-
HEALTHCARE BUSINESS INTELLIGENCE: THE CASE OF UNIVERSITYS HEALTH
CENTER
Ishola Dada Murainaa*, Azizah Ahmadb ab
ITU-UUM ASP CoE,
Universiti Utara Malaysia, 06010 Sintok Kedah Malaysia
*[email protected]
[email protected]
ABSTRACT
Organizations, private or public, feel increasing pressures,
forcing them to respond quickly to changing conditions and be
innovative in the way they operate. Such activities require
organizations to be agile and make frequent and strategic,
tactical, and operational decisions. Making such decision may
require considerable amounts of timely and relevant data,
information, and knowledge. Every semester university admits new
students; they do subject them to medical screening which sometimes
includes the staffs and returning students. However, the results of
the medical test from the laboratory technologists and the doctors,
such as patient diagnosis, treatment and medical prescription are
currently kept in the health center data repository for record
purposes without being further explored for their managerial
activities. Therefore, this paper applies Business Intelligence
(BI) method for exploring the university health center database
repository. The data warehouse was built for the activities in
university health center and a prototype was developed at the end,
while the system is evaluated by the prospective users of the
system. The result of this research helps the university health
center management by simplifying the technique needed for
managerial decision making and forecasting future activities that
would help the center. Also, the health care BI is also useful to
know the medical statistics of the patients in university community
and the drugs that need to be frequently ordered for.
Keywords: Business Intelligence (BI), University Utara Malaysia
(UUM), University Health Center Business Intelligence (PKUBI), Star
Schema
-
1. Introduction
The business environment in which organizations operate today is
becoming more complex and ever-changing. Organizations, private or
public, feel increasing pressures, forcing them to respond quickly
to changing conditions and be innovative in the way they operate.
Such activities require organizations to be agile and make frequent
and strategic, tactical, and operational decisions. Making such
decision may require considerable amounts of timely and relevant
data, information, and knowledge. Processing these in the framework
of the required decisions needs quick, frequent and some
computerized support which is traced to business intelligence (BI)
(Efraim, et al. 2008). Healthcare organizations are swimming in an
ever-deeper pool of data. But without a program in place to target,
gather, deliver and analyze the most relevant data, these
organizations will continue to be data rich but information poor
(Eckerson, 2003). Forward-thinking healthcare organizations realize
that data and, thus, BI is at the center of informed and precise
decision-making will improve patient and service outcomes in
ensuring their organizations future (Hyperion Solution Corporation,
2004). To achieve the full benefits of BI in healthcare
organizations, there must be a strategic approach to tactical
projects and realize that the greatest efficiencies come from
integrating data historically in operation and clinical systems
(Microsoft, 2009).
University Utara Malaysia (UUM) at Northern part of Malaysia is
a public university which is located in a small town called Sintok,
Kedah State. The university was officially established on February
16 th, 1984 and its mission is to play leadership role in
developing the country by providing high quality management
education in the country. On top of that, the university has
offered excellent education areas which are represented as College
of Business (COB), College of Arts and Sciences (CAS), and College
of Law, Government and International Studies (COLGIS).
In addition, UUM exempts itself in such a way that its students
and staffs health are guaranteed by conducting medical-up for them
at the University Utara Malaysia Health Center (PKU) during the new
intake admission and in the time of need to see the medical
officers. This medical related information is kept for record
purposes in the database repository of UUM. The data in the
database can be manipulated using BI techniques and tools in order
to provide PKU with faster and more accurate reporting, improved
decision making and better customer service, and eventually
increased revenue.
Furthermore, BI is operationally described as a collection of
data warehousing, data mining, analytics, reporting and
visualization technologies, tools, and practices to collect,
integrate, cleanse, and mine enterprise information for decision
making (Inmon, et al. 2000). Todays BI architecture is usually
designed for strategic decision making, where a small number of
expert users analyze historical data to prepare reports or build
models, and decision making cycles for previous activities
(Umeshwar, et al. 2009).
The major objective of BI is to enable interactive access to
data, to enable manipulation of data, and to give business managers
and analysts the ability to conduct appropriate analysis on the
historical and current data, which will reveal the situations and
performance for making better decision (Zaman, 2005). Data from
distributed sources such as online transaction processing (OLTP)
systems is periodically extracted, cleansed,
-
integrated, transformed, and loaded into a data warehouse (DW),
which in turn is queried by analytic applications (Chaudihuri, et
al. 2001).
Every semester UUM enrolls the new intakes with the target of
having 40% of graduate students and 60% of undergraduate students
as a research university, the PKU conducts medical test for the new
students before they are allowed to continue with their academic
activities. Both the staffs and returning students are subjected to
medical screening whenever they need medical attention from the
physicians throughout the semesters. The results of the medical
test from the laboratory technologists and the doctors, such as
patient diagnosis, treatment and medical prescription are currently
kept in the data repository for record purposes without being
further explored for their managerial activities. However, there
have been a fewer research that explores BI for PKU in managerial
decision making which should help in speeding-up the processes in
dealing with their patients and day-to-day activities of the PKU.
Therefore, this research intends to use the medical record of PKU
at UUM by performing business analytics which boosts their
services.
Finally, this paper produces answers to the following questions
like
i. What are the requirements needed for business intelligence in
UUM Health Center? ii. What type of model necessary for BI in
making decision at UUM Health Center?
iii. Which prototype needed for making decision in UUM Health
Center? iv. Does the UUM Health Centers BI system easy to use?
