International Journal of Information Management 34 (2014) 20–27

Contents lists available at ScienceDirect

International Journal of Information Management

journal homepage: www.elsevier.com/locate/ijinfomgt

Towards an implementation framework for business intelligence in

healthcare
Neil Foshay a,∗ , Craig Kuziemsky b
a
Schwartz School of Business, St. Francis Xavier University, PO Box 5000, Antigonish, NS B2G 2W5, Canada
b
Telfer School of Management, University of Ottawa, 55 Laurier Avenue East, Ottawa, ON K1N 6N5, Canada

a r t i c l e i n f o a b s t r a c t

Article history: As healthcare organizations continue to be asked to do more with less, access to information is essen-
Available online 1 November 2013 tial for sound evidence-based decision making. Business intelligence (BI) systems are designed to
deliver decision-support information and have been repeatedly shown to provide value to organizations.
Keywords: Many healthcare organizations have yet to implement BI systems and no existing research provides a
Healthcare healthcare-specific framework to guide implementation. To address this research gap, we employ a case
Business intelligence
study in a Canadian Health Authority in order to address three questions: (1) what are the most sig-
Implementation success factors
nificant adverse impacts to the organization’s decision processes and outcomes attributable to a lack of
Information quality
Organizational issues
decision-support capabilities? (2) what are the root causes of these impacts, and what workarounds do
they necessitate? and (3) in light of the issues identified, what are the key considerations for healthcare
organizations in the early stages of BI implementation? Using the concept of co-agency as a guide we iden-
tified significant decision-related adverse impacts and their root causes. We found strong management
support, the right skill sets and an information-oriented culture to be key implementation considerations.
Our major contribution is a framework for defining and prioritizing decision-support information needs
in the context of healthcare-specific processes.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction and motivation
Healthcare organizations are under ever increasing pressure to
do more with less and are continuously seeking ways to ensure
that resources are deployed as efficiently as possible while ensuring
high quality patient care (Hanson, 2011). Information is essential
to meeting these goals – it has been referred to as the lifeblood
of healthcare as it is essential for effective clinical and administra-
tive decision making (Pine et al., 2012; Toussaint & Coiera, 2005).
Healthcare decision making is complex and requires access to a
wide array of high-quality information (Sen, Banerjee, Sinha, &
Bansal, 2012). Business intelligence (BI) is defined as the use of
information and specialized analytical tools to enable informed
decision making in a variety of organizational contexts (Negash,
2004; Rohloff, 2011). A key characteristic of BI is that it integrates
data from a wide variety of internal and external sources, thus pro-
viding an effective information platform for healthcare decision
makers (Mettler & Vimarlund, 2009).
It is widely acknowledged that BI can provide benefits to health-
care organizations including improved patient care and outcomes
∗ Corresponding author. Tel.: +1 902 867 5425.
E-mail addresses: nfoshay@stfx.ca (N. Foshay),
Kuziemsky@telfer.uottawa.ca (C. Kuziemsky).
(Tremblay, Hevner, & Berndt, 2012), effective utilization of human
resources (Crist-Grundman & Mulrooney, 2011), improved process
efficiency (Flower, 2006) and cost avoidance (Pine et al., 2012;
Wang, Nayda, & Dettinger, 2007). Despite these potential bene-
fits, many healthcare organizations have not yet implemented BI
systems (Hanson, 2011) and there has been very limited research
on the factors that contribute to the successful implementation of
BI in a healthcare-specific context. Further, numerous studies have
highlighted the notion that information systems are notoriously
difficult to implement in healthcare organizations. The overarching
goal for our study is to identify the most critical factors that should
be addressed by healthcare organizations that are in the early stages
of BI system implementation, thus addressing a significant gap in
existing research.
To achieve our goal we employ a case study in the Guysborough
Antigonish Strait Area Health Authority (GASHA in Nova Scotia,
Canada. Through this case study, we seek to answer three research
questions:
1. What are most significant adverse impacts to the healthcare
organization’s decision processes and outcomes attributable to
a lack of decision-support capabilities?
2. What are the root causes of these adverse impacts, and what are
workarounds for dealing with them?

