Journal of Knowledge Management

Business analytics-enabled decision making effectiveness through knowledge absorptive capacity in

health care
Yichuan Wang Terry Anthony Byrd
Article information:
To cite this document:
Yichuan Wang Terry Anthony Byrd , (2017)," Business analytics-enabled decision making effectiveness through knowledge
absorptive capacity in health care ", Journal of Knowledge Management, Vol. 21 Iss 3 pp. -
Permanent link to this document:
http://dx.doi.org/10.1108/JKM-08-2015-0301
Downloaded on: 24 March 2017, At: 01:32 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 10 times since 2017*

Access to this document was granted through an Emerald subscription provided by emerald-srm:543096 []
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

 Business analytics-enabled decision making effectiveness through knowledge absorptive
capacity in health care

1. Introduction
A large amount of clinical data being produced by various healthcare systems (e.g.,
electronic health records system, pharmaceutical R&D reporting system, and insurance
claims/billing system) galvanizes healthcare organizations toward making considerable
investments in business analytics (BA) to strengthen their data governance and analysis
capabilities (Ghosh and Scott, 2011; Ferranti et al., 2010; Raghupathi and Raghupathi, 2014).
BA is consist of a number of different analytics techniques such as descriptive, predictive, and
prescriptive analytics (Delen, 2014) and visualization techniques that use to extract meaningful
information from “big data.” BA also holds the potential to help transform the healthcare
system (Chen et al., 2012; Watson, 2014) and to address the cost and service quality pressures,
including the reduction of readmission rate (Bardhan et al., 2015) and management cost (Bates
et al., 2014), and the improvement of patient safety (Ferranti et al., 2010).
However, despite leveraging BA to derive clinical performance is emerging as a top
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

priority for healthcare organizations, only 16% of healthcare organizations have substantial
experience using analytics across a broad range of functions (Cortada et al., 2012). Many
healthcare organizations are suffering from the lack understanding of how to transform clinical
data into insights, knowledge, and informed decisions (Raghupathi and Raghupathi, 2014;
Ward et al., 2014). An increasing number of studies have demonstrated the potentials of BA for
providing tailored, context-sensitive information to guide clinical practice (e.g., Bardhan et al.,
2015; Halamka, 2014; Spruit et al., 2014), and explored the impact of BA in terms of its
potential for harvesting data-driven insights, supporting evidence-based medicine, and
improving the quality of care at a lower cost (e.g., Bates et al., 2014; Foshay and Kuziemsky,
2014; Srinivasan and Arunasalam, 2013). Yet, these studies are mostly technology oriented
(Wamba et al., 2015) and their findings have not been collected in a comprehensive framework,
or validated on a broader empirical basis (Wang et al., 2016). Without discovering the
prominent paths driving BA to gain value, it is harder to convince healthcare practitioners to
adopt BA-related technologies for healthcare transformation, and may prevent them from fully
embracing BA (Murdoch and Detsky, 2013; Shah and Patak, 2014).
To address these gaps, this study aims to answer the following two research questions
recognized by the recent work in big data. First, what essential capabilities can be created from
the BA applications that healthcare organizations should acquire to succeed in driving sound
decisions (Ghosh and Scott, 2011; Phillips-Wren et al., 2015). Second, what organizational
capabilities enable healthcare organizations to effectively deliver knowledge, triggered by the
use of BA systems, to decision makers and other stakeholders (Ghosh and Scott, 2011; Sharma
et al., 2014; Phillips-Wren et al., 2015). More specifically, drawing on the resource-based
theory (RBT) and dynamic capabilities view, we first conceptualize BA capabilities in health
care and then consider the role of knowledge absorptive capacity in the relationship between
BA capabilities and decision making effectiveness that it plays an intermediary role in
transforming knowledge obtained from the BA applications into a useful decision making
resource.
The next section of this paper reviews the current research specifically focusing on
exploring the business value of BA. We then move on to examine the theoretical background
for BA capability, absorptive capacity, and decision making effectiveness to create a basis for
developing a research model with a series of hypotheses about the relationships between these
proposed constructs. After describing our research methodology and presenting the results, we
conclude by discussing our findings and their implications.

1
 2. Literature review on business analytics value creation
To unveil the role of BA in creating business value, lately there have been a number of
studies focused on developing BA enabled business value models that are generally grounded
on information processing view (IPV) and resource-based theory (RBT).
From an information processing view (IPV), several studies argue that BA can help
organizations process huge amounts of data to acquire meaningful insights that they can then
transform into organizational knowledge and actionable decisions (Cao et al., 2015;
Kowalczyk and Buxmann, 2014; Trkman et al., 2010). To facilitate decision-making quality,
organizations should design their organizational structure, mechanism, and business
processes in conjunction with data analysis processes that may reduce the environmental
uncertainty and ambiguity of the problem context (Kowalczyk and Buxmann, 2014; Sharma
et al., 2014). For supply chains, for example, Trkman et al. (2010) report that firms that have
the ability to analyze and utilize their information within the different stages of the supply
chain (i.e., plan, source, make, and deliver) enjoy a superior supply chain performance as a
result. In the same vein, Cao and colleagues (2015) have found that utilizing BA influences
information processing capability through the mediation of a data-driven environment, which
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

in turn has a positive effect on decision making effectiveness. These studies focus on
exploiting the use of information to improve decision making processes and outcomes. This
allows us to understand how business decisions are made through the joint effect of business
analytics and information processing mechanisms.
Grounded in the theoretical lens of the RBT and IT capability literature, few studies
explain how a firm’s unique business analytics capability can be constructed by the
configurations of available business analytics technological and organizational resources (e.g.,
Işık et al., 2013; Kwon et al., 2014; LaValle et al., 2011; Popovič et al., 2012; Wixom et al.,
2013). For example, Wixom et al. (2013) identify two key business analytics capabilities –
speed to insight and pervasive use – and their underlying dimension from BA resources as
playing a role in maximizing business value in the fashion retail industry. Popovič et al. (2012)
argue that mature business intelligence (BI) system with strong analytical capabilities and data
integration, along with knowledge workers who are capable of making full use of complex
business intelligence systems, can provide sufficient information to markedly improve
decision making processes. Işık et al. (2013) further demonstrate that technological capabilities
such as data quality, user access and the ability to integrate BI with other systems, and BI
flexibility are necessary for creating business value.
In the context of healthcare, Ghosh and Scott (2011) describe how analytical capabilities
facilitate data-driven decision making. Their case study shows that Veterans Health
Administration’s (VHA) BA systems support the physicians’ day-to-day clinical practices,
such as assessing the riskiness of a certain surgical procedure by providing the outputs
displayed in the dashboards and metrics. BA systems also allow aggregating patient data to
establish measurable improvements that help healthcare managers allocate resources (e.g.,
determine the resource utilization for the facility and geographic distribution of patients
support service needed) and choose future treatments and policies (e.g., assess the outcomes of
policy initiatives and develop medical protocols).