In addition, the application of BI in UUM health center is an
important measure which is used as a computerized support for
managerial decision making. This is achieved by viewing the
performance through the help of visualization of important data,
while the services receive by the students and staffs of UUM is
also increased. This research is also contributed to the body of
knowledge in the areas of healthcare and BI domains (helping the
PKU know the common diseases among the patients and medication that
need to be supplied frequently).
2. Business Intelligence (BI) Raisinghani (2004) described BI as
an umbrella term that includes architectures, tools, databases,
application, and methodologies. This means that BI is a
content-free expression that reflects different things to different
people. He added that BI enables interactive access to data,
enables manipulation of data and provides business managers and
analysts the ability to conduct appropriate analysis. Zaman (2005)
stressed that analysis in BI is based on the historical and current
data, situations, and performances which make decision makers to
get valuable insights upon which they can base more informed and
better decision. Therefore, the process of BI is based on the
transformation of data to information, to decisions and finally to
actions.
In industry, BI is finally achieving an increased prominence
record and existed in many forms for over three decades, initially
being called Decision Support Systems (DSS), while the umbrella
term is still most widely used in academia. The term Business
Intelligence has existed even longer, but in its present form has
been attributed to historical data (Watson & Wixom, 2007). In
addition, BI has evolved into a managerial philosophy and a
business tool, which can be referred to as an organized and
systematic process by which organizations acquire, analyze, and
disseminate information from both internal and external information
sources significant for their business activities and for decision
making (Lonnqvist & Pirttimaki, 2006).
-
3. EVOLUTION OF BUSINESS INTELLIGENCE The BI was brought newly
by the Gartner group in the mid 1990s from the area of management
information system (MIS) reporting systems of the 1970s. During the
introduction of BI, the reporting systems were static, having
two-dimensional features and had no analytical capabilities. In
early 1980s, the concept of executive information systems (EIS) was
emerged by expanding computerized support to top-level managers and
executives. The EIS concept consists of dynamic multidimensional
reporting systems, forecasting and predicting, trend analysis,
drilling down to details, status access and critical success
factors, while all of these features leaved to the mid 1990s. The
BI evolved with the existing capabilities, but built of the EIS
with few features and believes that all the information needed by
the executives can be in a BI-based enterprise information system.
In the year 2005, BI system started to include artificial
intelligence and powerful analytical capabilities (Efraim, et al.
2008). Figure 1 shows the interconnectivities of features that lead
to business intelligence (BI).
Fig. 1: Evolution of Business Intelligence (Source: Efraim, et
al. 2008)
4. CHARACTERISTICS OF BUSINESS INTELLIGENCE Inmon (2005)
submitted that set or tools and methodologies of BI have the
following characteristics:
i. Accessibility to Information: The business intelligence is
known to be flexible and allows end users to gain access to data
regardless of the source of data.
ii. Support in Decision Making: Business intelligence presents
the information and gives access to analysis tools that will allow
the users to select and manipulate data that are important to
them.
iii. Strategic Advantage: The business intelligence creates
fewer barriers to entry for new competitors to enter and possess
globalization features for readily available supply chain and
e-commerce.
-
5. BENEFITS OF BUSINESS INTELLIGENCE Eckerson (2003) reported
that many executives do not insist on a rigorous cost-justification
for business intelligence projects due to its numerous intangible
benefits, while Thompson (2004) noticed that the most common
application areas of BI are general reporting, sales and marketing
analysis, planning and forecasting, financial consolidation,
statutory reporting, budgeting and profitability analysis.
Meanwhile, Eckerson (2003) highlighted the benefits of BI as saves
time, improves strategies and plans, and improves tactical
decisions, more efficient in processes and cost saving.
In another contribution by Thompson (2004) shows that the major
benefits of BI are; faster, more accurate reporting, improves
decision making, improves customer service and increases
revenue.
6. TYPES OF BUSINESS INTELLIGENCE DATA SOURCES Adelman &
Larissa (2000) emphasized that one of the challenges in building a
BI decision-support environment is to merge data from different
types of data sources. The authors stated that there are three
major types of data sources: operational, private and external. 6.1
Operational Data Source
An operational data source is a database repository which can be
used to store and retain runtime data. The operational data source
application can read and write data to and from this data source
throughout the life of the application unlike the metadata data
source. The metadata source is primarily written during the
implementation stage of development and read mostly during startup
of the server. Also, because data stored in the operational data
source is of a transient nature, users do not need to back-up,
restore, or transfer the underlying database during upgrade. Both
online transaction processing (OLTP) and batch systems provide
internal operational data about subject areas, such as financial,
logistics, sales, order entry, personnel, billing, research and
engineering.
6.2 Private Data Source
Adelman & Larissa (2000) referred to internal departmental
data as the data that usually comes from the desktops and
workstations of business analysts, knowledge workers,
statisticians, and managers which include the Product Analysis
Spreadsheets, Regional Product Usage Spreadsheets and Prospective
Customer Databases.
6.3 External Data Source
Demarco (2001) submitted that organizations are often purchase
external data from vendors that specialize in collecting specific
industrial information that are available in the public domain like
Healthcare Statistics, Customer Profile Information, Customer
Catalog-Ordering Habits and Customer Credit Report.