0268-4012/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijinfomgt.2013.09.003
 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
3. In the context of questions 1 and 2, what are the most impor-
tant considerations for BI system implementations in healthcare
organizations?
By addressing question 1, we seek to identify some of the key
decision-oriented adverse impacts attributable to the absence of
decision-support capabilities within organizations like GASHA, and
by extension, to highlight the benefits that accrue to healthcare
organizations by implementing BI systems. Regarding question 2,
we draw upon the concept of co-agency to identify the information,
process and personnel-related issues that caused the impacts and
identify the workarounds that are required because of these issues.
Regarding question 3, we develop a framework which clearly iden-
tifies core processes and facilitates the definition and prioritization
of decision-support information needs in the context of these pro-
cesses.
2. Theoretical background
The ultimate goal for our study is to identify the most criti-
cal factors that should be addressed by healthcare organizations
that are in the early stages of BI system implementation. In this
section, we review relevant literature on Information Systems (IS)
implementation considerations factors in general terms and in the
specific context of healthcare. We then review BI implementation
considerations.
2.1. IS implementation considerations
A significant body of research has focused on IS implementa-
tion considerations in the context of user adoption. Socio-technical
and diffusion of innovation (DOI) theories are influential in this
research area. The socio-technical theory states that an information
system is composed of two interrelated subsystems: the technical
(technology and tasks required convert system inputs into out-
puts) and the social (users and their characteristics and needs, as
well as structure of the organization in which the system is being
implemented). In order to achieve the benefits anticipated from
the implementation of an IS, the socio-technical approach proposes
that (a) interdependence of the subsystems must be recognized
and (b) the design of the overall system must aim to jointly opti-
mize both subsystems (Bostrom & Heinen, 1977). System designers
must identify how the subsystems impact each other and must
ensure the subsystems work in harmony in order to fully realize
anticipated benefits (Mattia, 2011). Implicit to socio-technical the-
ory is that effective user participation in the systems development
process is essential to implementation success (Hartwick & Barki,
1994).
Numerous studies have leveraged Rogers’ (1995) diffusion of
innovations (DOIs) theory to explain factors that contribute to
implementation success. DOI theory proposes that innovations
are communicated throughout organizations through a variety of
formal and informal channels over time. Individuals in the orga-
nization are viewed as having different levels of willingness to
adopt the innovation. The speed of adoption is positively influenced
by four factors: relative advantage, compatibility, trialability and
observability and negatively by the complexity of the innovation.
DOI theory has been adapted for IS research in numerous ways. For
example, Moore and Benbasat (1991) expanded Rogers’ five fac-
tors to eight: voluntariness of use of the system, relative advantage,
task compatibility, system image, ease of use, result demonstrabil-
ity, visibility, and trialability. Agarwal and Prasad (1998) argue that
the specific characteristics of the innovation (e.g. relative advan-
tage, visibility and result demonstrability) as well as the degree
of voluntariness of use of the system are the key determinants of
21
implementation success. IS Research has consistently found that
technical compatibility, technical complexity, and relative advan-
tage are important antecedents to system adoption (Bradford &
Florin, 2003).
2.2. Health information system implementation considerations
While IS implementation in many fields has achieved a cer-
tain level of success as described above, healthcare information
system (HIS) implementation has proven problematic. While stud-
ies have advocated positive outcomes from HIS (McKibbon et al.,
2012), there is a far more substantial body of research reporting
on negative outcomes including workflow, communication, and
safety issues (Ash, Berg, & Colera, 2004; Harrison, Koppel, & Bar-Lev,
2007; Kaplan & Harris-Salamone, 2009). Negative consequences
occur because a HIS interacts with the people and processes that
are part of the environment in which the HIS is implemented.
While HISs are designed to improve specific processes (i.e. deci-
sion making), studies have shown that underlying information and
process issues can impede HISs from achieving their full potential
and that these issues must be managed pre-implementation (Tariq,
Georgiou, & Westbrook, 2013). A systematic review on HIS diffusion
identified a lack of studies on readiness (i.e. what an organization
can do to assess and anticipate the impact of HIS implementa-
tion) as a key research gap (Greenhalgh, Robert, Macfarlane, Bate,
& Kyriakidou, 2004). In other words the environment needs to be
studied pre-implementation to identify any issues that will lead
to negative unintended consequences. Understanding underlying
HIS implementation issues requires a multi-dimensional approach.
The concept of co-agency (Thraen, Byron, Mullin, & Weir, 2012)
suggests that healthcare processes (i.e. decision making) cannot be
viewed as discrete or isolated events but rather need to be studied
as a set of interactions between processes, people and technol-
ogy. We employ the co-agency concept in this study to identify
key implementation considerations for BI in healthcare.
2.3. BI implementation considerations
Wixom and Watson (2001) developed a BI-specific implemen-
tation success model that proposes that implementation factors
(strong management support, a visible business champion, suffi-
cient resources, effective user participation, appropriate technical
team skills and source system data quality) serve to positively
influence implementation success from three perspectives: organi-
zational, project and technical. They make the point that BI systems
are not IT applications in the traditional sense; rather they are often
an enabler of different applications. Seah, Hsieh, and Weng (2010)
also highlight strong support and leadership from top management
as a key success factor.
Ramamurthy, Sen, and Sinha (2008) draw on DOI theory and
view BI as a major IT infrastructure innovation. They propose
that implementation success is dependent upon organizational
factors such as management commitment, organization size and
absorptive capacity as well as characteristics of the innovation
(the BI system) including relative advantage and low complexity.
Isik, Jones, and Sidorova (2011) contend that successful BI imple-
mentations require specific capabilities including high quality
data, appropriate user access and effective integration with other
systems. Further, the specific decision environment (types of deci-
sions, operational context) in which the BI system is implemented
must be cleared defined and understood. Yeoh and Koronios (2010)