3. Research model and hypotheses development

Drawing on the RBT (Barney, 1991; 2001) and IT capability literature (Bharadwaj, 2000),
we first conceptualize BA capability by arguing that BA resources – that is, its BA architectural
components (i.e., data aggregation, data analysis, and data interpretation) can create
BA-specific IT capabilities. The logic behind this argument is that IT capability literature
generally adopts a RBT (Barney, 1991; 2001) to argue that a firm’s unique IT capability can be
constructed by the configurations of its available tangible and intangible IT resources or the

2
 synergetic combination of its non-valuable, rare, imperfectly imitable and non-substitutable
(VRIN) resources (Santhanam and Hartono, 2003). For example, Pavlou and El Sawy (2006)
propose three key dimensions of IT capability that can be identified from new product
development (NPD) systems: effective use of project and resource management systems,
effective use of knowledge management systems, and effective use of cooperative work
systems. With this logic, BA capability could be viewed as a specific type of IT capability.
While we are mindful of developing BA capabilities, we are also concerned for the role of
dynamic capability in driving business value from BA. Dynamic capability is a firm’s
organizational ability to sense and shape opportunities and threats, to seize market
opportunities and to maintain competitiveness (Barreto, 2010; Teece, 2007). Dynamic
capability view explains how organizations integrate, reconfigure, gain and renew IT resources
to match rapidly-changing market environments (Eisenhardt and Martin, 2000; Helfat and
Peteraf, 2003; Pavlou and El Sawy, 2006; Teece et al., 1997). We follow this view and the
arguments from prior BA literature (Cao et al., 2015; Popovič et al., 2012; Trkman et al., 2010)
to consider a mediating role of dynamic capability between BA related constructs and
organizational performance since IT alone do not unequivocally facilitate organizational
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

performance (El Sawy et al., 2010).

Organizational scholars have viewed absorptive capacity as a specific type of dynamic
capability (Liu et al., 2013). From an organizational learning perspective, Lichtenthaler (2009)
defines it as the ability to assimilate and transform valuable IS knowledge, or to combine new
knowledge with existing knowledge by communicating with other organizational members.
Fink et al. (2016) and Sharma et al. (2014) argue that this capability plays an important role in
transforming the insights triggered by BA into business value. As noted by Ross et al. (2013),
BA per se may not magically create benefits until an organization learns how to turn insights or
knowledge discovered from data analytics into competitive advantage. Cultivating the ability
to identify, extract, transform, and utilize knowledge is essential for healthcare since healthcare
decision making is a complex process and heavily dependent on access to knowledge.
Schneeweiss (2014) has agreed on this view, suggesting that healthcare organizations must
develop the ability to absorb medical knowledge discovered by BA systems and obtain new
insights, and apply them to clinical practices in order to create a high quality evidence-based
medicine. We thus focus on examining the role of absorptive capacity in the relationship
between BA capabilities and decision making effectiveness.
Overall, this study develops a research model (see Figure 1) to represent the mechanisms
by which BA capabilities (i.e., the effective use of data warehouse tools, analytics tools, and
data visualization tools) in healthcare units can be shown to indirectly influence decision
making effectiveness through a key mediating link: absorptive capacity. The following
sections will discuss the constructs being used in our research model and associated hypotheses
guiding this research.

-------------------
Insert Figure 1 here
-------------------
3.1 Decision making effectiveness
Strategic management scholars define decision making effectiveness as the extent to
which a decision achieves the objectives established by management at the time it is made
(Dean and Sharfman, 1996, p. 372). In the IS literature, decision making effectiveness is an
important indicator of IS success and is generally viewed as the dependent variable in IS
adoption model (DeLone and McLean, 1992). IS scholars (e.g., Meador et al., 1984; Sanders
and Courtney, 1985) have used decision making effectiveness to measure the performance of
an information systems.

3
 In the BA literature, decision making effectiveness can be achieved by boosting the speed
of a decision (Wixom et al., 2013) and the extent to which organizations understand their
customers (Cao et al., 2015; LaValle et al., 2011). These two outcomes have been emphasized
in the context of analytics-based healthcare systems and individually linked to improved
quality of patient care (Barjis et al., 2013; Foshay and Kuziemsky, 2014). Therefore, this study
chose enhanced decision making effectiveness as signifying BA success in the healthcare
context. The following sections describe the roles of BA capabilities and absorptive capacity,
which are proposed to influence decision marking effectiveness.

3.2 Business Analytics Capabilities

Following Pavlou and El Sawy’s (2006) reasoning, we identify the key dimensions of BA
capability from the tools and functionalities of BA systems. To this end, we reviewed the
relevant academic literature (e.g., Raghupathi and Raghupathi, 2014; Ward et al., 2014),
technology tutorials (Hu et al., 2014; Watson, 2014), and case descriptions regarding applying
BA systems in healthcare settings. Our starting point was Ward et al.’s (2014) proposed BA
architectural framework for health care that elucidates how decisions are made in terms of four
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

architectural layers that begin with data generation and continue through data extraction and
data analysis to visualization and reporting, listing the tools and functionalities that are used in
each architectural layer. With these dimensions in mind, over 60 big data implementation cases
from diverse resources such as major IT vendors, academic journal databases, and healthcare
institute reports were reviewed to include, integrate, or drop the items. This review generally
affirmed Ward et al.’s framework, apart from the need to integrate data generation and data
extraction under a single dimension – data aggregation – because BA systems typically use
data warehousing tools to capture, aggregate and ready data from various sources for
processing (Raghupathi and Raghupathi, 2014). Based on the results of this review, we propose
three key dimensions of BA capability in healthcare: (1) the effective use of data aggregation
tools, (2) the effective use of data analysis tools, and (3) the effective use of data interpretation
tools, as described below in more detail and summarized in Table 1.

