EFFECTIVENESS OF A BUSINESS INTELLIGENCE SOLUTION
TO MANAGE THE ANTIRETROVIRAL THERAPY
PROGRAMME
Eduan Kotzé and Theo McDonald
Department of Computer Science and Informatics, University of the Free State, Bloemfontein, South Africa
Keywords: HIV/AIDS, Antiretroviral therapy, Data warehousing, Business intelligence, OLAP.
Abstract: The Human Immunodeficiency Virus and Acquired Immune Deficiency Syndrome (HIV/AIDS) has caused
the death of millions of people worldwide. To combat the effect of HIV/AIDS, the South African
government started with the provisioning of Antiretroviral Therapy (ART) in the public health sector.
Monitoring and evaluating the effectiveness of this ART programme is of the utmost importance. A
business intelligence approach was followed that first of all integrated several independent operational
sources into one data warehouse and then delivered strategic management information with easy to use
business intelligence tools that was developed and deployed for the users. The business intelligence solution
was then evaluated by the users of the system and the results indicated that the users deemed the solution to
be an effective way to obtain strategic information on the rollout of the ARV treatment programme.
1 INTRODUCTION
In response to the HIV/AIDS epidemic the South
African Government created the HIV/AIDS and
Sexually Transmitted Disease (STD) Strategic Plan.
In November 2003, after considerable sustained
pressure from advocacy groups, the government
adopted the Operational Plan for Comprehensive
HIV and AIDS Treatment and Care, which included
the provisioning of Antiretroviral Therapy (ART) in
the public health sector (National Department of
Health, 2003).
Monitoring and evaluating the effectiveness of
this ART programme is of the utmost importance. In
this endeavour strategic information plays a major
role. The Free State Department of Health (FSDOH)
decided that strategic information must be provided
through a business intelligence solution. How this
business intelligence solution was developed and
evaluated to determine its effectiveness forms the
basis of this paper.
2 RESEARCH METHODOLOGY
The research methodology used by this study was
action research. Butler, et al. (2006) noted that in
action research projects, researchers collaborate with
practitioners to solve practical problems while
expanding scientific knowledge. Baskerville (1999)
characterizes information system action research as
an increased understanding of an immediate social
situation, with emphasis on the complex and
multivariate nature of this social setting in the
information systems (IS) domain. It simultaneously
assists in practical problem solving and expands
scientific knowledge.
The action research description (Susman &
Evered, 1978) details a five phases, cyclical process.
The approach first requires the establishment of a
client-system infrastructure or research environment.
Then, five identifiable phases are iterated: 1)
diagnosing, (2) action planning, (3) action taking,
(4) evaluating and (5) specifying learning.
Baskerville (1999) provides an explanation of these
components. The client-system infrastructure is the
specification and agreement that constitutes the
research environment and provides the conditions
under which action and change may be specified.
The client will be the FSDOH and the researcher
will be the author of this paper. The research
environment will be the Free State Province and will
include staff working for the FSDOH at different
levels of management.
543
Kotzé E. and McDonald T..
EFFECTIVENESS OF A BUSINESS INTELLIGENCE SOLUTION TO MANAGE THE ANTIRETROVIRAL THERAPY PROGRAMME.
DOI: 10.5220/0003150905430546
In Proceedings of the International Conference on Health Informatics (HEALTHINF-2011), pages 543-546
ISBN: 978-989-8425-34-8
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
3 BACKGROUND
Strategic information is an essential requirement in
the fight against HIV/AIDS. The lack of strategic
information is, however, apparent if a closer look is
taken at the Free State antiretroviral (ARV) patient
information system. The patient information system
was a traditional online clinical system, dealing with
operational issues of accumulating data on a patient.
Very little functionality was provided to deal with
the complexities of managing the clinical outcomes
of the ART programme. To add to the problem,
other independent online transaction processing
(OLTP) systems, all closely related to HIV/AIDS,
had to be interrogated to gain an understanding of
the impact the rollout of ARV drugs had. Examples
of these systems were standalone human resource
systems, information systems accumulating data on
tuberculosis, patient admissions, notifiable diseases
and blood tests.
As a first step to obtain strategic information on
the rollout of the ARV treatment programme, the
Hospital Information System and all the other
disparate information systems were integrated in one
data warehouse. This had the advantage that all the
disparate systems could be interrogated using the
same tool. Next a BI analytical tool was used to
provide strategic information for the FSDOH ART
management team. The research question, therefore,
was to determine if this BI solution was effective.
4 BUSINESS INTELLIGENCE (BI)
APPROACH
The current study is the first, as far as can be
determined, that attempted a data warehouse that
integrated several standalone systems which are all
closely related to ART.