describe a similar concept, “business orientation” – the alignment
of BI with business goals and strategy, as an important success
factor.
Popovic, Hackney, Coelho, and Jaklic (2012) identify BI matu-
rity and identified analytical decision-making and culture as key
22 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
success factors. The authors measure BI maturity in terms of data
integration effectiveness (the ability to collect, cleanse and trans-
form data from disparate sources) and analytic capabilities (support
for a wide variety of analytic use-cases, for example OLAP, data
mining and reporting). Sen, Ramamurthy, and Sinha (2012) also
view BI maturity as a success factor. They state that ‘process matu-
rity’ is critically important and propose a Capability Maturity Model
(CMM)-inspired five stage model covering both the development
and on-going operation of BI systems.
While studies of BI in healthcare exist (e.g., Ferrand, Amyot, &
Corrales, 2010; Tremblay et al., 2012), they have focused primarily
on the intended outcomes of using BI (i.e. improved decision mak-
ing) or tools for supporting BI. However, there are no studies that
have looked at implementation considerations for BI systems in a
healthcare context. BI systems are designed to improve decision
making capabilities but a significant shortcoming is that existing
healthcare studies make the assumption that the decision making
processes to be supported by BI systems (and required informa-
tion) are in place. The literature reviewed in Section 2.2 suggests
that this assumption may be invalid. Based on previous research in
HISs, people, process and information issues often cause negative
unintended consequences for HIS implementation.
2.4. Summary of literature findings
In this section we reviewed information system implementa-
tion considerations both in general and in healthcare, followed
by a review of BI implementation considerations, again in general
and in healthcare. Given the problematic history of HIS imple-
mentation it has been suggested that that a key research need is
studies that assess people and process readiness for HIS imple-
mentation (Greenhalgh et al., 2004). Our literature review did not
identify any pre-implementation studies for BI in healthcare. Given
the potential value that BI can provide to healthcare organiza-
tions coupled with the difficulty of implementing HISs, we believe
that a significant research gap is that no one has identified pre BI-
implementation considerations in healthcare, i.e. what a healthcare
organization needs to consider prior to implementing a BI system. We
address this gap in our study.
3. Materials and methods
3.1. Data source
Our study took place in the Guysborough, Antigonish, Strait-
Area Health Authority (GASHA), one of nine district health
authorities in the province of Nova Scotia, Canada. It serves a pop-
ulation of approximately 44,500 residents (approximately 5% of
the population of Nova Scotia) covering the Antigonish Town and
County, Guysborough County, Richmond County and the south-
ern portion of Inverness County through one primary facility and
four smaller community-based hospitals (GASHA Annual report,
2011). In recent years, GASHA has implemented two major enter-
prise systems: Meditech, which is used to support a wide variety
of operational processes within GASHA facilities and SAP, which is
used primarily for HR and accounting purposes. In addition to these
systems, the district has a 3M Health Record Management System
responsible for the capture of health information.
During preliminary discussions with the Authority’s senior
management team we learned that there were numerous deci-
sion support issues faced by decision makers. In order to explore
these issues, we conducted 40 face to face interviews from June
to August of 2011. Each interview was approximately an hour
in duration. Table 1 presents a summary of participants by func-
tional area and participant type. We identified three distinct groups
Table 1
Participant roles relative to information consumption and production (empty cells
represent a value of 0).
Participant roles Information Information Consumer/
consumers providers providers
Senior administration 5
Clinical managers 10
Diagnostic imaging 1
Finance/accounting/HR/procurement 4
Health records 3 1
IT support 2
Laboratory 1
Pharmacy 1
Physicians/Clinical Directors 8
Public health 1
Quality/risk/infection control 3
Totals 28 5 7
for our interview population (Information Consumers, Information
Producers and consumer/producers) and designed distinct inter-
view guides for each group. Information Consumers have jobs that
involve accessing information that has been provided by someone
else for analysis and management decision making purposes. Infor-
mation Producers are responsible for collecting, consolidating and
formatting data (e.g., into reports, Excel spreadsheets and Access
databases) in order to provide information to one or more manage-
ment decision makers, but does not play a decision making role.
Information consumer/producers are responsible for both roles.
Note that empty cells represent a value of 0. For example, all five
Senior Administrators in the study identified themselves as infor-
mation consumers – none were providers or consumer/providers.
Table 2 presents a number of key characteristics regarding our
participant group. Overall, participants were very experienced with
an average tenure at GASHA of 18.3 years. Participants were asked
to rate themselves on computer/information access proficiency on
a 5 point scale (1 = Novice, 5 = Expert). The average for the group
was 3.8, indicating a reasonably high overall proficiency. Partici-
pants were also asked to describe their decision making context:
whether they made decisions individually or as part of a group. We
found that decision context is split fairly evenly between individual
decision making and decisions made in a group setting.
3.2. Data analysis
Each interview was transcribed into a Word document, and
returned to participants to verify its completeness and accuracy.
Responses were then anonymized and consolidated in an Excel
workbook so that all responses to each question were grouped
together on individual sheets. Using this consolidated data, we per-
formed a detailed qualitative content analysis (Hseih & Shannon,
2005). We used an interpretative analytical approach, drawing
upon the principles of Miles and Huberman (1994), in order to study
BI implementation from the perspective of co-agency (Thraen et al.,
2012). Our analysis focused on identifying the agency (i.e. infor-
mation, people and process) issues that caused decision making
anomalies and the impacts (i.e. workarounds) that were caused by
these issues thus enabling us to propose BI-implementation success
factors for healthcare settings.
4. Results
Fig. 1 provides a map of how our results are presented in this sec-
tion. As per the description in Section 3.2 we first analyzed the data
to identify adverse impacts to the decision making process. We then
reverse engineered these adverse impacts to identify the informa-
tion, people and process (co-agency) issues that lead to the adverse
events and the implications of the adverse events. The numbers
 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27 23