Therefore, Kimball & Richard (2000) added that external data
is usually clustered around the following categories:
i. Sales and Marketing Data: this is the list of prospective
customers. ii. Credit Data: individual credit ratings, business
viability assessments.
iii. Competitive Data: products, services, prices, sales
promotions, mergers, takeovers. iv. Industry Data: technology
trends, marketing trends, management science and trade
information.
-
v. Economic Data: currency fluctuations, political indicators,
interest rate movements, stock and bond prices.
vi. Economic Data: income groups and customer behavior. vii.
Demographic Data: age profiles and population density.
viii. Community Data: raw material prices. ix. Psychometric
Data: consumer profiling. x. Meteorological Data: weather
conditions, rainfall, temperature (for agriculture and
travelling industry).
7. BUSINESS INTELLIGENCE (BI) ARCHITECTURE According to Eckerson
(2003), BI consists of four major components that are merged
together to form BI architecture; a data warehouse (DW), business
analytics (BA), business performance management (BPM) and a user
interface. The BI architecture with its components is shown in
Figure 2 below.
Fig. 2: A high level architecture of BI (Source: Eckerson, W.
2003)
7.1 Data Warehouse
Data warehouse and its variants is the cornerstone of any
medium-to-large BI system. Initially, data warehouse included only
the historical data that are organized and summarized, and allows
end users to easily view or manipulate data and information. While
data warehouse in the new versions is known to include current data
in order to provide real-time decision support.
-
7.2 Business Analytics
Efraim, et al. (2008) added that business analysis as one of the
components of BI architecture is a collection of tools for
manipulating and analyzing the data in the data warehouse without
segregating the data mining. The author said business analytic
makes it possible for the end users to work with data and
information in a data ware house by using a variety of tools and
techniques that are classified into three categories below:
i. Reporting and Queries: This includes both static and dynamic
reporting, all types of queries, and discovery of information,
multidimensional view and drill-down to details.
ii. Advanced Analytics: Advance analytics includes statistical,
financial, and mathematical and models used in analyzing data and
information.
iii. Data, Text and Web Mining: This is a process of searching
for unknown relationships or information in large database or data
warehouse, using intelligent tools like neural computing and
predictive analytics techniques.
7.3 Business Performance Management (BPM) Business performance
management (BMP) is always referred to as corporate performance
management (CPM) which is an immerging portfolio of applications
and methodology that contains BI architecture and tools. BPM does
measure, monitor, comparing of sales, profit, cost and other
performance indicators by introducing the concepts of management
and feedback. It also performs functions like planning and
forecasting as the core belief of a business strategy. Unlike DSS,
EIS and BI which support the bottom-top extraction of information
from data, BPM provides a top-down enforcement of corporate-wide
strategy.
7.4 The User Interface (Dashboards and Other Information
Broadcasting Tools) Dashboards provide a comprehensive visual view
of corporate performance measures, such as (Key Performance
Indicators) trends and exceptions from multiple business areas. The
graphs do show actual performance versus desired metrics and
provide at-a-glance view of the health of the organization.
John (2007) stated that dashboards provide an at a glance view
of business performance for many individuals in an organization.
They give companies a factual and timely window into performance
and help to identify anomalies that could turn into significant
business issues, and therefore provide an entry point for digging
deeper into root causes. Meanwhile, data consistency is crucial to
the success of any dashboard solution which means that no matter
how spectacular the interface is, it has to be fed with trusted
data from an enterprise-class platform. Without reliable and
consistent data, the value of predictive analysis is limited at
best. Moreover, John (2007) added that the following as the
characteristics of BI dashboards:
i. BI dashboards deliver a high degree of visualization with
graphs, gauges and charts.
ii. BI dashboards offer personalized views of trusted key
information. iii. BI dashboards can easily be delivered in multiple
formats to suit specific needs of
business users. iv. They are easy to manage from an IT
perspective.
-
8. HEALTHCARE ORGANIZATIONS
Maria & Abdel-Badeeh (2010) stressed that Healthcare
organizations (HCOs) are information-intensive enterprises, while
Healthcare personnel requires sufficient data and information
management tools to make appropriate decisions. Clinicians assess
patients status, plan patients care, administer appropriate
treatments, and educate patients and families regarding clinical
management of various conditions. Primary-care physicians and care
managers assess the health status of new members of a health plan.
Medical directors evaluate the clinical outcomes, quality, and cost
of health services provided. Administrators determine appropriate
staffing levels, manage inventories of drugs and supplies, and
negotiate payment contracts for services. Governing boards make
decisions about investing in new business lines, partnering with
other organizations, and eliminating underutilized services.
Collectively, healthcare professionals comprise a heterogeneous
group with diverse objectives and information requirements.
In addition, the authors submitted that the main objective of
HCO in a highly competitive environment is to reduce operating
costs while maintaining a consistently acceptable level of patient
treatment. Reduce operating costs at all levels, such as:
i. Cost of healthcare Professionals. ii. Cost of lab equipment
& consumables.
iii. Cost of pharmaceuticals / medical material. iv. Cost of a
treatment per Diagnosis related grouping (DRG). v. Cost per type of
medical intervention (specific medical operation).
Meanwhile, an acceptable level of patient treatment involves: i.
Evidence based medicine, accurate diagnosis and efficient
treatment.
ii. On time admittance in the Hospital and healthcare treatment.
iii. Treatment with respect for the Patient- analysis of options.
iv. Reduction of risks during treatment. v. Capture of medical
history of the patient in to support evidence based medicine.
