! k!
then for any time interval
" #
(λ N )k+m
∞ ∞ (λ N )k−1+m
d ( k, t) = e−λ N(t−T) λ N {1 − r γ( k)} (−1)m (−1)m
X X
+ r γ( k − 1)
m=0 m! k ! m=0 m!( k − 1)!
∞ (λ N ) m
(−1)m
X
+ {1 − r γ( k)} d ( k + 1, t − T )
m=0 m!
" #
k k+1− i + m k− i + m
∞ (λ N ) ∞ (λ N )
(−1)m (−1)m
X X X
+ {1 − r γ( k)} + r γ( k − 1) d ( i, t − T ).
i =1 m=0 m !( k + 1 − i )! m=0 m!( k − i )!
The above equation is new appointment probabilities.

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al. 651

4. Result and Numerical Illustration of the Model


If n = total number of patient in the our model, then,
(i) Expected queue length for waiting patient in a queue is measured by
µ ¶s
1 λ µλ
· ¸
Lq = p0
( s − 1)! µ (µ s − λ)2
 
1 λs+1 µλ
= ´2  p 0
 
( s − 1)! µs−1
³
µ s 1 − µλs

1 λs+1 1
· ¸
= p0
( s − 1)! µs s(1 − λ)2
1 λs+1 1
· ¸
= p0
s( s − 1)! µs (1 − λ)2
1 λs+1 1
· ¸
= p0 . (4.1)
s! µs (1 − λ)2
(ii) Expected waiting time for the patient in the queue
Lq
Wq = . (4.2)
λ
(iii) Expected queue length of waiting patient in the model
λ
Lm = Lq + . (4.3)
µ
(iv) Expected waiting time of the patient in the model
Lm
Wm = . (4.4)
λ
The performance measures using queuing analysis and queuing simulation of single server and
multiple server queuing model at a government hospital using arrival rate λ, service rate µ and
number of server.

Measure PRD OPD Pharmacy


(for two (only single
window) window)
(i) Arrival rate (λ) of patient 70/hour 22/hour 42
(ii) Service rate (µ) of patient 38/hour 15/hour 45
(iii) Model Utilization (ρ ) 92.10% 73.34% 93.33%
(iv) Probability (P0 ) when model is idle 4.46% 15.55% 9.56%
(v) mean (L m ) 8.18 0.517 7.609
(vi) mean (L q ) 8.16 0.516 7.60
(vii) mean time of patient spend in queue (Wm -hour) 0.108 2.35 0.181
(viii) mean time of patient spend in queue (Wq -hour) 0.109 2.28 0.180

The queued features in a district hospital of Raipur C.G. are analyzed using the queue
analysis and linear simulation in three departments. Single servers and multiple server queue

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
652 Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al.

models have been used for these analyzes. The present paper compared to other departments,
waiting time for patients in the queue in both days is more in the pharmacy department. It
has also been observed that waiting time for patients can be reduced by using multiple servers
instead of single server queued model. Lastly, numerical Illustration of the model has been
provided. The proposed result would be useful for academic literature, queuing scientists and
practitioners.

Notations
If n = total number of patient in the our model;
S = Number of doctor’s;
λ = Arrival rate /hour;
µ = Serving rate/ hour;
sµ = Service rate when s > 1;
Arrival rate/hour λ
ρ = model utilization, i.e., ρ = or ρ = sµ .
Service rate

Acknowledgement
The authors are very thankful to the unknown referees whose suggestions have helped in
improving the paper.

Competing Interests
The authors declare that they have no competing interests.

Authors’ Contributions
All the authors contributed significantly in writing this article. The authors read and approved
the final manuscript.