Table 2
Participant profile.

Participant details Decision contexts

Participant group (# in group) Average years @ GASHA Computer/information access Individual Group
proficiency (1-Novice; 5 = Expert)

Senior administration (5) 24.7 3.4 47% 53%

Clinical Managers (10) 20.7 3.6 43% 57%
Diagnostic Imaging (1) 3.5 3.0 50% 50%
Finance/accounting/HR/procurement (4) 21.0 4.0 77% 23%
Health records (4) 8.5 5.0 80% 20%
Laboratory (1) 20.0 3.5 20% 80%
Pharmacy (1) 6.0 4.0 65% 35%
Physician/Clinical Directors (7) 19.0 3.4 19% 81%
Public health (1) 8.0 4.0 80% 20%
Public relations (1) 4.0 4.0 N/A N/A
Quality/infection control (3) 21.8 3.3 48% 52%
18.3 3.8 45% 55%

Italic values are overall averages for the 4 columns.

in Fig. 1 refer to specific sections in the text. We use quotations
to supplement the results. Each quotation is tagged with a label
identifying the source of the quotation from the participant list in
Table 1.
4.1. Adverse impacts to the decision making process
Our analysis identified numerous adverse impacts to the deci-
sion making processes within the health district attributable to
a lack of decision-support capabilities within GASHA. These can
be grouped into two overarching themes: decision confidence and
decision timeliness.
Physician/Clinical Director: The length of stay of patients is not
being managed as effectively as possible. The quality of patient
care can suffer as a result (surgeries postponed) and overall costs
are higher than they need to be. At present, the issue is difficult
to address due to lack of timely access to current information.
The next step of our analysis was to reverse engineer the two
above impacts by further analyzing the data using co-agency to
identify the information, process and personnel root causes of the
adverse impacts. In the following sections each co-agency issue is
broken down into themes and sub-themes. We also identified the
workarounds that were caused by each of the issues.