9. HEALTHCARE ACTIVITIES, SERVICES AND PROCESSES IN THE CONTEXT
OF BUSINESS INTELLIGENCE
Health care organizations typically prescribe how their
processes have to be performed; especially those processes that
represent complex routine work, that involve many persons and
organizational units and that are in general frequently performed
(Yorozu et al., 1987). In the context of BI, medical processes are
those activities and work practices within a HCO and focused on the
health services delivery (nursing and medical treatment). Business
processes comprise activities that are needed to effectively run
the health care organization. Support processes are used from both
kinds of processes but only have an indirect impact on medical and
business activities (supply of materials) as shown in Figure 3.
-
Fig. 3: Healthcare Process in the Context of BI
10. TECHNOLOGICAL COMPONENTS OF BI IN HEALTHCARE
Maria & Abdel-Badeeh (2010) submitted that intelligent
technologies can be seen as enabler for managing, storing,
analyzing, visualizing, and giving access to a great amount of data
in the context of BI. For this purpose, a wide range of intelligent
technologies such as; expert systems, online analytical processing,
data mining and knowledge discovery, grid computing, cloud
computing are used in developing BI system in healthcare sector.
Technology is required to provide an integrated view of both,
internal and external data (data warehouse) which is regarded as
the base for BI. 10.1 Types of Operational Databases Maria &
Abdel-Badeeh (2010) highlighted the essential parts of database
technologies and intelligent technologies from BI perspective, and
concluded that three types of operational databases should be
created in healthcare organization (HCO) as shown in Figure 4.
i. Clinical Operational Databases (CODB): these include all kind
of medical data which is needed for health care service delivery to
the patients; such as electronic patient records, and laboratory
results.
ii. Administrative Operational Databases (AODB): these contain
the entire business data which is required for running the health
care organization; like personnel data, and financial data.
iii. External Operational Databases (EODB): these can either be
clinical or business data from an external provider (medical
reports, insurance forms, and statistical data).
-
Fig. 4: Technology of Business Intelligence in Healthcare
11. BENEFITS OF BI IN HEALTHCARE INDUSTRY
The benefits from applying BI in the healthcare environment can
be tremendous. BI serves an increasingly wide variety of
departments in the provider market with an assortment of unique
reporting and analysis applications (Kornack & Rakic, 2001).
Thus, the authors added that a robust BI environment offers
healthcare organizations a host of business benefits including,
which include the following:
i. The ability to optimize resources (including physical space,
equipment and devices, staff and supplies) in individual
departments such as Surgical Services.
ii. The ability to develop and monitor key performance
indicators and clinical indicators to improve performance and
quality.
iii. The ability to conduct planning, budgeting, and forecasting
more efficiently and accurately across large organizations.
iv. The ability to effectively understand and manage the supply
chain and logistics to contain costs and ensure consistent
supply.
v. The ability to better ensure patient safety through efficient
diagnostics and the identification and enforcement of best practice
treatment protocols.
-
vi. The ability to contain costs and improve performance and
quality through human resources management and physician
profiling.
12. RELATED WORK ON BI
Diana et al. (2010) submitted that a disorder characterized by
an excessive sweating was treated by endoscopic thoracic
sympathectomy which improved the patient overall quality of life.
Therefore, the patients daily activities are not affected, or are
less affected, by this disorder, and their emotional state verifies
a significant improvement, from a situation of shame and
self-punishing to what we could say a normal life. The authors
presented the analysis of the quality of life of 227 patients that
were treated by an endoscopic thoracic sympathectomy, using
business intelligence technologies which allowed the storage, the
analysis and the reporting of all the relevant findings. Meanwhile,
the authors illustrated that database and data analysis
developments were needed in a specific healthcare application
domain. Such as, a data mart (data storage) which was designed to
address the relevant attributes. Also, On-line analytical
processing and data mining (data analysis) technologies were used
to show the evolution of the patients health conditions.
Furthermore, Maria & Abdel-Badeeh (2010) stressed that
Business intelligence is a new methodology to maximize the benefits
for healthcare organization. Business intelligence also provides an
integrated view of data that can be used to monitor, key
performance indicators, identify hidden patterns in diagnosis and
identify variations in cost factors. Therefore, intelligent
techniques provide an effective computational methods and robust
environment for business intelligence in the healthcare domain.
Christos et al. (2008) stressed that On-Line Analytical
Processing (OLAP) tools use multidimensional views to provide quick
access to information. Therefore, these have become the existing
standard in the business world for analytical databases. In health
care, care givers and managers could benefit from being able to
perform interactive data exploration, while ad-hoc analysis and
possibly discover hidden trends and patterns in health data.
Jim & Achim (2008) argued that paper still plays an
important role in todays and future BI-lifecycle. This makes the
authors to hypothesis that working with a digital pen together with
the specially designed graphical elements and forms can enhance
work in the healthcare domain, because of the digitalization of the
contents. The investigations were operated in the context of a
Business Intelligence (BI) lifecycle, where content is captured,
analyzed, utilized, and re-annotated on paper printouts. The
authors developed architecture for acoustic digital pen integration
into a healthcare environment and implemented a prototype for
evaluation reasons. Form elements were classified while current
processing of digital pen forms was also analyzed. The whole system
was evaluated to gather first impressions from end users. The basic
question arises, how forms can be designed that the process of note
taking is getting faster, lesser and well arranged so that nurses
and physicians have more time for the patients. In addition, the
authors came out that when designing forms for healthcare, three
parties need to be involved; the nurses or whoever has to work with
the form to assure ergonomic and usability requirements, the
management or whoever has knowledge about data integration and
process optimizing opportunities and the physician or whoever is
responsible for quality management and legal concerns. This will
make the acoustic digital pens seem to provide a cost-effective
opportunity to bridge the gap between physical paper records and
the digital representation of them.