-------------------
Insert Table 1 here
-------------------

3.2.1 Effective use of data aggregation tools

Data aggregation tools are capable of transforming different types of healthcare data (e.g.,
electronic health records; EHRs, diagnostic or monitoring instrument data, web and social
media data, insurance claims/transaction data, pharmacy data, patient-generated data) into a
data format that can be read by the data analysis platform. As Raghupathi and Raghupathi
(2014) stated, data is intelligently aggregated by three key functionalities in data aggregation
tools: acquisition, transformation, and storage.
The primary goal of data acquisition is to collect data from external sources and all the
various system components across the healthcare organization. During the data transformation
process, transformation engines move, clean, split, translate, merge, sort, and validate the data,
as needed. Structured data such as that typically contained in an eclectic medical record is
extracted from EHR systems and converted into a specific standard data format, sorted by
specified criterion (e.g., patient name, location, or medical history), and then the record
validated against data quality rules. Finally, the data are loaded into target databases such as
Hadoop distributed file systems (HDFS) or stored in a Hadoop cloud for further processing and
analysis. The data storage principles are established based on compliance regulations, data
policies and access controls, and data storage methods can be implemented and completed in

4
 batch processes or in real time. Since these three functionalities support health care service in
value-adding ways, the effective use of data aggregation tools is viewed as a key element of
BA capability in health care.

3.2.2 Effective use of data analysis tools

Data analysis tools process all kinds of data and perform appropriate analyses to harvest
insights (Wald et al., 2014). This is particularly important for transforming patient data into
meaningful information that supports evidence-based decision making and useful practices for
healthcare organizations. The simple taxonomy of analytics developed by Delen (2014) lists
three main kinds of analytics, descriptive, predictive, and prescriptive, each of which is
distinguished by the type of data and the purpose of the analysis.
Descriptive analytics has been widely used in both business intelligence systems and BA
systems (Watson, 2014). In hospital settings, descriptive analytics is useful because it allows
healthcare practitioners to understand past patient behaviors and how these behaviors might
affect outcomes from their EHR database. It also provides high-speed parallel processing,
scalability, and optimization features geared toward BA, and offers a private and secure
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

environment for confidential patient records (Wang et al., 2014; 2015).

Predictive analytics allows users to predict or forecast the future for a specific variable
based on probability estimation (Phillips-Wren et al., 2015; Watson, 2014).
Hadoop/MapReduce, one of the most commonly used predictive analytics-based software
products, integrates analytical approaches such as natural language processing (NLP), text
mining, and natural networks in a massively parallel processing (MPP) environment. In
general, predictive analytics enables users to develop predictive models in a flexible and
interactive manner to identity causalities, patterns and hidden relationships between the target
variables for future predictions. Applying this to a healthcare context, predictive analytics
helps managers disentangle the complex structure of clinical cost, identify best clinical
practices, and gain a broader understanding of future healthcare trends based on knowledge of
patients’ lifestyles, habits, disease management and surveillance (Groves et al., 2013).
Predictive analytics also can be used to analyze social media data. Prior research has indicated
that this analysis could benefit a healthcare organization in various ways, including helping
track and even predict the course of illness through a population, providing non-official
channels for disease reporting, and facilitating conversations and interactions with patients
(Ward et al., 2014).
Prescriptive analytics is a relatively new kind of analytics that uses a combination of
optimization-, simulation-, and heuristics-based predictive modeling techniques such as
business rules, algorithms, machine learning and computational modeling procedures (Delen,
2014). Whereas predictive analytics suggests “what will occur in the future” (Watson, 2014, p.
1251), prescriptive analytics offers optimal solutions or possible courses of action to help users
decide what to do in the future (Phillips-Wren et al., 2015; Watson, 2014). Prescriptive
analytics continually re-predicts and automatically improves prediction accuracy by importing
and incorporating new datasets (a combination of structured and unstructured data and
business rules) to aid decision makers in solving problems (Riabacke et al., 2012).
Combining these functionalities of data analysis can help increase the efficiency of health
care delivery, and we thus proposed the effective use of data analysis tools as a key dimension
of BA capability.

3.2.3 Effective use of data interpretation tools

Data interpretation tools can be used to produce reports about daily healthcare services to
aid managers’ decisions and actions. Three key functionalities are involved. The first
functionality yields general clinical summaries such as historical reporting, statistical analyses,

5
 and time series comparisons and can be utilized to provide a comprehensive view that supports
the implementation of evidence-based medicine (Ghosh and Scott, 2011), provides advanced
warnings for disease surveillance (Jardine et al., 2014), and guides diagnostic and treatment
decisions (Fihn et al., 2014).
Second, data visualization, which is a critical BA feature, facilitates the extraction of
meaning from external data by creating helpful visualizations of the information, generally in
the form of interactive dashboards and charts. In healthcare, these visualization reports support
physicians and nurses’ daily operations and help them to make faster and more rational
evidence-based decisions (Roski et al., 2014). For example, a Dutch long-term care institution
has visualized the number of incidents, the locations where the incidents occurred, and the type
of physical damage that resulted by mining a collection of 5,692 incidents over a certain time
period (Spruit et al., 2014). Displaying frequency tables in the form of visual dashboards has
enabled this Dutch long-term care institution to improve patient safety throughout the hospital.
Third, real-time reporting, such as alerts and proactive notifications, real time data navigation,
and operational key performance indicators (KPIs) can be sent to interested users or made
available in the form of dashboards in real time. Since these three functionalities support
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

clinical decision making, the effective use of data interpretation tools is viewed as a key
element of BA capability.

