Rosina Surovi Khan et al.

/ (IJCSE) International Journal on Computer Science and Engineering

Vol. 02, No. 08, 2010, 2616-2621

Design of a Hospital-Based Database System

(A Case Study of BIRDEM)
Rosina Surovi Khan
Department of Computer Science and Engineering Mohammad Saber
Ahsanullah University of Science and Technology (AUST) Department of Computer Science and Engineering
Dhaka, Bangladesh Ahsanullah University of Science and Technology (AUST)
Email: [email protected] Dhaka, Bangladesh
Email: [email protected]
Abstract— As technology advances, information in different
organizations of Bangladesh can no more be maintained I. EXTRACTION OF INFORMATION FROM HOSPITAL
manually. There is a growing need for the information to become MANAGEMENT (BIRDEM)
computerized so that it can be suitably stored. This is where BIRDEM was visited and information was gathered;
databases come into the picture. Databases are convenient demos were seen and its websites visited from the internet.
storage systems which can store large amounts of data and Now Ibrahim Medical College is founded in BIRDEM for the
together with application programs such as interfaces they can MBBS degree. A lot of students pass MBBS from this medical
aid in faster retrieval of data. An initiative was taken to design a college. [3,4]
complete database system for a hospital management such as
Bangladesh Institute of Research and rehabilitation in Diabetes, A database system was designed based on a case study of
Endocrine and Metabolic disorders (BIRDEM) in Dhaka so that BIRDEM Hospital via Entity Relationship Diagram (ERD),
its information can be stored, maintained, updated and retrieved Relational Model , Normalization of tables and
conveniently and efficiently. The existing information in Implementation in SQL server. [1,2,5] The ERD is outlined
BIRDEM is partly computerized via databases only in patients’ below.
admissions, doctors’ appointments and medical tests and reports
sections. A partly slow and tedious manual system still exists in II. DATABASE DESIGN
BIRDEM for example, in record of ambulances in service,
assigning ward boys and nurses to rooms, the billing process and A. Entity Relationship Model (ERM)
record of doctors’ prescriptions etc. However, this paper outlines 1) ER Diagram
one complete database design for the entire BIRDEM hospital in In the ER diagram, we can view the entities- Patient,
which data maintenance and retrieval are in perfect harmony Doctor, Receptionist, Department, Medicine, Test, OT
and speedy. Sample SQL-based queries executed on the designed
(Operation Theater), Room, Nurse, Ward_boy, Driver,
system are also demonstrated.
Ambulance, Carriers, Accountant and Bill. Among these
entities, relationships exist which connect all the entities in the
Keywords- Database system, entity relationship diagram,
diagram. For example, Patient, Doctor and Receptionist are
relational model, normalization, SQL-based queries. connected via the relationship Appointment. In other words, a
receptionist will set up a doctor’s appointment for a patient.
Introduction Similarly, Doctor, Patient and Medicine are connected via the
relationship Prescription. Here, a doctor may prescribe one or
Bangladesh Institute of Research and rehabilitation in more medicine to a patient. In a similar way, other entities are
Diabetes, Endocrine and Metabolic disorders, abbreviated as connected via relationships in a meaningful way. Cardinality
BIRDEM, is a Research Center for Diabetes and also a ratios [1] for the entities connected to a relationship are
hospital. There are doctors, patients, and employees like explained in the next section.
nurses, ward boys, ambulance carriers which are considered as
entities in the designed system. It becomes a tough, tedious 2) Cardinalities
and comparatively slow process to store the information partly Binary Relationship
manually and partly computerized. But having all the
We can see a binary relationship in the ER diagram.
necessary information stored in one database, it not only helps
in orderly maintenance but very speedy retrieval of data.

The work in this paper has been supervised by Rosina Surovi

Khan for her student Syed Mahboob Nur and his group as a thesis
work in Year 4, Semester 1, CSE, AUST in Fall 2009.

ISSN : 0975-3397 1
 Figure 1. ER diagram of BIRDEM

As can be seen in Figure 2, 1 Doctor can be from 1 or

many Departments . 1 Department may have 1 or many Ternary Relationship
Doctors . So it is a many to many relationship named Doctor
from Department (in the diamond). As seen in Figure 3, 1 Receptionist can admit 1 Patient in 1
Room in a certain date and time. 1 Receptionist can admit in 1
Room 1 Patient in a certain date and time. In 1 Room, 1
Patient is admitted by 1 Receptionist in a certain date and
time. So the relationship is a ternary relationship named
Admission (in the diamond) with the cardinality ratio from
Patient to Receptionist to Room being 1 to 1 to 1.
Similar logic applies to other ternary relationships in the
ER diagram of Figure1.