-
Lihong, et al (2010) submitted that Extract-Transform-Loading
(ETL) tools integrate data from source side to target in building
data warehouse. However data structure and semantic heterogeneity
exits widely in the enterprise information systems. On the purpose
of eliminate data heterogeneity so as to construct data warehouse,
the authors introduced domain ontology into ETL process of finding
the data sources, defining the rules of data transformation, and
eliminating the heterogeneity. They embedded domain ontology in the
metadata of the data warehouse which led to data record mapped from
data bases to ontology classes of Web Ontology Language (OWL). This
resulted to access information resources more efficiently. The
authors tested the method in a hospital data warehouse project, and
the result shows that ontology method plays an important role in
the process of data integration by providing common descriptions of
the concepts and relationships of data items, and medical domain
ontology in the ETL process is of practical feasibility.
Xuezhong, et al (2008) suggested that the clinical data from the
daily clinical process, which keeps to traditional Chinese medicine
(TCM) theories and principles, is the core empirical knowledge
source for TCM researches. The authors introduced a data warehouse
system, which is based on the structured electronic medical record
system and daily clinical data, for TCM clinical researches and
medical knowledge discovery. The system consists of several key
components: clinical data schema, extraction-transformation-loading
tool, online analytical analysis (OLAP) based on Business Objects
(a commercial business intelligence software), and integrated data
mining functionalities. Their data warehouse is currently contained
20,000 inpatient data of diabetes, coronary heart disease and
stroke, and more than 20,000 outpatient data. Conclusively, their
analysis applications showed that the developed clinical data
warehouse platform is promising to build the bridge for TCM
clinical practice and theoretical research, which will promote the
related TCM researches.
13. METHODS Therefore, paper uses the Business Intelligence
Roadmap methods in designing the Healthcare BI for PKU in UUM
(Larissa & Shaku, 2003). The adopted method consists of six
stages; Justification Stage, Planning Stage, Business Analysis
Stage, Design Stage, Construction Stage and Deployment Stage, shown
in Figure 5.
Fig. 5: Business Intelligence Roadmap (Sources: Larissa &
Shaku, 2003)
-
The PKUBI system starts with investigation of the highlighted
problem of making decision by the policy makers (Chief Medical
Officer, Doctors, matrons and medical laboratory officers) in UUM
and business opportunity that need BI solution which was discovered
during the interview with the chief medical officer of PKU, Dr.
Sanuri. Each BI application should be cost-justified and should
clearly define the benefits of either solving a business problem or
taking advantage of a business opportunity in PKU. This method
proceeds to the certification of infrastructure that PKU in UUM has
on ground for the development of BI and preparation of the needs
for the application. The infrastructure may include hardware,
software, middleware, Meta data repository and network components.
In addition, organizations of the staffs, budgets, and
technologies, business representatives of PKU in UUM which must be
in detail are closely reported.
Furthermore, the business analysis stage for PKU in UUM has four
phases, such as project requirement definition, data analysis, and
application prototyping and Meta data repository analysis.
Moreover, the design stage consists of database design, extract
transform load (ETL) design and Meta data repository design for the
PKU BI system. The design has to meet the requirements of the
logical Meta model and take processes of SQL Server Integration
Services (SSIS), SQL Server Analytical Services (SSAS) and SQL
Server Report Services (SSRS). The conclusion part has to do with
the development of ETL, application, data mining and the Meta data
repository for the PKU BI application in UUM. Therefore, the BI
system then deploys for evaluation by the doctors, matrons and the
laboratory technologists during the deployment period.
14. ANALYSIS AND DESIGN OF PKUBI This is the processes of
analyzing and designing BI system for PKU in UUM. It displays
stepwise development of BI application and answers the measures
from the fact table of the dimension tables. Therefore, System
analysis is the dissection of a system into its components for the
purposes of studying how those components interact and work.
14.1 Dimensional Model IBM (2006) submitted that to overcome
performance issues for large queries in the data warehouse, we use
dimensional models. The dimensional modeling approach provides a
way to improve query performance for summary reports without
affecting data integrity. A dimensional model is also commonly
called a star schema. This type of model is very popular in data
warehousing because it can provide much better query performance,
especially on very large queries, than an E/R model. However, it
also has the major benefit of being easier to understand. It
consists, typically, of a large table of facts (known as a fact
table), with a number of other tables surrounding it that contain
descriptive data, called dimensions. When it is drawn, it resembles
the shape of a star. The dimensional model consists of two types of
tables having different characteristics like Fact table and
Dimension table.
14.2 Dimensional Tables and Models for PKUBI Data warehouses are
built using dimensional data models which consist of fact and
dimension tables. Dimension tables are used to describe dimensions;
they contain dimension keys, values and attributes. For example,
the time dimension would contain every hour, day, week, month,
quarter and year that has occurred since the beginning of business
operations. Product dimension could contain a name and description
of products you sell, their unit price, color, weight and other
attributes as applicable. Meanwhile, Dimension tables are typically
small, ranging from a few to several thousand rows which can grow
fairly large (sqlserverpedia.com). Therefore, the dimensional
tables for PKUBI in UUM are students
-
patient, staffs patient, familys patient, publics patient and
diseases patient, time, doctor and laboratory technologist.