3.3 Absorptive capacity

Absorptive capacity was conceptualized by Cohen and Levinthal (1990) to describe how a
firm absorbs relevant knowledge. Absorptive capacity is believed to be beneficial for firms
since it allows them to identify the value of new information gathered from internal and
external source, absorb it, and apply it to support their business decisions (Cohen and Levinthal,
1990; Saraf et al., 2013). In IT business value literature, researchers suggest that absorptive
capacity acts as a key driver of transforming IT into business value because organizations have
to make a considerable effort to acquire and internalize new knowledge from IT (Joshi et al.,
2010; Malhotra et al., 2005). With absorptive capacities, a firm can proactively make proper
and fast decisions on business strategies than their competitors (Elbashir et al., 2011;
Francalanci and Morabito, 2008). In the context of new product development, for example,
firms can make timely decisions related to product development and more effectively
commercialize innovative ideas into new products if they can create new knowledge more
efficiently than other competitors (Lin et al., 2015).
Although the dimensions of absorptive capacity are well-defined, and the importance of
absorptive capacity is established in various contexts, it has not been significantly applied to
healthcare sectors (Kash et al., 2013). Cohen and Levinthal (1990) originally identify three
dimensions of absorptive capacity: identification, assimilation and exploitation. This was later
expended to four dimensions: acquisition, assimilation, transformation, and exploitation of
knowledge (Flatten et al., 2011; Zahra and George, 2002). We follow this extended
conceptualization provided by Flatten et al. (2011) and consider these four capacities together
to represent the absorptive capacity of the healthcare organization. Acquisition reflects the
process of identifying valuable clinical knowledge from internal resources, such as diagnostic
or monitoring instrument data and patient behavioral data and external resources, such as
insurance claims/transaction data, pharmacy and lab data. Assimilation means the process of
understanding or interpreting the meaning of the clinical knowledge, while transformation
refers to the integration of new knowledge with current knowledge, thus preparing the
knowledge for application (Zahra and George, 2002). Finally, exploitation illustrates the
process of using the integrated knowledge to improve the healthcare organization’s existing
performance and generate new value (Wu and Hu, 2012). Together, these four capacities

6
 reflect healthcare organizations ability to highlight and apply new clinical knowledge, which is
critical to clinical performance.
As importance of knowledge management and organizational learning has been
acknowledged, prior research indicates that higher absorptive capacity can lead to better
healthcare quality improvement (Kash et al., 2013; Lev et al., 2009; Wu and Hu, 2012).
Evidence from 12 Israli hospitals, Lev et al. (2009) argue that absorptive capacity is the key to
integrate with existing knowledge and improve knowledge flow for its further sharing in order
to achieve better performance in a turbulent environment. Taking electronic medical record
(EMR) adoption as an example, with absorptive capacities, the members within a healthcare
organization can extend their ability from using EMR information properly to identifying how
to more fully use the EMR to attain additional financial, quality, or other strategic goals (Kash
et al., 2013). Wu and Hu (2012) have found the impact of absorptive capacity with four
underlying dimensions (i.e., knowledge acquisition, transfer, integration, and application) on
hospital process capabilities and financial and patient performance. Based on these arguments,
we believe that a high level of absorptive capacity enables healthcare organizations to
transform clinical data into insights that speed up the decision making process and enable
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

medical staffs to respond quickly to customer needs. Hence, the following hypothesis was
developed:

Hypothesis 1 (H1): Absorptive capacity will have a positive impact on decision making
effectiveness in health care.

3.4 The effect of BA capabilities on absorptive capacity

In health care, several studies have reported that BA capability offers several benefits,
including the ability to gather data from current patients to gain useful knowledge for decision
making (Ghosh and Scott, 2011), the ability to predict patient behavior via predictive analytics,
and the option to retain valuable customers by providing real-time offers (Bardhan et al., 2015;
Srinivasan and Arunasalam, 2013). Although acquiring and extracting knowledge from patient
data appears to be a challenge due to the need to preserve privacy and maintain trust in the
health infrastructure (Chen et al., 2012), several studies have explored ways through which BA
capabilities can help healthcare organizations improve their absorptive capacity (e.g., Bates et
al., 2014; Groves et al., 2013; Lin et al., 2011). First, the effective use of data aggregation tools
can track healthcare data from external sources and the system’s IT components throughout the
organization’s units. Healthcare-related data such as activity and cost data, clinical data,
pharmaceutical R&D data, patient behavior and sentiment data are commonly collected in real
time or near real time from payers, healthcare services, pharmaceutical companies, consumers
and stakeholders outside healthcare (Groves et al., 2013). Thus, knowledge related to patient’
needs is likely to be acquired when the ability to collect, store, and disseminate the data are
sufficient.
Second, since significant clinical knowledge and a deeper understanding of patient
disease patterns can be gathered from the analysis of EHRs (Lin et al., 2011), data analysis has
become an important tool to identify patterns of care and discover associations from massive
healthcare records, thus providing a broad overview for evidence-based clinical practice. In
hospital settings, the clinical analysis tools in large longitudinal healthcare databases can be
used to identify knowledge about drug risk, for example. By integrating BA algorithms into
their legacy IT systems, medical staffs can automatically acquire information relating to drug
safety decompensation, and treatment optimization by analyzing warning signals triggered by
alarm systems (Bates et al., 2014). In addition to clinical analyses, social media analytics allow
healthcare organizations to discover knowledge from online healthcare communities (Fan and
Gordon, 2014). Social media and its content generated by social interactions and

7
 communications among patients not only makes it possible to explore incredible business
values, but can also serve as a vital knowledge base for improving healthcare quality and
patient satisfaction.
Third, the effective use of data interpretation tools can yield sharable information and
knowledge in the form of historical reports, executive summaries, and drill-down queries in an
interoperable BA platform. BA has the potential to equip organizations with the reporting
systems they need to harness the mountains of heterogeneous data, information, and
knowledge that they routinely gather, disentangle intricate customer networks and develop a
new portfolio of business strategies for products and services. For example Premier, a
healthcare alliance of approximately 3,000 U.S. hospitals, collects data from different
departmental systems and sends it to a central data warehouse. After near-real-time data
processing, comprehensive and comparable clinical reports of resource utilization and
transaction level cost are generated and used to help hospital managers to recognize emerging
healthcare issues such as patient safety and inappropriate medication use.
Given the increasing embeddedness of BA tools in healthcare operational process, the
extent to which a healthcare organization can rapidly acquire, assimilate, and exploit
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

knowledge across its boundaries appears to be primarily dependent upon its ability to leverage
and implement BA tools, which is reflected in its BA capabilities. Hence, we developed the
following set of hypotheses:

Hypothesis 2 (H2): The effective use of data aggregation tools has a positive impact on
absorptive capacity in health care.