Figure 2. Example of Binary Relationship

 vii) Bill(Bill_id, Bill_purpose, Bill_total)
Here Bill_purpose refers to the cause e.g blood test for
which the bill is paid.
viii) Accountant(Acct_id, Acct_name, Age, Address,
MOB, Working_time, Acct_salary)
ix)Payment Table (Pay_id, Bill_id, Pat_id, Acct_id, Pay_type,
(fk) (fk)
Pay_date)
This is a junction table between Patient, Bill & Accountant
tables. Primary key of the Patient Table goes to Payment table
as foreign key. Primary key of the Accountant Table goes to
Payment table as foreign key. Primary Key of the Bill Table
becomes a part of the primary key in Payment table. (Pay_id,
Bill_id) is the primary key in the Payment table. Since the
Figure 3. Example of ternary relationship cardinality ratio from Patient to Accountant to Bill is 1 to 1 to
C (which stands for many), Pat_Id from Patient table and
Acct_id from Accountant table become foreign Keys in the
Payment table while Bill_id from Bill table becomes a part of
B. Relational Model the primary key in the Payment table. In this table, pay_type
The entire ER diagram in sec A.1 can be converted to a refers to the type of payment such as cash, pay order, check or
Relational Model (relational tables) [1,2,5] as shown below. credit card.
The attribute(s) of a relation which serve as a primary key of
the table are underlined. Those attribute(s) which represent vii) Medicine(Mdcn_id, Mdcn_name, company, m_date,
foreign keys are indicated as fk underneath . e_date, price)

i)Patient(Pat_id, Pat_name, age, sex, Address, DOB, MOB) m_date refers to manufacture date. e_date refers to expiry
date.
Here DOB refers to patient’s date of birth and MOB refers
to mobile number. viii) Prescription( Prs_id, Doc_id, Mdcn_id, Pat_id, date, fee)
(fk) (fk)
ii) Room(Room_id, Room_No, Room_type, Room_cost) This is a junction table between Patient, Doctor & Medicine
iii)Admission(admsn_id, Pat_id, Room_id, Rcp_id, date, tables.

time) ix) Test(Test_id, Test_name, date, rep_date, fee)