Student Patient Dimensional Table Table 1: Student Patient
Dimensional Model Definition Attribute Description
Student_patient_id The unique identification for the students
patient. Student_patient_name This is the name of students patient.
Student_patient_age This is the age of students patient.
To generate the report of students patient based on age. To give
a correct treatment based on age.
Student_patient_gender This is the gender of students patient.
To generate the report of students patient based on gender. To give
a correct treatment based on gender.
Student_patient_college This is the college of study of students
patient. To ease the contact of students patient in terms of
need.
Student_patient_natinality This is the country o origin of
studentpatient. To know if the students patient ailment is peculiar
to his or her country of origin.
Student_patient_contactnum Contact number of students patient.
To contact students patient for further treatment through
telephone.
Student_patient_treatment To identify number of students patient
with similar diseases and current syndrome.
Student_patient_appointment To record any appointment.
To easily track the health history of students patient.
Student_patient_department This department o the students
patient.
Staffs Patient Dimensional Table Table 2: Staffs Patient
Dimensional Model Definition Attribute Description
Staff_patient_id The unique identification for the staffs
patient. Staff_patient_name This is the name of Staffs patient.
Staff_patient_address This is the address of staffs patient in
order to follow-up through hard copy. Staff_patient_age This is the
age of staffs patient.
To generate the report of staffs patient based on age.
-
To give a correct treatment based on age.
Staff_patient_gender This is the gender of staffs patient. To
generate the report of staffs patient based on gender. To give a
correct treatment based on gender.
Staff_patient_contactnum Contact number of staffs patient. To
contact staffs patient for further treatment through telephone.
Staff_patient_treatment To identify number of patient with
similar diseases and current syndrome.
Staff_patient_appointment To record any appointment.
To easily track the health history of staffs patient.
Familys Patient Dimensional Table Table 3: Familys Patient
Dimensional Model Definition Attribute Description
family_patient_id The unique identification for the familys
patient. family_patient_name This is the name of familys patient.
family_patient_address This is the address of familys patient in
order to follow-up through hard copy. family_patient_age This is
the age of familys patient.
To generate the report of familys patient based on age. To give
a correct treatment based on age.
family_patient_gender This is the gender of familys patient. To
generate the report of familys patient based on gender. To give a
correct treatment based on gender.
family_patient_contactnum Contact number of familys patient. To
contact familys patient for further treatment through
telephone.
family_patient_treatment To identify number of patient with
similar diseases and current syndrome.
family_patient_appointment To record any appointment.
To easily track the health history of familys patient.
Publics Patient Dimensional Table Table 4: Publics Patient
Dimensional Model Definition Attribute Description
public_patient_id The unique identification for the publics
patient. public_patient_name This is the name of publics patient.
public_patient_address This is the address of publics patient in
order to follow-up through hard copy.
-
public_patient_age This is the age of publics patient. To
generate the report of publics patient based on age. To give a
correct treatment based on age.
public_patient_gender This is the gender of publics patient. To
generate the report of publics patient based on gender. To give a
correct treatment based on gender.
public_patient_contactnum Contact number of publics patient. To
contact publics patient for further treatment through
telephone.
public_patient_treatment To identify number of patient with
similar diseases and current syndrome.
public_patient_appointment To record any appointment.
To easily track the health history of publics patient.
Disease Dimensional Table Table 5: Disease Dimensional Model
Definition
Attribute Description
Disease_id The unique identification for the diseases.
Disease_name The name of the diseases for proper
identification.
Disease_type To know the group the diseases belong to.
Disease_class To know the class that diseases belong to.
To know either the disease is major or minor.
Disease_control To know the specific drug for controlling the
diseases.
Time Dimensional Table Table 6: Time Dimensional Model
Definition
Attribute Description
Time _id The unique identification for each patient that
receives treatment.
Year The year that the patient receives treatment.
To generate report based on year.
Month The month that the patient receives treatment.
To generate report based on month.
Day The day that the patient receives treatment.
To generate report based on day.
-
Doctor Dimensional Table Table 7: Doctor Dimensional Model
Definition
Attribute Description
Doctor_id Unique identification for each doctor.
Doctor_name Name of doctor.
Doctor_gender To assign doctor for each patients treatment. (a
special case which is requested by patient)
Doctor_department To know the department of each doctor.
Doctor_specialisation To know the specialization of each
doctor.
To assign a correct doctor to a correct patient.
Laboratory Technologist Dimensional Table Table 8: Laboratory
Technologist Dimensional Model Definition
Attribute Description
technologist_id Unique identification for each technologist.
technologist_name Name of the technologist that attend to
patient.
technologist_gender To assign technologist for each patients
diagnosis. technologist_specialisation To know the specialization
of each technologist.
To assign the right technologist to the right patient.