4.1.1. Decision confidence 4.2. Information-related issues

The first impact is a lack of confidence in decisions that are made
within the district. A lack of access and poor information qual- Two predominant information related issues emerged from our
ity eroded the confidence of decisions made as illustrated by the analysis – poor information quality and ineffective access and dis-
following comment: semination of information.
Senior Administrator: [We lack] confidence in
decisions. . .they are sometimes subjective due to lack of
facts and data to support.. . .. 4.2.1. Information quality issues
Two information quality-related issues were identified: oper-
ational data quality and issues with the format of available
4.1.2. Decision timeliness
decision-support information. Information is data that has been
The second impact is that people cannot respond to situations
manipulated in a way to make it useful for a specific purpose, e.g.,
and make decisions in a timely manner. Unlike confidence issues,
for decision support. At GASHA, data is supplied from a number
in this circumstance people have the information they need, they
of operational systems including Meditech and SAP. In order to be
just do not have access to it when needed:
useful, data must be of sufficient quality for its intended purpose
Clinical Director: [We] cannot react to situations as quickly (Wang & Strong, 1996). Our findings indicate that in many cases,
as necessary – impacts timeliness of decisions, and the overall GASHA’s operational data is not of sufficient quality for decision-
department efficiency. support purposes. 42% of interviewees cited poor data quality as
a major barrier to effective decision making. The most significant
data quality issues were related to format, currency and accuracy.
Regarding the format of data, respondents indicated that a good
deal of required data was not stored electronically (it was in paper
form only), necessitating extensive manual input and assembly in
order to generate the information needed for decision support as
described by the following comments:
Clinical Manager: It takes a great deal of manual effort to get the
information needed. There are many different places that infor-
mation is stored and important information is in paper form
(e.g., staff schedule information).
Physician/Clinical Director: Much of a patient’s chart informa-
tion is on paper only. When needed for management purposes,
requires extensive manual effort to extract (by reviewing charts
Fig. 1. Map of results by section. manually).
24 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
Table 3
FTE cost of manual data integration.
Consumer or consumer/producer FTE Provider FTE Total FTE
4.2 2.7 6.8
In terms of data stored electronically in operational systems,
data quality issues were seen to be related to the currency and
accuracy of data, as illustrated by these quotations:
Heath Records Professional: Some data (e.g. discharge sum-
maries) is not current – there is a lag of several months.
Senior Administrator: timeliness issues mean that information
is often not available in ‘decision windows’.
Clinical Manager: Access to ‘clean’ accurate data is an issue.
For example, patients should be assigned triage scores when
entering the ER. This information is often not entered or entered
incorrectly, resulting in the appearance that patients have very
simple or minor health issues when this is not the case.
With regard to available decision-support information, GASHA
staff was forced to access operations directly from operational sys-
tems such as Meditech and SAP. Often, this data was viewed to be
stored in isolated siloes as well as being poorly formatted for deci-
sion support, particularly for clinical decision makers, as illustrated
by these quotations:
Clinical Manager: Information is ‘siloed’ in different places.
Information is not appropriately disseminated, not everyone
who needs the information is receiving it.
Senior Administrator: SAP is not a very user-friendly tool; the
information that can be accessed is not formatted in a way that
makes sense to staff outside of the finance and procurement.
4.2.2. Impacts of information quality issues
The implications of poor data quality are significant. During the
interviews we learned that in order to obtain value from available
data, a two-stage workaround was used. First, a number of reports
were run and then manually exported into Excel where the data
was manipulated into useful form. We also learned that this pro-
cess, while marginally effective in some cases, was labor intensive
and time consuming. To understand the impact of these manual
workarounds one of our interview questions asked:
On a monthly basis, what percentage of your time is spent on
manually compiling information (perhaps using tools such as
Excel) to support management decision-making by you or oth-
ers?
The results of this question are presented in Table 3. Based on
the percentages supplied, we calculated the ‘full time equivalent’
(FTE) cost of manual information compilation.1 FTE being 100% of
a person’s time or 37.5 h per week. Overall, almost 7 FTEs of effort
are currently expended by the 40 interviewees. This translates to
approximately 17% of total working time for the study participants.
4.2.3. Information access and dissemination
Access and dissemination refers to an individual’s ability to
locate, retrieve, manipulate and share information. 76% of inter-
viewees identified access to information as a major barrier to
effective decision making. Our interview respondents stated that
while a great deal of data was being collected by GASHA systems
it was often difficult to access for various issues. The most signif-
icant access issue was a lack of a defined dissemination strategy.
There are two basic strategies for information dissemination: pull
and push. A ‘pull’ strategy means that users request information
as they need it. A ‘push’ strategy, on the other hand, means that
information is proactively sent to users automatically. From our
interviews it was apparent that a defined dissemination strategy
did not exist, resulting in inconsistent information dissemination,
and more importantly, people not having necessary information
when needed, as illustrated by the following comments:
Clinical Director: . . . you must pull to get the information
instead of it being pushed to you.
Public Relations Manager: Information is not appropriately
disseminated . . .. there is no automated push of information
to the leadership team.
4.2.4. Impact of information access and dissemination issues
Information access issues, particularly a lack of a dissemination
strategy, resulted in significant workarounds. Some of the indi-
viduals we interviewed stated they would actually spend more of
their time on manual information extraction, but over time they
have learned that the information they need is excessively diffi-
cult to access and have essentially given up on acquiring it. We
asked study participants to provide a percentage breakdown of how
information is obtained for group decision processes and how infor-
mation should be provided ideally. We learned that almost 100% of
information must be requested (pulled) using the ad hoc process
described earlier. In fact, across all interviews, we identified only
one report (relating to daily bed utilization/patient census), that
was provided using a push strategy. Study participants indicated
that an ideal mix of push vs. pull information delivery would be a
situation where approximately 55% of information is pushed auto-
matically to the appropriate groups, while 45% would be pulled as
required.
4.3. Process and personnel issues
Effective decision making requires well defined processes for
both producing and consuming information. This was not the