(fk) (fk) (fk) Here rep-date refers to the date the report of the test will be
This is a junction table among Patient, Room and given.
Receptionist tables. Primary key of the Patient table goes to
Admission table as foreign key. Primary key of the Room x) Assist(Srl_no, Pat_id, Doc_id, Test_id, time,
table goes to Admission table as foreign key. Primary key of date) (fk) (fk)
the Receptionist table goes to Admission table as foreign key. This is a junction table between Patient, Doctor & Test
Admsn_id is a primary key in the Admission table. Since the tables. Srl_no corresponds to serial number of conducted test.
cardinality ratio from Patient to Receptionist to Room is 1 to 1 xi) OT (Ot_id, Ot-room_no)
to 1, Pat_Id from Patient table, Room_Id from Room table and
Rcp_Id from Receptionist table become foreign keys in the Here OT refers to Operation Theater.
Admission junction table. xii) Operation(Op_id, Doc_id, Pat_id, Ot_id, Op_date,
iv)Receptionist(rcp_id, rcp_name, Age, Address, MOB, (fk) (fk)
Shifting) Op_time)
This is a junction table between Patient, Doctor & OT
Here Shifting refers to morning, afternoon or night shifts. tables.
v)Doctor Table(Doc_id, Doc_name, Age, Address, Salary, xiii) Department(Dept_id, Dept_name, treatment)
MOB, Designation, Passed_from)
xiv) Doctor_from_Department(Dfd_id, Doc_id, Dept_id)
Here Passed_from indicates passed from which institution.
This is a junction table between Doctor & Department
vi)Appointment(Ap_id, Pat_id, Doc_id, Rcp_id, Ap_date, tables.
(fk) (fk) (fk)
xv) Nurse(Nrs_id, Nrs_name, Age, Address, MoB,
Ap_time) Nrs_wo_shift, experience, salary)
This is a junction table between Patient, Receptionist &
Doctor tables.
Here Nrs_wo_shift refers to the working shift of the nurse such as morning, afternoon or night shifts.
xvi) Nursing_Service(Ns_id, Pat_id_, Nrs_id, Room_id) The relation is in 1NF because the attributes of the relation
(fk) do not have sub attributes.
This is a junction table between Patient, Room & Nurse
tables. The relation is in 2NF because non-primary keys are fully
functionally dependent on primary key.
xvii) Ward_Boy(Wb_id, Wb_name, MoB, w_shift,
The relation is in 3NF because no transitive dependency
Salary) Here w_shift refers to working shift of a exists from non-primary key to primary key.
ward boy. The relation is in BCNF because there is no part of primary
xviii) Cleaning _Service(Cls_id, Pat_id, Wb_id, Room_id) key that is fully functionally dependent on non primary key.
(fk) Similarly, all other relations of the system follow Codd’s
This is a junction table between Patient , Room & rules of normalization except the Payment table which is
Ward_Boy tables. explained below.
xix) Driver(Dr_id, Dr_name, MoB, Address, Shift,
Salary) xxiii)Ambulance(Amb_id, Amb_num, Capacity)
2) Violation of Normal forms
xxiv)Ambulance_Service(As_id, Pat_id, Dr_id, Amb_id)
(fk) (fk) (fk)
Payment Table :-
This is a junction table between Patient, Driver &
Ambulance tables. This is a junction table between Patient, Bill & Accountant
tables.
xxv) Carriers(Cr_id, Cr_name, MOB, Address, Salary)
Payment(Pay_id, Pat_id, Bill_id, Acct_id, Pay_type,Pay_date)
Carriers are those persons who carry patients from the
ambulance to the hospital premises. For the above relation, the following functional
xxvi) Carrying_Service Table(CS_id, Cr_id, Amb_id, Pat_id) dependencies exist:
(fk) (fk) {Pay_id}=> {Pay_type, Pay_date, Pat_id}
This is a junction table between Patient, Ambulance &
Carriers tables. Two different patient ids, payment dates and payment
types cannot correspond to the same payment id. So Pay_type,
Pay- date and Pat_id are fully functionally dependent on
Pay_id.
C. Functional Dependencies and Normalization
1) Fulfillment of all normal forms {Bill_id}=> {Acct_id , Pat_id}
Similarly two different accountant ids and patient ids
Example: cannot correspond to the same bill id. So Acct_Id and Pat_id
are fully functionally dependent on Bill_id.
Room Table:-
Based on the above functional
Room(Room_id, Room_No, Room_type, Room_cost) dependencies: The relation is in 1NF.
The relation is not in 2NF because all non-primary keys are
{Room_id} => {Room_No}
not fully functionally dependent on the primary key (Pay_id,
Here Room-id corresponds to serial numbers like 1, 2, 3, 4 Bill_id). So we split the relation to make it 2NF.
etc. Room_No corresponds to designated numbers of rooms
Payment1(Pay_id, Pay_Type, Pay_date, Pat_id)
like 206, 307, 508 etc. A Functional Dependency exists from
Room_No to Room_id because two different Room_Nos Payment2 (BP_id, Bill_id, Acct_id, Pat_id) {BP_id stands
cannot correspond to the same Room_id. for Bill Pay id and corresponds to serial numbers of all bills
such as 1, 2, 3 etc in ascending order. Bill_id on the other hand
Similarly, the following functional dependencies exist: corresponds to the bill code numbers}.
{Room_id} => {Room_type}
The relations are now in
{Room_id} => {Room_cost} 2NF. The relations are in
Considering the above full functional dependencies, we 3NF.
now verify whether the relation fulfills Codd’s stringent rules
of normalization. [1,2,5] The relations are in BCNF.