Nationality Dimensional Table Table 9: Nationality Dimensional
Model Definition
Attribute Description
Nationality_id Unique identification for each nationality
Nationality_name Name of country of origin of the patient
College Dimensional Table Table 10: College Dimensional Model
Definition
Attribute Description
College_id Unique identification for each college
College_name Name of college of the patient
-
Drug Dimensional Table Table 11: Drug Dimensional Model
Definition
Attribute Description
Drug_id Unique identification for drug given to the patient
Drug_name Name of the recommended drug
Drug_expiringdate The period the drug will be expired
14.3 Requirements for PKUBI An interview was granted the chief
medical officer, Dr. Sanuri and Madam Ashia of PKU University Utara
Malaysia about their expectation on the requirements of the PKUBI.
These requirements are the measures of the PKUBI for making a
managerial decision on the important records in the data warehouse
repository of PKU. The following are the requirements gathered from
the management of the PKU:
Table 12: Requirements Detail for PKUBI
No Analysis Pass Fail
1 Which of the drug consume most by the patient? 1 0
2 Which of the drug is expiring in the next three months? 1
0
3 Which set of patients patronize PKU most? 1 0
4 What is the department of the most patronized patient in PKU?
1 0
5 What is the nationality of the most patronized patient in PKU?
1 0
6 What is the college of the most patronized patient in PKU? 1
0
7 How many patients have skin disease in a period of time? 1
0
8 Which of the diseases is most common among the patients in
PKU? 1 0
-
14.4 PKUBI Star Schema A Star schema has one fact table and
several dimension tables based on the PKUBI requirements.
Fig. 6: A PKU Star Schema
-
15. RESULTS
Fig. 7: Distribution of Patients According To the
Nationality
The Figure 7 above shows the distribution of patients that have
visited the PKU according to their nationality from 2006 to 2010.
It shows that Nigerians visited the PKU 32, 20, 24, 43 and 60 times
from the year 2006, 2007, 2008 and 2010 respectively, while
Malaysian visited PKU in 68, 46, 56, 65, and 80 times from 2006 to
2010. Also, Thai students visited PKU for treatment in 56, 23, 56,
76 and 87 times in the year 2006, 2007, 2008, 2009 and 2010
respectively.
Fig. 8: The Cube for the Patients According To the Nationality
in PKU
The Figure 8 shows how the important data like patients and
nationality distribution are filled in the cube. This is generated
from the reporting tool of the SQL server 2008 used for this
research.
-
Drug According to the Nationality
Fig. 9: Distribution of Use of Drug According To the
Nationality
Fig. 10: The Cube for the Use of Drug According To the
Nationality in PKU
-
communicable disease Non communicable disease
Fig. 11: Distribution of College According To the Diseases
Fig. 12: the Cube for the college according To the Diseases in
PKU
16. THE PKUBI DEPLOYMENT The PKUBI System has been successfully
implemented. All the functional requirements described before have
been fully achieved. The prototype initially developed for testing
has been fully converted to a working system. Front Page 2003 is
used as the Integrated Development Environment (IDE) and the back
end database was developed using Microsoft SQL Server 2008. The
Figures 13-17 (screenshots) show a sample of user interface.
-
Fig. 13: Login page of PKUBI
Fig. 14: Login page of PKUBI
The Figures 13 and 14 display the login page of the PKUBI. This
page allows the officer and the administrator in PKU to access some
functions of the PKUBI. Once the login button has been clicked, the
user will move to the home page of the system.
-
Fig. 15: Home page of the PKUBI
The home page of the PKUBI contains the pages that the users can
have access to; view page, report page and update page.
Fig. 16: Viewing page of the PKUBI
This page allows the administrator and the officer (authorized)
to view the record or profile of the patients through the PKUBI
system.
-
Fig. 17: Update page of the PKUBI
Figure 17 shows the update page which allows the user (database
officer) in PKU to update records of the patients in order for the
system to be effective.
17. RECOMMENDATION The importance of the PKUBI in decision
making cannot be overemphasized in achieving effective healthcare
set-up. Therefore, this calls for immediate recommendation of this
research in PKU. This research has helped in the style of service
delivery to the patients in PKU and helps in forecasting and
drilling of the drugs that need to be ordered for in large quantity
for pharmacy department at PKU.
18. CONCLUSION The design of BI system for PKU in UUM helps the
management by simplifying the technique needed for managerial
decision making and forecasting future activities that would help
the PKU. The PKUBI will also be useful to know the medical
statistics of the patients in UUM and the drugs that need to be
frequently ordered for. Moreover, this research has helped to
determine the diseases that require a crucial attention among the
patients at PKU in UUM.
-
REFERENCES
Adelman, S., & Larissa, T. M. (2000). Data Warehouse Project
Management. Boston, M. A: Addison-Wesley.
Berndt, D. J., Hevner, A. R., & Studnicki, J. (2000).
Hospital Discharge Transactions: A Data Warehouse Component.
Proceedings of the Hawaii International Conference on System
Sciences. 28-45.
Chaudhuri, S., & Dayal, U. (1997). Data Warehousing and OLAP
for Decision Support. ACM SIGMOD International Conference on
Management of Data. Tuscon, Arizona. 123 135.
Chaudhuri, S., Dayal, U., & Ganti, V. (2001). Database
Technology for Decision Support System. IEEE Explores, 34(12), 48
55.
Christos, V., Elia, E., & Penagiotis, C. (2006). An
OLAP-Enabled Software Environment for Modeling Patient Flow. 3rd
International IEEE Conference Intelligent Systems. 261 266.
Codd, E. F., Codd, S. B., & Salley, C. T. (1993). Providing
OLAP to User-Analysts: An IT Mandate. California, Codd &
Associates, Sunnyvale.