Hypothesis 3 (H3): The effective use of data analysis tools has a positive impact on absorptive
capacity in health care.

Hypothesis 4 (H4): The effective use of data interpretation tools has a positive impact on
absorptive capacity in health care.

3.5 The mediating role of absorptive capability

Absorptive capacity can be conceptualized as a higher-order organizational capability
(Liu et al., 2013; Roberts et al., 2012) that enable firms to identify, assimilate, and exploit
lower-order capabilities (e.g., IT capability and operational capability) to help organizations
acquire and sustain a competitive advantage (Cohen and Levinthal, 1990; Grewal and
Slotegraaf, 2007; Zahra and George, 2002). IT capabilities, on the other hand, can be viewed as
lower-order capabilities that triggering by higher-order capabilities (Pavlou and El Sawy, 2006;
2010). Pavlou and El Sawy (2006) and Roberts et al. (2012) agree, arguing that absorptive
capacity serves as a complement to IT capability in creating business value and emphasizing
that obtaining capabilities from the use of IT to increase organizational performance cannot be
guaranteed unless organizations have sufficient capacity to identify, absorb, transform, and
exploit the knowledge that is generated from IT. For instance, Pavlou and El Sawy (2006)
contend that the pivotal role of absorptive capability, triggered by the effective use of new
product development (NPD) systems, has become the source of competitive advantage in the
NPD context. Pavlou and El Sawy’s study extends RBT by considering the effect of absorptive
capability as a mediating factor linking the impact of NPD related systems with competitive
advantage.
Following this logic, few studies view BA capabilities as lower-order capabilities that
enabling the development of higher-order organizational capabilities, such as BA-enabled
organizational capabilities and dynamic capabilities (Knabke and Olbrich, 2015; Shanks and
Bekmanedova, 2012) and adaptive capabilities (Erevelles et al., 2016). In their longitudinal

8
 case study of a large financial institution, Shanks and Bekmanedova (2012) found evidence to
suggest that BA systems creates firm performance by orchestrating BA enabled organizational
capabilities and dynamic capabilities over time. Most recently, Erevelles et al. (2016)
integrates RBT with dynamic capability to develop a BA enabled competitive advantage model.
Their model not only argues that organizational BA resources allow firms to transform
marketing data into consumer insights, but also underscores the realization that dynamic and
adaptive capabilities will be triggered by these BA resources, thereby creating marketing
value.
Thus, conceptual arguments from prior literature suggest that absorptive capacity
mediates the relationship between a healthcare organization’s BA capability and
decision-marking effectiveness. High levels of BA capability could enable healthcare
organizations to support their decision making. Improved absorptive capacity provides an
opportunity for them to speed up their decision making processes, enhance the quality of
decision making, and deepen their understanding of their patients’ needs. In contrast, without it
they are less likely to achieve superior decision making effectiveness. We therefore propose
indirect impacts of BA in healthcare on decision making effectiveness through the mediating
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

role of absorptive capacity, expressed by the following hypotheses:

Hypothesis 5a (H5a): Absorptive capacity mediates the impact of effective use of data
aggregation tools on decision making effectiveness in health care.

Hypothesis 5b (H5b): Absorptive capacity mediates the impact of effective use of data analysis
tools on decision making effectiveness in health care.

Hypothesis 5c (H5c): Absorptive capacity mediates the impact of effective use of data
interpretation tools on decision making effectiveness in health care.

4. Research Methodology
4.1 Sampling frame and data collection
This study employed a survey method to collect primary data from Taiwan's healthcare
industry. The sample population consisted of Taiwan’s hospitals from the most recently
available list of hospitals published by the Joint Commission of Taiwan (JCT). The qualifying
hospitals should have experience of BA investment for the management and development of
healthcare services. We posited that larger hospitals would be more likely to perform BA
activities, so to be included in our study, a hospital had to be classified as either a medical
center, regional hospital or district hospital and have at least 100 in-patient beds. Local clinics
and psychiatric hospitals were excluded because they are generally too small to invest in BA.
In all, 424 hospitals satisfied all the above criteria and were included in the survey.
This study focuses on whether organizations’ decision making effectiveness can be
influenced by the use of BA systems. Thus, C-suite business executives, IT managers or senior
IT staffs who were actively involved in BA activities were the subjects in this survey. As Wu
and Hu (2012) noted, the implementation of knowledge management practice should be
supported by the role of IT. Senior IT employees such as IT managers and senior IT staffs in
Taiwan’s hospitals are often most responsible for knowledge management practice. Thus, they
are knowledgeable and should be considered to be important subjects in this survey. We mailed
one questionnaire to each hospital’s primary contact, with a follow-up reminder two weeks
later to non-respondents; in total, 424 questionnaires were sent to potential participants. Of the
155 responses received, three were incomplete, giving a 35.84% response rate with 152 valid
data points. Of these respondents, 26.97% (n=41) were from C-suite business executives,
including 17 CEOs and 24 CIOs, 47.37% (n=72) were IT managers and 25.66% (n=39) were

9
 senior IT staffs. With respect to hospital size, 76.32 % (n=116) of the participating hospitals
had at least 200 employees. We recognized the difficulty and importance of finding
respondents who can provide insights into various factors and so built in a selection filter by
asking the participants to self-check against their level of experience regarding BA before
taking the survey. The responses revealed that 78.94 % (n=120) of the participants had been
working on BA projects for at least five years, 12.50% of the participants (n=19) had been
working on BA projects for at least three years, and 8.56% of the participants (n=13) had at
least one year BA experience. Since the primary focus of the present study is at the
organizational level, the respondents’ abundant experience in this area should provide some
valuable insights.