The main focus was Antiretroviral Treatment and
therefore the data warehouse had to contain an
antiretroviral clinical data mart. Data marts for other
systems closely related to ART (i.e. Blood Tests,
TB, Notifiable Diseases and Hospitalization) also
had to be included in the data warehouse. To
complicate matters, the FSDOH had a separate
challenge in optimizing key internal business
processes such as human resource management and
revenue collection. Data marts for these processes
were also required and had to be developed.
Strategic information was required by the respective
managers to improve the business performance for
each of these business processes. Figure 1 illustrates
diagrammatically the data warehouse, resulting from
several interdependent data marts (Gray and Watson,
1998:104-105).
Figure 1: Data warehouse framework.
The idea behind business intelligence (BI) is to
turn data into information and then into knowledge
(Golfarelli, Rizzi and Cella, 2004). Analytical tools
do this by accessing and analyzing information
contained in the data warehouse. These tools include
reporting, querying, on-line analysis and
exploration, visualization, decision modeling and
planning, and data mining tools.
The FSDOH BI solution can be divided into two
parts: ad-hoc query and OLAP cube analysis. Ad-
hoc query functionality was made available using
Cognos Query Studio and Cognos Report Studio.
OLAP functionality was made available using the
Web-based front-end interface of Cognos PowerPlay
Enterprise Server.
By deploying Cognos BI functionality, security
and privacy was also addressed by using Cognos
Framework Manager. Cognos Framework Manager
captures the BI metadata (including the dimensional
model and database model) as well as which user is
allowed to access which portion of the BI metadata.
Cognos role-based authentication was used for any
user logging into Cognos Portal. The cubes and drill-
down reports were only made available to users that
successfully logged into Cognos Portal and the
user’s credentials were matched in the background
by Cognos Access Manager to determine what the
user was allowed to browse and drilldown into.
HEALTHINF 2011 - International Conference on Health Informatics
544
5 BI EVALUATION
Wixom and Watson (2001) investigated several
implementation success factors affecting data
warehouse success. Shin (2003) expanded on the
work done by Wixom and Watson (2001) and
examined the success factors in data warehousing by
using system quality, information quality, service
quality and user satisfaction as variables.
A questionnaire was used as the instrument to
evaluate the BI implementation. Some of the
questions of Shin (2003) were re-used in the
questionnaire and new ones were added that were
applicable to the FSDOH data warehouse evaluation.
The questionnaire utilized a 5 point Likert scale
and consisted of three sections. Section 1 covered all
the basic demographics of the respondents. Section 2
looked at how the respondents used the existing data
warehouse while section 3 covered the perceptions
of respondents on the information from the data
warehouse.
Data was collected from data warehouse users in
the FSDOH which included ART, human resources,
revenue collection and hospital managers. All users
received a questionnaire if they were either using the
data warehouse themselves or request information
from the knowledge workers who extract
information for them from the data warehouse on a
regular basis. The selected group was well
represented over the five layers of employment at
the FSDOH namely: production workers,
supervisors, assistant managers, middle managers
and top managers.
A total of 87 questionnaires were sent to this
selected group of whom 51 responded. This
translated into a response rate of 58.6%. Three (3) of
the 51 respondents’ questionnaires were incomplete
and discarded from the study. The final number of
completed questionnaires to be used for analysis was
48.
6 RESULTS
Most users (56.2%) accessed the system either
monthly or quarterly. Six users used the system on a
daily basis. Eight main organizational tasks were
included in the survey. The first four (decision-
making support, status monitoring, planning, and
forecasting) were considered more unstructured than
the others (administration, accounting, resource
allocation/budgeting and personnel management).
Personnel management (62.5%) stood out as the
most frequent task while forecasting (45.9%) was
the least performed task.
Direct and indirect usage was also investigated.
Most users (67.2%) indicated that they make use of
either an assistant or knowledge workers at head
office to obtain the information from the data
warehouse for them. It is worth mentioning that the
knowledge workers at head office were provided
with certified training that was offered by Cognos
South Africa. These courses empowered them to
assist users with analysis requests.
The remaining users (32.8%) retrieve the data by
themselves and perform their own analysis. Most of
these users attended an in-house business
intelligence course, which introduced them to basics
of Cognos reporting and Cognos cube analysis.
For the unstructured tasks, most of the survey
respondents would use the data warehouse
sometimes or never while for the structured tasks the
usage would be from very frequently to sometimes.
According to the study done by Shin (2003), more
users were using the data warehouse for unstructured
duties rather than for routine or administrative
responsibilities. For this study the weight tends to be
for structured tasks instead of unstructured tasks.
This was an unexpected finding. A possible
explanation for this finding could be that the BI
maturity of the survey respondents was much lower
than the respondents used in the study by Shin
(2003). This possibility is supported with the finding
that 67.2% of respondents indicated they make use
of either an assistant or knowledge worker to obtain
information from the data warehouse instead of by
them self.