Demarco, T. (2001). Slack: Getting Past Burnout, Busywork, and
the Myth of Total Efficiency. New York: Broadway Books.
Diana, G., Maribel, Y. S., & Jorge, C. (2010). Analysis of
Quality Of Life after an Endoscopic Thoracic Sympathectomy: A
Business Intelligence Approach. Second International Conference on
Advances in Databases, Knowledge, and Data Applications. 1 6.
Ebidia, A., Mulder, C., Tripp, B., & Morgan, M. W. (1999).
Getting Data Out Of the Electronic Patient Record: Critical Steps
in Building a Data Warehouse For Decision Support. Boston, M. A:
Addison-Wesley.
Eckerson, W. (2003). Smart Companies in the 21st Century: The
Secrets of Creating Successful Business Intelligent Solutions. The
Data Warehousing Institute: Seattle, WA.
Efraim, T., Ramesh, S., Jay, E. A., & David, K. (2008).
Business Intelligence: A Managerial Approach. 1st Edition, Upper
Saddle River New Jersey: Pearson Prentice Hall.
Ewen, E. F., Medsker, C. E., & Dusterhoft, L. E. (1998).
Data Warehousing in an Integrated Health System, Building the
Business Case. International Conference of DOLAP, Washington DC.,
USA. 58.
Hristovski, D., Rogac, M., & Markota, M. (2000). Using Data
Warehousing and OLAP in Public Health Care. 3rd Edition, Upper
Saddle River New Jersey: Pearson Prentice Hall.
Hyperion Solution Coorporation, (2004). Management A Strategic
For Performance. Business Intelligence Journal, 2, 23.
Inmon, W. H. (1996). Building the Data Warehouse. New York: John
Wiley & Sons. Inmon, W. H. (2005). Building the Data Warehouse.
3rd Edition, New York: John Wiley & Sons. Inmon, W. H.,
Claudia, I., & Ryan, S. (2000). Corporate Information Factory.
2nd Edition, New York: John
Wiley & Sons. Jim, S., & Achim, P. K. (2008). Seamless
User-Interface for Business Intelligence. IEEE International
Conference on e-Business Engineering, 5, 728 731. John, K.
(2007). Cognos 8 Business Intelligence Dashboards. IBM Software
Magazine, 1. Kimball, R., & Richard, M. (2000). The Data
Warehouse Toolkit: Building the Web-Enabled Data
Warehouse. New York: John Wiley & Sons. Larissa, T. M.,
& Shaku, A. (2003). Business Intelligence Roadmap: The Complete
Project Lifecycle for
Decision Support Application. Information Technology Series,
Pearson Education Inc., USA. Addison-Wesley.
-
Ledbetter, C. S., & Morgan, M. W. (2001). Toward Best
Practice: Leveraging the Electronic Patient Record as a Clinical
Data Warehouse. Journal of Healthcare Information Management, Vol.
15, 119-131.
Lihong, J., Hongming, C., & Boyi, X. (2010). A Domain
Ontology Approach in the ETL Process of Data Warehousing. IEEE
International Conference on e-Business Engineering, 3, 30 35.
Lonnqvist, A., & Pirttimaki, V. (2006). The Measurement of
Business Intelligence. Information Systems Management Journal,
32-40.
Maria, A. M., & Abdel-Badeeh, M. S. (2010). Intelligent
Techniques for Business Intelligence in Healthcare. IEEE Explore,
545 550.
Microsoft, (2009). An Exclusive Report from Business Week
Research Services. Microsoft Magazine, 1 12.
Pedersen, T. B., & Jensen, C. S. (2001). Multidimensional
Database Technology. Journal of Computer, 34, 40-46.
Raisinghani, M. (2004). Business Intelligence in the Digital
Economy. Hershey PA: The Idea Group. Scheese, R. (1998). Data
Warehousing As a Healthcare Business Solution. Healthcare
Financial
Management, 52, 56-59. SQL Server Information. URL:
www.sqlserverpedia.com. Retrieved on 19/04/2011. Thompson, O.
(2004). Business Intelligence Success, Lessons Learned. Technology
Evaluation. China:
China Translation and Printing Limited. Umeshwar, D., Malu, C.,
Alkis, S., & Kevin, W. (2009). Data Integration Flows for
Business Intelligence.
New York: John Wiley & Sons, 1- 11.
Vassiliadis, P., & Sellis, T. (1999). A Survey of Logical
Models for OLAP databases. ACMSIGMOD Record, 28, 64-69.
Watson, H.J., & Wixom, B.H. (2007). The Current State of
Business Intelligence. IEEE Explore, 96-99. Xuezhong, Z., Baoyan,
L., Yinghui, W., Runsun, Z., Ping, L., Shibo, C.,Yufeng, G., Zhuye,
G., & Hua, Z.
(2008). International Conference on BioMedical Engineering and
Informatics, 5, 615 620. Yorozu, Y., Hirano, M., Oka, K., &
Tagawa, Y. (1987). Electron Spectroscopy Studies On
Magneto-Optical
Media and Plastic Substrate Interface. IEEE Transl. Journal
Japan, 2, 740741. Zaman, M. (2005). Business Intelligence: Its Ins
and Outs. URL:
www.technologyevaluationcenters.com/search/ retrieved on
05/03/2003.
Zhu, F. (2007). A Framework to Develop a University Information
Portal. IEEE Explores, 506-509.