4.2 Measurement
We developed a series of multi-item measures by either adopting scales that had been
previously validated from the existing literature and modifying them appropriately to fit our
research context or by developing new scales where there was no existing validated scale.
Since the target healthcare organizations are located in Taiwan, all the survey questions were
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

translated into Chinese by one of the authors. The translated copies of questionnaires
double-checked by two researchers who are familiar with both languages for more confident.
Following the survey translation procedure recommended by Brislin (1990), the final draft
translated to English again for comparing with original one. A few changes to the BA
capability scale were made in order to be consistent with the terminology. This process of back
translation ensures the face validity and accuracy of the items. Appendix 1 lists the
measurement items used. Responses to all the multi-item measures were captured using
seven-point Likert-type scales.
Decision making effectiveness: The measurement of this construct was based on reports in
the relevant literature, suitably adapted to the context of health care (Cao et al., 2015, LaValle
et al. 2011; Wixom et al., 2013). The speed with which a decision is reached is a key
component of decision making effectiveness expected from BA (Wixom et al., 2013), while
understanding customers refers to the extent to which organizations understand their customers
(Cao et al., 2015; LaValle et al., 2011). The quality of decision making was included based on
Sanders and Courtney’s (1985) suggestions. The resulting 3-item scale was used to capture
responses by asking about whether the decision making effectiveness can be satisfied with the
aid of BA, with responses ranging from 1 = completely dissatisfied through 7 = completely
satisfied.
Business analytics capabilities: As BA is still in its infancy in the IS field, there are no
validated measurement items for BA capability, so to develop and validate an instrument for
BA capability, we incorporated scale development procedures and recommendations from
Lewis, Templeton and Byrd (2005) and Mackenzie et al. (2011) as our guidelines. First, we
selected appropriate constructs and underlying items by reviewing academic research,
technical reports, and case studies. From a system functionality perspective, BA capabilities
are operationalized into three dimensions: the effective use of data aggregation tools, the
effective use of data analysis tools, and the effective use of data interpretation tools. These
initial items aim to assess the extent to which each BA tool is used effectively in healthcare
services. Next, content validity was verified and achieved through a pre-test. A small panel of
three CIOs who work for healthcare organizations, five MIS researchers, and seven doctoral
students in the MIS program were recruited as our content evaluation panel to review our
instrument in terms of format, content, understandability, terminology, and ease and speed of
completion. This panel was asked to act as judges by sorting items into groups and then
critiquing the proposed items. We also asked the judges to identify specific items that should be
added or deleted from the instrument, and to provide suggestions for improvement generally.

10
 Seven items were modified in accordance with their suggestions. A seven-point Likert-type
scale was used for all the BA capability dimensions to capture responses by asking “please rate
the effectiveness by which your organization uses the following BA tools in healthcare
services”, ranging from 1= poorly developed to 7 = well developed.
Absorptive capacity: The measurement of this construct was adopted from Pavlou and El
Sawy (2010), and modified to fit the context of health care. A 4-item scale was used to rate the
effectiveness by which an organization can acquire, assimilate, transform, and exploit
knowledge with the aid of BA. A seven-point Likert-type scale was again used to capture the
responses, ranging from 1= strongly disagree to 7 = strongly agree.

4.3 Non-response bias and common method bias

Non-response bias. This aspect was assessed by comparing the early (those who
responded to the first mailing) and late respondents (those who responded after the reminder),
in terms of the number of employees using t-tests. The results showed no statistically
significant difference between these two groups, indicating that non-response bias did not
present a problem for this study.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

Common method bias. To reduce common method bias, Podsakoff et al. (2003) suggest
the use of specific procedures during both the design and data collection processes. Following
these guidelines, we protected respondent-researcher anonymity, provided clear directions to
the best of our ability, and proximally separated independent and dependent variables
(Podsakoff et al., 2003). We then tested for bias statistically. First, Harman’s one factor test
(Brewer et al., 1970) was used to determine whether common method bias would pose a threat
to the validity of this study’s results. The results showed that five factors emerged with
eigenvalues greater than 1. Of these, the first component accounted for 31.41% of the total
variance and the unrotated factor solution indicated that no factor accounted for 50% or more
of the variance. Second, following a procedure suggested by Pavlou et al. (2007), we compared
correlations among the constructs. The results revealed no constructs with correlations over 0.7,
whereas evidence of common method bias ought to have shown considerably higher
correlations (r>.90). Consequently, these tests suggest that that common method bias is
unlikely to pose a significant threat to the validity of this study.

5. Analysis and results

Given our research model and objectives, structural equation modeling (SEM) was used
to conduct data analysis. Three reasons drove this choice. First, SEM can examine proposed
causal paths among constructs (Gefen et al., 2011). Second, the model does not include
second-order formative constructs. Each indicator was modeled in a reflective manner. Third,
our mediating variable, absorptive capacity was measured using multiple items which have to
model the measurement error. Thus, SEM is more appropriate than PLS. We analyzed the data
using IBM Amos 20.

5.1 Descriptive statistics and reliability and validity of scale

Table 2 presents the means, standard deviations, Cronbach’s alphas, average variance
extracted (AVE), Composite reliability, and construct correlations. The Cronbach’s alphas
(ranging from 0.80 to 0.91) indicate a satisfactory degree of internal consistency and reliability
for the measures (Bollen and Lennox, 1991), with all values well above .70 (Nunnally and
Bernstein, 1994). Construct reliability was assessed based on the composite construct
reliabilities (CR) (Hair et al., 2010, p. 687). As shown in Table 3, the CRs ranged from 0.93 and
0.98, well over the commonly accepted cutoff value of .70 (Hair et al., 2010), thus
demonstrating the adequate reliability of the measures.

11
 -------------------
Insert Table 2 here
-------------------

Discriminant validity was first assessed by examining the construct correlations.

Although there are no firm rules, inter-construct correlations below |.7| are generally
considered to provide evidence of measure distinctness, and thus discriminant validity. None
of the construct correlations were greater than .7, which demonstrates discriminant validity
(see Table 3). Another way to examine discriminant validity is to compare the AVE to the
squared inter-construct correlation. When the AVE is larger than the corresponding squared
inter-construct correlation estimates, this suggests that the indicators have more in common
with the construct they are associated with than they do with other constructs, which again
provides evidence of discriminant validity. The data shown in Table 3 suggests the adequate
divergent validity of the measures.