References
[1] D. Ben-Tovim, J. Filar, P. Hakendorf, S. Qin, C. Thompson and D. Ward, Hospital event simulation
model: Arrivals to discharge–design, development and application, Simulation Modeling Practice
and Theory 68 (2016), 80 – 94, DOI: 10.1016/j.simpat.2016.07.004.
[2] P. Bhattacharjee and P. K. Ray, Simulation modeling and analysis of appointment system
performance for multiple classes of patients in a hospital: a case study, Operations Research
for Health Care 8 (2016), 71 – 84, DOI: 10.1016/j.orhc.2015.07.005.
[3] S. C. Brailsford, P. R. Harper, B. Patel and M. Pitt, An analysis of the academic literature
on simulation and modeling in health care, Journal of Simulation 3(3) (2009), 130 – 140,
DOI: 10.1057/jos.2009.10.
[4] D. Cocchi, F. Frosini, E. Ciagli, P. Tortoli, C. Carpini, D. Cirone and A. Belardinelli, Discrete event
simulation model for the analysis of centralized front office service in a regional hub hospital, in
World Congress on Medical Physics and Biomedical Engineering (2018), pp. 559 – 562, Springer,
Singapore, DOI: 10.1007/978-981-10-9035-6_103.

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021
Application of G k /G d /1 Queuing Model to Patient Flow at Hospital: M. K. Pandey et al. 653

[5] M. M. Günal and M. Pidd, Discrete event simulation for performance modeling in health care: a
review of the literature, Journal of Simulation 4(1) (2010), 42 – 51, DOI: 10.1057/jos.2009.25.
[6] X. Hu, S. Barnes and B. Golden, Applying queueing theory to the study of emergency department
operations: a survey and a discussion of comparable simulation studies, International Transactions
in Operational Research 25(1) (2018), 7 – 49, DOI: 10.1111/itor.12400.
[7] H. Jiang, M. Wang and Y. Mao, MRI allocation analysis of regional hospitals based on queuing
simulation model, China Medical Equipment 13(6) (2016), 101 – 103.
[8] Y. H. Kuo, O. Rado, B. Lupia, J. M. Leung and C. A. Graham, Improving the efficiency of a hospital
emergency department: a simulation study with indirectly imputed service-time distributions,
Flexible Services and Manufacturing Journal 28(1-2) (2016), 120 – 147, DOI: 10.1007/s10696-014-
9198-7.
[9] D. C. Lane, C. Monefeldt and E. Husemann, Client involvement in simulation model building: hints
and insights from a case study in a London hospital, Health Care Management Science 6(2) (2003),
105 – 116, DOI: 10.1023/A:1023385019514.
[10] M .K. Pandey and D. K. Gangeshwer, A queuing model with heterogeneous servers for specific
service on health sector, International Journal of Applied Mathematics & Statistics 58(1) (2019),
28 – 35, URL: http://www.ceser.in/ceserp/index.php/ijamas/article/view/5991.
[11] L. R. Pinto, P. M. S. Silva and T. P. Young, A generic method to develop simulation models
for ambulance systems, Simulation Modeling Practice and Theory 51 (2015), 170 – 183,
DOI: 10.1016/j.simpat.2014.12.001.
[12] F. Rodrigues, G. S. Zaric and D. A. Stanford, Discrete event simulation model for planning Level
2 “step-down” bed needs using NEMS, Operations Research for Health Care 17 (2018), 42 – 54,
DOI: 10.1016/j.orhc.2017.10.001.
[13] M. Saima and S. Nisha, A comparison of single server and multiple server queuing models
in different departments of hospitals, Punjab University Journal of Mathematics 47(1) (2015),
http://journals.pu.edu.pk/journals/index.php/pujm/article/viewArticle/3598.
[14] N. Sharma, Analysis of Simulation for Queuing Models, Ph.D Thesis, Department of Statistics,
Maharshi Dayanand University Rohtak, Haryana, India (2011), URL: http://hdl.handle.net/
10603/7835.
[15] J. Shi, E. Erdem, Y. Peng, P. Woodbridge and C. Masek, Performance analysis and improvement of a
typical telephone response system of VA hospitals: a discrete event simulation study, International
Journal of Operations & Production Management 35(8) (2015), 1098 – 1124, DOI: 10.1108/IJOPM-
01-2014-0016.
[16] G. Vassilacopoulos, A simulation model for bed allocation to hospital inpatient departments,
Simulation 45(5) (1985), 233 – 241, DOI: 10.1177%2F003754978504500502.

Communications in Mathematics and Applications, Vol. 12, No. 3, pp. 645–653, 2021

You might also like