case within GASHA; our analysis revealed significant process and
personnel issues both with the production and consumption of
information.
4.3.1. Process issues
We identified both production and consumption process issues
at GASHA. From a production perspective, the most significant
issue identified was the ad hoc nature of the process. Numerous
respondents stated that there were no standard procedures and
documentation for requesting information – all requests for infor-
mation were ad hoc in nature.
In terms of decision support information consumption, respon-
dents indicated that there were no standards in place in terms of
how the core processes within the health district should be empir-
ically measured and a lack of understanding of what information is
needed for this purpose, as illustrated by these comments:
Clinical Manager: [There is a] lack of tools and no provincial
standards – do not know how other districts do measure-
ments/collect statistics. Situation makes benchmarking very
difficult. [It is] difficult to justify resource allocation/workloads.
Senior Administrator: [There is a lack of] workload indicators,
pertaining to staffing information. We don’t currently have sys-
tems in all parts of the business to capture the information.
4.3.2. Impacts of process issues
Processes relating to the production and consumption of
decision-support information are quite immature and this lack
of maturity has significant implications. Regarding information
production, the lack of a standard process and documentation
 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
of requests and request fulfillment, the work done on a given
request is not effectively leveraged to respond to subsequent,
similar requests. Secondly, those individuals responding to the
request must spend a significant amount of time dialoging with
the requestor to ensure a sound understanding of the information
needed. Finally, the ad hoc nature of the process contributes to a
lengthy turnaround time meaning information is often not received
in a timely manner, i.e., within the window of time available to
make decisions.
Regarding the consumption of decision support information, the
lack of definition of the information needed to monitor and mea-
sure core processes is a significant issue. Management guru Peter
Drucker often stated that “if you can’t measure it, you can’t man-
age it”. This was true for GASHA as they did not formally measure
many core processes, making it extremely difficult to assess their
efficiency and effectiveness.
4.3.3. Personnel issues
Personnel issues were identified both associated with the pro-
duction and consumption of decision-support information. At the
time of the study no dedicated decision support roles existed within
GASHA, nor did any job descriptions contain explicit decision sup-
port responsibilities. Rather, this information was provided ‘off
the side of the desk’ by a number of individuals who have dis-
played aptitude for decision support related work. This situation
was described in many of our interviews.
Heath Records Professional: Lack of dedicated decision sup-
port resources [is an issue]: [We] do not have someone who is
designated to write reports; this task is spread throughout many
different people
Senior Administrator: [There is] no centralized decision sup-
port resource: There is not one person who acts as the decision
support person – several different people provide support “from
the side of their desk”
4.3.4. Impact of personnel issues
The lack of relevant skills and knowledge, particularly as it per-
tains to the management team, is a significant issue – it implies
that even if the district had access to an effective BI platform, they
would not be able to exploit it, given that a significant portion
of the management team lacked data analysis skills. Many of the
health district management personnel are clinicians who moved
into management roles. As such, many of these individuals have
not had exposure to data analysis strategies and techniques.
To illustrate, only 15% of our participants were familiar with
the use of Microsoft Excel pivot tables. The major impact of this is
there are numerous individuals that lack the analytical capabilities
to perform their required duties.
Clinical Manager: [There is a] Lack of skilled staff to
retrieve/compile required information. Example: current cler-
ical staff does not have Excel skills – had to hire someone to
create Excel spreadsheets.
Clinical Director: Lack of management information and a lack
of ability to use the information [is a barrier to effective decision
making]
5. Discussion
Our study asked three questions. The first question relates
to the adverse decision-related impacts attributable to a lack
of decision-support capabilities. The second examines the infor-
mation, process and personnel-oriented root causes of these
impacts. Finally, we wanted to identify the factors that can assist
25
healthcare organizations implement BI systems, in light of the
issues that we discovered. As such, this work focuses strictly on
foundational implementation success factors, the starting point to
the ultimate success of a BI system.
Regarding question one, our analysis revealed that a lack of
decision-support capabilities had significant negative implications
for the organization. In the absence of timely access to the right
information, managers lacked confidence in their decisions, and
decision making took longer than required. Further, given the dif-
ficulty and length of time it took to obtain information, some
managers indicated that, in certain situations, they had given up on
getting the facts and data needed to make decisions, forcing them
to rely solely on experience and intuition. These findings make it
clear that decision processes in the organization are significantly
compromised due in large part to a lack of information. Address-
ing this issue through the implementation of effective BI systems
should have significant and tangible positive consequences.
However in using co-agency to answer question two we iden-
tified number of informational, process and personnel issues that
contributed to sub-optimal decision making processes and neces-
sitated inefficient workarounds. We identified information-related
issues within GASHA both in terms of operational data qual-
ity and effective access to and dissemination of decision-support
information. Providing decision makers with convenient access to
high quality information is critical for BI system success (Yeoh &
Koronios, 2010). It is important to note, however, that information
quality is a consequence of successful BI implementation and not
an antecedent (Wixom & Watson, 2001). Therefore addressing data
quality issues that exist in operational systems is an important pre-
BI implementation success factor that must be addressed (Isik et al.,
2011).
Many of the process and personnel issues we identified are
consistent with those of other studies, For example, we found a
lack of skills and knowledge required for effective consumption
of decision-support information, echoing studies by Ramamurthy
et al. (2008) and Popovic et al. (2012), who highlighted the need
for organizations to develop the right IS skills and knowledge to
develop and effectively use BI capabilities. However our study
extended existing research by identifying that is not just IS skills
that are needed but it is also critically important that manage-
ment personnel possess data analysis skills and that processes
be in place to enable effective dissemination of information. Our
analysis identified that processes related to creating and dissem-
inating decision-support information were highly immature with
a sub-optimal, informal mix of information pushing and pulling
processes. This is clearly an issue to be addressed as mature BI-