5.2 Exploratory factor analysis

Downloaded by Fudan University At 01:32 24 March 2017 (PT)

For the measurement property evaluation, exploratory factor analysis (EFA) was
conducted to explore the factor structure. Before performing the factor analysis, we verified
that the data were appropriate for factor analysis using the Kaiser–Meyer–Olkin (KMO) test
and the Bartlett sphericity test. The results of both tests indicated that a factor analysis would
be useful given our data (KMO=0.815; χ2 = 1502.457; df =136 p < .000). The initial factor
analysis using principal components analysis extracted five factors that were evident on the
scree plot, all with an eigenvalue greater than one. Factor loadings for the effective use of the
data aggregation block ranged from 0.894 to 0.928, the effective use of the data analysis block
ranged from 0.675 to 0.865, the effective use of data interpretation from 0.819 to 0.910, the
absorptive capacity block ranged from 0.686 to 0.850, and the decision making effectiveness
block ranged from 0.673 to 0.857. Overall, the results for EFA achieved standard factor
loadings of 0.5 as the cut-off significance, confirming that individual factors can indeed be
identified in a given block of dimensions.

5.3 Confirmatory factor analyses and measurement model

A measurement model was then analyzed to assess the measurement quality of the
constructs using a confirmatory factor analysis (CFA). The measurement model consisted of
five factors. The loading ranges for these five factors were as follow: the effective use of data
aggregation, 0.816 to 0.932; the effective use of data analysis, .574 to .825; the effective use of
data interpretation, 0.830 to 0.945; absorptive capacity, 0.674 to 0.845; and decision making
effectiveness, 0.700 to 0.793. The model chi-square was not statistically significant (χ2 (109) =
143.117, p > .05), which indicates that the exact fit hypothesis should be accepted. The
comparative fit index (CFI) was 0.976, which exceeds the cutoff value of .80, and the
standardized root mean square residual (SRMR) was .0557. The root mean square error of the
approximation (RMSEA) was .046, which is less than .08. Thus, we concluded that our data
adequately fit the measurement model.

5.4 Mediating effect testing

To test the mediating effects of absorptive capacity, we compared five alternative models
in terms of their fit statistics and path coefficients. The fit statistics for the models are shown in
Table 4. First, the proposed model (Model A) in which the path coefficients among the five
latent variables were freely estimated was tested. The absolute value of and CFI was well
above .95 and SRMR and RMSEA were both less than .08 for Model A. Then, a series of
alternative structural models were tested against each other. After comparing Model B, in

12
 which all path coefficients among the five latent variables were constrained to zero, to the
direct model (Model C), in which all path coefficients to and from absorptive capacity were
constrained to zero, we found that Model C produced a significantly better fit to the data
compared to Model B. In Model C, we examined the impact of BA captivity alone on decision
making effectiveness. The results revealed that the path coefficient was significant from the
effective use of data interpretation tools to decision making effectiveness, but insignificant
from the effective use data analysis and aggregation tools to decision making effectiveness.
Next, Model D, in which all path coefficients from the three forms of BA capabilities were
constrained to zero, was also compared to the baseline model (Model B). Hypothesis 1 was
supported because Model D produced a significantly better fit to the data compared to Model B
and the path coefficient from absorptive capacity to decision making effectiveness was
significant.
The full mediation model (Model E), in which all path coefficients from the three forms of
BA capabilities to decision making effectiveness were constrained to zero, was then compared
to Model C and Model D. The results showed that Model E produced a significantly better fit to
the data compared to either Model C or Model D, indicating that the effective use of data
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

analysis and interpretation tools positively affects absorptive capacity. Thus, Hypothesis 3 and
Hypothesis 4 were supported, but Hypothesis 2 was not supported. Finally, the proposed
model (Model A) was compared to Model E; the results showed that Model A fit the data
slightly better than Model E. We thus concluded that our proposed model (Model A) provided
the most parsimonious fit to the data.
The paths and parameter estimates for the proposed model (Model A) are shown in Figure
2, which indicates that absorptive capacity had the greatest association with decision making
effectiveness and the path coefficients from business capabilities to absorptive capacity
became insignificant after adding a mediator (in this case, absorptive capacity). While it
mediated the relationships between the effective use of data analysis tools and both the
effective use of data interpretation tools and the decision making effectiveness, it failed to
mediate the relationship between the effective use of data aggregation tools and decision
making effectiveness because the path coefficient between effective use of data aggregation
tools and absorptive capacity was not significant. As the direct effects of the effective use of
data analysis tools on decision making effectiveness was not significant, this indicates that
absorptive capacity fully mediated the relationship between them. However, as the direct
effects of effective use of data interpretation tools on decision making effectiveness was
significant, the absorptive capacity only partially mediated the relationship between them.
To further confirm the mediating role of absorptive capacity, a bootstrapping analysis was
used to assess the significance of each indirect effect. As recommended by Cheung and Lau
(2008), we set the number of bootstrap samples as 1,000. The results showed that the two-sided
bias-corrected bootstrap confidence interval for the indirect effect of data interpretation tools
on decision making effectiveness through absorptive capacity was [0.269, 0.511], that for the
indirect effect of data aggregation tools on decision making effectiveness was [-0.016, 0.0149]
and for the indirect effect of data analysis tools on decision making effectiveness it was [0.018,
0.314]. Thus, the indirect (mediated) effects of data analysis and interpretation tools on
decision making effectiveness were both significant, whereas the indirect effect of data
aggregation tools on the decision making effectiveness was not significant, consistent with the
aforementioned results. Thus, Hypotheses 5b and 5c were supported, but Hypothesis 5a was
not supported.

-------------------
Insert Table 3 here
-------------------

13
 -------------------
Insert Figure 2 here
-------------------

6. Discussion
Big data related technologies have become the most influential IT innovations in the last
decade. Medical professionals urge their peer institutions to leverage the new data governance,
collection and analysis approaches for gaining a holistic understanding of health from
large-scale patient data that goes beyond the current state of knowledge about treatments and
diseases. The main objective of this research was to advance our understanding of the way BA
enables healthcare units to enjoy better decision making through the absorption of the new
knowledge provided by the BA systems. By applying the RBT and the dynamic capability
view, this study proposes a conceptual model in which BA capabilities, as lower-order IT
capabilities, exert influence on decision making effectiveness through a higher-order
capability, namely absorptive capacity. The empirical evidence collected for this study
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

supports five key findings.