Respondents were on the whole very positive
about data quality, levels of detail and accuracy.
Most respondents (89.6%) agreed that the data in the
data warehouse is current enough to meet work
needs. That was matched by 75% who disagreed that
the data warehouse was out of date for a similar
question that was negatively phrased.
A total of 79.2% of the respondents indicated
that the data warehouse maintains data at an
appropriate level of detail to perform their tasks.
This was matched by 68.1% who disagreed that the
data warehouse does not have enough detail to make
them more productive.
Most respondents (70.3%) indicated that the data
in the data warehouse is accurate and reliable and
this was matched by (81.9%) who were either unsure
or disagreed that the data is inconsistent.
Next the data warehouse was evaluated in terms
of functionality, flexibility, processing speed and
ease of use. Respondents were on the whole very
EFFECTIVENESS OF A BUSINESS INTELLIGENCE SOLUTION TO MANAGE THE ANTIRETROVIRAL
THERAPY PROGRAMME
545
positive about these aspects. A large number
(78.9%) of respondents indicated that they were
satisfied with the overall functionality of the data
warehouse. This was matched by 80.4% of
respondents who indicated that they were either
unsure or did not agree that the data warehouse had
no functional value to them.
Again a large number (68.8%) of respondents
indicated they were satisfied with the overall
flexibility of the data warehouse and that was
matched by 68.7% who disagreed that they cannot
perform their own analysis. A total of 68.7% of the
respondents indicated that the data warehouse
processing speed is good, but the negative question
had a relative high number (41.7%) who indicated
they were unsure. This could be due to the fact that
network speed and bandwidth restrictions placed an
uncertainty in the users’ minds. A large number
(79.2%) of respondents indicated that the data
warehouse is convenient and easy to use, but
interestingly 78.7% also indicated that more training
is needed to find, understand and use the data
warehouse.
Finally, 93.8% of the respondents indicated that
overall the data warehouse is a valuable asset for the
FSDOH and it is recognized as being a critically
important tool to improve the productivity of
knowledge workers by providing strategic
information.
7 CONCLUSIONS
Strategic information is an essential requirement in
the fight against HIV/AIDS. In the Free State a
business intelligence approach was followed in order
to provide strategic information for the management
of the ART programme in the province. This was
achieved by integrating several disparate systems
into one appropriately designed data warehouse,
specifically focused on ART.
By using OLAP technology it was possible for
the FSDOH ART management team to extract
strategic information by means of easy to use
analytical tools. Finally, the business intelligence
solution was evaluated by the users of the system.
The results indicated that the users were on the
whole very positive about the data quality,
responsiveness, functionality, flexibility and ease of
use. Overall the system was deemed a huge success.
Taking all the results into consideration, it can
therefore be concluded that BI solution was
effective.
REFERENCES
Baskerville, R. L. 1999. Investigating Information
Systems with Action Research. Communications of the
Association for Information Systems, 2(19). Retrieved
17 September 2007 from the World Wide Web:
http://cais.isworld.org/articles/2-19/default.asp?
View=pdf&x=78&y=17.
Butler, T., Feller, J., Pope, A., Murphy, C. and Emerson,
B. 2006. An action research study on the design and
development of core IT artifacts for knowledge
management systems. http://csrc.lse.ac.uk/asp/aspecis
/20060044.pdf Accessed 8 October 2007.
Golfarelli, M., Rizzi, S., Cella, I. 2004. Beyond data
warehousing: What’s next in business intelligence?
Proceedings of the 7
th
ACM international workshop on
Data warehousing and OLAP. New York, pp1-6.
Retrieved 28 November 2006 from the World Wide
Web: http://doi.acm.org/10.1145/1031763.1031765.
Gray, P. and Watson, H. J. 1998. Decision Support in the
Data Warehouse. Upper Saddle River, NJ, Prentice-
Hall Publishing.
National Department of Health, South Africa. 2003.
Cabinet’s decision on the operational plan for
comprehensive care and treatment of people living
with HIV and AIDS. Pretoria. South Africa. Retrieved
16 July 2008 from the World Wide Web:
http://www.doh.gov.za/docs/pr/2003/pr1119.html.
Shin, B. 2003. An Exploratory Investigation of System
Success Factors in Data Warehousing. Journal of the
Association for Information Systems, 4, 141-170.
Susman, G. and Evered, R. 1978. An Assessment of the
Scientific Merits of Action Research. Administrative
Science Quarterly, 23(4), 582-603.
Wixom, B. and Watson, H. 2001. An empirical
investigation of the factors affecting data warehousing
success. MIS Quaterly, 28(1), 17-41.
HEALTHINF 2011 - International Conference on Health Informatics
546