related processes such as information dissemination are essential
for successful implementation of BI systems (Sen, Ramamurthy,
et al., 2012). We also identified a lack of maturity in GASHA’s core
operational processes from the perspective of how they are moni-
tored and managed. Study participants indicated that virtually no
decision-support information was available for this purpose, and
that this was a significant concern. This is not to say that existing
processes are necessarily inefficient or ineffective, but rather it is
difficult to assess and effectively manage them. A primary objec-
tive of BI systems is to support effective process management. In
order to do so, processes, and the information required to effec-
tively manage them, must be well-defined (Bucher, Gericke, & Sigg,
2009; Ferrand et al., 2010).
Our final question seeks to clearly define the factors, informed
by the co-agency issues we identified, that healthcare organizations
in the early stages of BI system implementation should consider.
To answer that question we develop a framework which clearly
identifies core processes and facilitates the definition and prior-
itization of decision-support information needs in the context of
these processes. These information needs would include, but not
26 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
Table 4
Framework for information needs identification to support core healthcare processes.
Core process examples Overall strategy/goals
Medical
Surgery services Information needs
Emergency room services Information needs
Inpatient care – medical Information needs
Lab services Information needs
Diagnostic imaging services Information needs
Pharmacy services Information needs
Business
Facilities management Information needs
Inventory management Information needs
Support
Communications Information needs
Finance and accounting Information needs
Human resources Information needs
be limited to, specific measures of performance, referred to as met-
rics and performance indicators. Table 4 presents our proposed
process framework for information needs identification to support
core healthcare processes. Mettler and Vimarlund (2009) identified
three distinct classes of health system processes: medical, business
and support. Medical processes are the most mission-critical as
they have a direct impact on the delivery of patient care and include
diagnostic, inpatient care, surgery, and emergency services, among
others. We view business processes as those that directly support
the delivery of medical services (e.g., medical supply inventory
management). Support processes indirectly contribute to patient
care and include financial, accounting and human resource man-
agement. We identified core processes for each category from our
analysis of the GASHA organization and included them in Table 4
in the left vertical column.
Barone, Yu, Won, Jiang, and Mylopoulos (2010) suggest that
effective process management requires multiple perspectives
including strategy and goals, resource management, and operations
management. The strategic perspective involves defining over-
all objectives for the process. The goal of resource management
is the optimal allocation of an organization’s human and capital
resources. Operations management is concerned with measuring
process effectiveness and efficiency. We suggest the need for a
fourth perspective: quality and risk management, as it is funda-
mentally important to the delivery of patient care (Effken et al.,
2005; Travaglia, Westbrook, & Braithwaite, 2009). As illustrated in
Table 4, each management perspective for each core process would
have distinct information needs.
The framework we developed is specific to healthcare BI and
highlights that information requirements must be identified and
prioritized as a precursor to implementing BI. While the framework
we developed is inspired by the study in this paper, the co-agency
concepts (i.e. process examples and information needs mapping)
are general to allow it to be used in different healthcare contexts.
When considering the implementation of a BI system, we pro-
pose that a healthcare organization should begin by utilizing our
framework to identify and prioritize their core medical, business,
and support processes. Efforts can then shift to clearly defin-
ing strategic, resource management, operations management, and
quality and risk management objectives for each process (starting
with those of highest priority) and specifying the decision-support
information needed to manage the processes from these four per-
spectives. As previously discussed, addressing operational data
quality issues is a key implementation success factor. A significant
benefit of our proposed framework is that it allows organizations
to take a phased approach to dealing with these issues by address-
ing them on a process by process basis and focusing on the specific
data required to meet decision-support information needs.
Resource management Operations management Quality and risk management
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
Information needs Information needs Information needs
To illustrate how our framework might be used, consider the
example of emergency room services. From a strategic perspective,
a healthcare organization might establish overall objectives related
to patient satisfaction and cost per patient visit and establish a set of
metrics to monitor and measure progress towards these objectives.
In terms of resource management, the organization might measure
staff, facilities and equipment utilization vs. budgets and examine
trends in these areas over time. Operational measures would be
established to assess both efficiency and effectiveness of various
aspects of ER operations including lab test turnaround times, wait
times for inpatient beds, wait times to see a specialist, etc. Quality
and risk metrics would be established to measure infection rates,
repeat visits, etc.
6. Conclusion
This paper addressed a key research need by identifying infor-
mation, personnel, and process issues that must be addressed
prior to implementing BI systems in healthcare. We also provided
a co-agency inspired healthcare-specific process framework for
identifying and prioritizing decision-support information require-
ments. Our framework offers a practical means for health systems
to begin their journey towards BI implementation by crystalizing
what their most important processes are and defining the informa-
tion needed to support these processes.
A limitation of our study is that it is a case study of a single
organization. However our current work offers the foundation for
future research. For example, we are currently building on the work
from this paper by using the process framework to identify how
BI can be deployed in conjunction with electronic health records
(EHR) systems.
References
Agarwal, R., & Prasad, J. (1998). A conceptual and operational definition of personal
innovativeness in the domain of information technology. Information Systems
Research, 9(2), 204–215.
Ash, J., Berg, M., & Colera, E. (2004). Some unintended consequences of information
technology in health care: The nature of patient care information system-related
errors. Journal of the American Medical Informatics Association, 11(2), 104–112.
Barone, D., Yu, E., Won, J., Jiang, L., & Mylopoulos, J. (2010). The practice of enterprise
modeling. Lecture Notes in Business Information Processing, 68, 31–45.
Bostrom, R., & Heinen, J. S. (1977). MIS problems and failures: A socio-technical
perspective. MIS Quarterly, 1(3), 17–32.
Bradford, M., & Florin, J. (2003). Examining the role of innovation diffusion fac-
tors on the implementation success of enterprise resource planning systems.
International Journal of Accounting Information Systems, 4(3), 205–225.