D. Implementation in SQL Server

SQL server is a modern software where we can store huge
amount of information via a database. In fact we have
 Figure 4. SQL Server Implementation Diagram
Query 1:
Select Pat_name, Doc_name, Test_name from tbl_Patient,
implemented our database system in SQL server. In this server tbl_Doctor, tbl_Test, tbl_Assist where Doc_name = 'Selima'
we can store data easily, retrieve data speedily and execute the and tbl_Doctor.Doc_id = tbl_Assist.Doc_id and
queries conveniently by SQL query language. We can create tbl_Patient.Pat_id = tbl_Assist.Pat_id and tbl_Test.Test_id =
relations easily by code and also manually (drag and drop tbl_Assist.Test_id
method like MS.Access). Output 1:

III. USAGE OF THE SYSTEM

We can easily retrieve various data based on our demands
using SQL Queries. Sample data has been entered in the
relational tables. Some useful queries that can be imposed on
our designed system but cannot be retrieved that speedily from
the existing partly manual and partly computerized
information system of BIRDEM are: Question 2
a) Ambulance Id 100 carried which patients to BIRDEM Which doctors prescribed which medicine to patient
whom receptionist Sadia gave doctors’ appointment for heart Mamun ?
ailments?
Query 2:
b) Which patients were billed for above Tk5000 having
Select Pat_name, Doc_name, Mdcn_name from
gone through kidney tests suggested by Doctor Selima?
tbl_Patient, tbl_Doctor, tbl_Medicine, tbl_Prescription where
c) Doctors from which departments prescribed mainly Pat_name = 'Mamun' and tbl_Patient.Pat_id =
antibiotics to patients? tbl_Prescription.Pat_id and tbl_Doctor.Doc_id =
tbl_Prescription.Doc_id and tbl_Medicine.Mdcn_id =
d) How many patients with irregular ECG reports went tbl_Prescription .Mdcn_id
though a heart surgery in Cardiology Department?
Output 2:
Sample queries directed on our system together with SQL
expressions and results follow next.
Question 1
Which tests are suggested by doctor Selima to which
patients ?

Question 3 :

ISSN : 0975-3397 2620

 Which doctors are from which departments and they Applied Sciences, Germany. She also completed PhD
passed from which college and got salaries below Tk 20000 ? coursework in Computer Science from the University of Texas
at Arlington, USA.
Query 3
The second author of this paper, Mohammad Saber is also
Select Doc_name, Passed_from, Dept_name from currently working as an Assistant Professor in the Department
tbl_Doctor, tbl_Department, tbl_DFD where Salary <20000 of Computer Science and Engineering, Ahsanullah University
and tbl_Doctor.Doc_id = tbl_DFD.Doc_id and of Science and Technology, Dhaka, Bangladesh. He obtained
tbl_Department.Dept_id = tbl_DFD.Dept_id his MSc in Computer Science and Engineering from
(Table tbl_DFD in SQL Server implementation diagram Bangladesh University of Engineering and Technology
corresponds to the relationship Doctor From Department in (BUET).
ER diagram).
Output 3:

IV. CONCLUSION
Our database contains all the information needed to be
maintained in a BIRDEM hospital. As we have computerized
the entire system via a database, the maintenance is very
convenient and efficient and also retrieval of data according to
demand is speedy. The existing system of BIRDEM is partly
manual and partly computerized and it becomes a tedious
process to keep track of all the information partly in paper files
and partly on computers. Therefore, our designed system is a
good and useful implementation. We can further improvise it
by enhancing its security. An initiative has also been taken to
use Microsoft Visual Studio 2008 and the programming
language C# for developing user friendly interfaces to the
current database system. That way a software has been
developed which is used to interface with the SQL Server and
hence data accessed, retrieved and searched for far better in a
more efficient and convenient form.
REFERENCES
[1] A. Silbershatz, H. Korth., S. Sudarshan, Database System Concepts,
McGraw Hill, 5th Edition, 2005
[2] Database System Concepts Slides: http://www.cse.iitb.ac.in/sudarsha/db-
book/slide-dir/ch1.pdf, Accessed 2010
[3] Diabetic Association of Bangladesh: http://www.dab-bd.org/, Accessed
2010
[4] BIRDEM hospital: http://www.dab-bd.org/project/birdem/birdem.php,
Accessed 2010
[5] R. Elmasri., S. B. Navathe, Fundamentals of Database Systems,
Addison-Wesley, 3rd Edition, 2000.

AUTHORS’ PROFILES

The first author of the paper, Rosina Surovi Khan is

currently working as an Assistant Professor in the Department
of Computer Science and Engineering, Ahsanullah University
of Science and Technology, Dhaka, Bangladesh. Her medium
of instruction has always been in English. She obtained her
MSc in Software Technology from Stuttgart University of

ISSN : 0975-3397 2621