Bucher, T., Gericke, A., & Sigg, S. (2009). Process-centric business intelligence. Busi-
ness Process Management Journal, 15(3), 408–429.
Crist-Grundman, D., & Mulrooney, G. (2011). Effective workforce management
starts with leveraging technology, while staffing optimization requires true
col*laboration. Nursing Economics, 29(4), 195–200.
 N. Foshay, C. Kuziemsky / International Journal of Information Management 34 (2014) 20–27
Effken, J., Brewer, B., Patil, A., Lamb, G., Verran, J., & Carley, K. (2005). Using OrgAhead,
a computational modeling program, to improve patient care unit safety and
quality outcomes. International Journal of Medical Informatics, 74(7–8), 605–613.
Ferrand, D., Amyot, D., & Corrales, C. V. (2010). Towards a business intelligence
framework for healthcare safety. Journal of Internet Banking and Commerce, 15(3),
1–9.
Flower, J. (2006). Who owns health care’s most valuable information? Physician
Executive, 32(5), 54–55.
GASHA Annual report, 2011, Retrieved from: http://www.gasha.nshealth.ca/
admin/sources/editor/assets/annualreport20102011.pdf
Greenhalgh, T., Robert, G., Macfarlane, F., Bate, P., & Kyriakidou, O. (2004). Diffusion of
innovations in service organizations: Systematic review and recommendations.
Milbank Quarterly, 82(4), 581–629.
Hanson, R. (2011). Good health information – An asset not a burden!. Australian
Health Review, 35(1), 9–13.
Harrison, M. I., Koppel, R., & Bar-Lev, S. (2007). Unintended consequences of informa-
tion technologies in health care: An interactive sociotechnical analysis. Journal
of the American Medical Informatics Association, 14(5), 542–549.
Hartwick, J., & Barki, H. (1994). Explaining the role of user participation in informa-
tion system use. Management Science, 40(4), 440–465.
Hseih, H. F., & Shannon, S. E. (2005). Three approaches to qualitative content analysis.
Qualitative Health Research, 15(9), 1277–1288.
Isik, O., Jones, M. C., & Sidorova, A. (2011). Business intelligence (BI) success and the
role of BI capabilities. Intelligent Systems in Accounting, Finance and Management,
18(4), 161–176.
Kaplan, B., & Harris-Salamone, K. D. (2009). Health IT success and failure: Recom-
mendations from literature and an AMIA workshop. Journal of the American
Medical Informatics Association, 16(3), 291–299.
Mattia, A. (2011). A multi-dimensional view of socio-technical information systems
research and technochange. The Review of Business Information Systems, 15(4),
11–18.
McKibbon, K. A., Lokker, C., Handler, S. M., Dolovich, L. R., Holbrook, A. M., O’Reilly,
D., et al. (2012). The effectiveness of integrated health information technologies
across the phases of medication management: A systematic review of random-
ized controlled trials. Journal of the American Medical Informatics Association,
19(1), 22–30.
Mettler, T., & Vimarlund, V. (2009). Understanding business intelligence in the con-
text of healthcare. Health Informatics Journal, 15(3), 254–264.
Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis (2nd ed.). Thousand
Oaks, CA: Sage.
Moore, G. C., & Benbasat, I. (1991). Development of an instrument to measure the
perceptions of adopting an information technology innovation. Information Sys-
tems Research, 2(3), 192–222.
Negash, S. (2004). Business intelligence. Communications of the Association for Infor-
mation Systems, 13, 177–195.
Pine, M., Sonneborn, M., Schindler, J., Stanek, M., Maeda, J., Hanlon, C., et al. (2012).
Harnessing the power of enhanced data for healthcare quality improvement:
Lessons from a Minnesota hospital association pilot Project/PRACTITIONER
APPLICATION. Journal of Healthcare Management, 57(6), 406–418.
Popovic, A., Hackney, R., Coelho, S., & Jaklic, J. (2012). Towards business intelligence
systems success: Effects of maturity and culture on analytical decision making.
Decision Support Systems, 54(1), 729–739.
Ramamurthy, K., Sen, A., & Sinha, A. P. (2008). An empirical investigation of the
key determinants of data warehouse adoption. Decision Support Systems, 44(4),
817–841.
27
Rogers, E. M. (1995). Diffusion of innovations (4th ed.). New York: Free Press.
Rohloff, R. (2011). Health-care BI: A tool for meaningful analysis. Healthcare Financial
Management, 65(5), 100–108.
Seah, M., Hsieh, M. H., & Weng, P. (2010). A case analysis of Savecom: The role of
indigenous leadership in implementing a business intelligence system. Interna-
tional Journal of Information Management, 30(4), 368–373.
Sen, A., Banerjee, A., Sinha, A. P., & Bansal, M. (2012). Clinical decision support:
Converging toward an integrated architecture. Journal of Biomedical Informatics,
45(5), 1009–1017.
Sen, A., Ramamurthy, K., & Sinha, A. P. (2012). A model of data warehousing process
maturity. IEEE Transactions on Software Engineering, 38(2), 336–353.
Tariq, A., Georgiou, A., & Westbrook, J. (2013). Understanding the information
dynamics of medication administration in residential aged care facilities
(RACFs): A prerequisite for design of effective ICT systems. Studies in Health
Technology and Informatics, 194, 188–194.
Thraen, I., Byron, B., Mullin, S., & Weir, C. (2012). Characterizing “information trans-
fer” by using a Joing Cognitive Systems model to improve continuity of care in
the aged. International Journal of Medical Informatics, 81, 435–441.
Travaglia, J., Westbrook, M., & Braithwaite, J. (2009). Implementation of a patient
safety incident management system as viewed by doctors, nurses and allied
health professionals. Health, 13(3), 277–296.
Toussaint, P. J., & Coiera, E. (2005). Supporting communication in health care. Inter-
national Journal of Medical Informatics, 74(10), 779–781.
Tremblay, M., Hevner, A., & Berndt, D. (2012). Design of an information volatil-
ity measure for health care decision making. Decision Support Systems, 52(2),
331–341.
Wang, R., & Strong, D. M. (1996). Beyond accuracy: What data quality means to data
consumers. Journal of Management Information Systems, 12(4), 5–33.
Wang, X. S., Nayda, L., & Dettinger, R. (2007). Infrastructure for a clinical-decision-
intelligence system. IBM Systems Journal, 46(1), 151–169.
Wixom, B. H., & Watson, H. J. (2001). An empirical investigation of the factors affect-
ing data warehousing success. MIS Quarterly, 25(1), 17–41.
Yeoh, W., & Koronios, A. (2010). Critical success factors for business intelligence
systems. Journal of Computer Information Systems, 50(3), 23–32.
Neil Foshay is a tenured faculty member in the Department of Information Systems
in the Schwartz School of Business at St. Francis Xavier University in Nova Scotia
Canada. His research interests are in the areas of business intelligence (BI), metadata
management and user adoption of technology. Dr. Foshay completed his PhD at the
Bradford University School of Management in 2008 and received an MBA from the
University of British Columbia in 1990. He has over 25 years of IT consulting experi-
ence including 17 years focused on business intelligence and has provided strategic
BI consulting services to organizations such as Apple, IBM, Honda, Caterpillar and
Walt Disney World.
Craig Kuziemsky, PhD, is an Associate Professor and director of the Master of Science
in Health Systems Program in the Telfer School of Management at the University
of Ottawa. He joined the University of Ottawa in 2007. He completed his PhD in
Health Information Science from the University of Victoria in 2006. He also received
Bachelor of Commerce and Science Degrees from the University of Alberta in 1993
and 2000 respectively. Craig’s research focuses on developing new approaches for
modeling collaborative healthcare delivery so we can better design information and
communication technology to support collaboration. He is also researching health-
care interoperability with a focus on people and process interoperability.