First, the results strongly support the claim that healthcare organizations’ BA capabilities
– both the effective use of data analysis and interpretation tools – can help improve its
absorptive capacity. This finding is consistent with prior studies that emphasized the notion
that the amplifying role of lower-order IT or operational capabilities can be developed to
improve organizational capabilities (e.g., Liu et al., 2013; Pavlou and El Sawy, 2006; 2010).
Second, with respect to indirect effect, the effective use of interpretation tools in
healthcare units indirectly influences decision making effectiveness, an impact that is mediated
by absorptive capacity. This means that hospitals are likely to create valuable knowledge to
make sound clinical decisions as they utilize visual dashboards and metrics effectively (Jardine
et al., 2014; Spruit et al., 2014).
Third, consistent with the dynamic capabilities view which contends the effective of IT
capability on competitive advantage are fully mediated by dynamic capability (Pavlou and El
Sawy, 2006; 2010), the mediating effect test indicates the full mediation of absorptive capacity
on the relationship between the effective use of data analysis tools and decision making
effective. Fourth, as we expected, the findings have highlighted the critical role of absorptive
capacity in achieving decision making effectiveness in health care.
Finally, the results of this study do not support the hypothesis on the association of the
effective use of data aggregation tools and absorptive capacity as well as decision making
effectiveness. A possible explanation is that the majority of respondents (74.34%) were from
the top-level management (i.e., CEO, CIO) and middle-level management (i.e., IT managers).
They are not responsible for aggregating and dealing with patient data in back-end systems.
Although data aggregation is a precursor requirement to data analysis and interpretation and
remains important, many care providers are suffering from data aggregation-related issues such
as the lack of data standards and data integration, data overload issues, and barriers to the
collection of high-quality data (Ashrafi et al., 2014; Shah and Patak, 2014; Ward et al., 2014)
Thus, healthcare managers have to be aware of the importance of data aggregation tools as
implementing BA systems. Based on these findings, we can offer some useful insights
regarding the theoretical and managerial implications of these findings.

6.1 Theoretical contributions and implications

One of the most discussed questions in the fields of IS and healthcare informatics is how
to use patient data and analysis effectively to support its clinical decisions. This study makes
three main contributions towards this question. First, the conceptualization and

14
 operationalization of the construct of BA capability has contributed to the development of a
deeper understanding of BA. Previous research has sought to measure BA capability by
modeling it as a one-dimension construct, instead choosing to focus solely on examinations of
the data analysis process (Cao et al., 2015). However, such approaches may unintentionally
overlook other important facets of BA capability, such as its ability to visualize data. Thus, the
construct of BA capability draws on a broader view of IT functionality that allows us to capture
the concept of BA more fully by reviewing its functionalities and how it is actually
implemented in real-world healthcare units to conceptualize the BA capability. This
conceptualization is the first step towards building a much needed body of knowledge on the
business value of BA and provides researchers with a useful lens through which to examine the
effectiveness of BA systems in supporting various organizational practices.
Second, in this study, a theoretical basis for the relationship between BA capability and
decision making effectiveness was elucidated by adopting a knowledge absorptive capacity
perspective that is rooted in RBT and dynamic capability view. Our results demonstrate how
knowledge absorption matters when implementing BA to the decision making process by
examining its mediation role. Specifically, our finding suggests that the effective use of data
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

analysis and aggregation tools has no business value, thus affirming the commonly held view
of the IT productivity paradox in the healthcare context (Jones et al., 2012). However, the
mediating role of absorptive capacity not only provides a mechanism by which BA can
contribute to decision making practices, but also offers a new solution to the puzzle of the IT
productivity paradox in healthcare settings.
Finally, previous studies have developed the BA value creation models to demonstrate the
managerial, economic, and strategic impacts of BA from the different theoretical perspectives.
Yet, these are generic and do not meet the healthcare industry’s particular requirements
(Foshay and Kuziemsky, 2014). Thus, our proposed model would help healthcare
organizations recognize the business value of BA and guide them through the process to
effectively utilize BA for decision making in clinical settings.

6.2 Implications for practice

For project leaders who are responsible for BA adoption, this study provides a set of
interesting insights that may affect the scope of their current BA projects. First, even if IT
vendors have enthusiastically advocated the potential benefits of BA when used for various
business practices, BA implementation requires organizational changes if it is to be effective.
In addition to the technological issues of BA, managers must also turn their attention to
integrating knowledge management into BA initiatives, focusing particularly on ways to
harness BA-generated knowledge. Healthcare organizations must constantly seek and
disseminate new knowledge obtained by analyzing patient data to respond to industry
regulations and market needs. Our findings suggest that the ability to obtain and apply
knowledge becomes critical for healthcare organizations since patient data per se cannot
generate value. Thus, a strong knowledge management protocol could add tremendous value
during BA implementation.
Second, our results show that data interpretation is a crucial capability that directly
impacts decision making effectiveness. Although BA can create convenient summarized
reports or charts, the key to making these reports meaningfully is to equip managers and
employees with relevant professional skills. Incorrect interpretation of the reports generated
could lead to serious errors of judgment and questionable decisions. Managers should provide
suitable analytical training courses for the employees who will play a critical support role in the
new information-rich work environment if organizations are to make best use of their new
opportunities to transfer data into knowledge.

15
 7. Conclusion
Notwithstanding the above-mentioned contributions and implications, our study is
inevitably subject to some limitations. First, different industries have different needs, goals and
expectations when implementing BA solutions. We targeted healthcare industries for this study,
so the generalizability of the results is limited, because data were only collected from a limited
sample consisting of large hospitals in Taiwan. Thus, our findings are not applicable to
healthcare industries in other countries. Second, the sample size used for validating the BA
capability scale was relatively small, although the representativeness of our sample may
overcome the sample size issue to some extent. More than 70% of the participants in this study
served as senior IT executives and were thus able to provide strategic overviews of the BA
implementation in their healthcare organizations. Meanwhile, by carefully taking various steps
for scale development, we tried to minimize the potential bias. Third, in order to make stronger
conclusions from research, further empirical research should validate the scales of business
analytics capability by utilizing larger samples. Finally, given its exclusive focus on BA
capability, our study does not consider other possible factors contributing to BA success.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

In response to the limitations of the current study, we offer some suggestions for future
research. A more comprehensive study is now needed that examines other factors that may
serve as enablers or moderating or mediating roles for this path. As the business value of IT
research suggests, several human IT resource (e.g., the analytical personnel’s skills), other
organizational capability factors (e.g., dynamic capability, improvisational capabilities),
organizational complementary resources (data government, synergy, and culture), and
environmental factors (market and environmental turbulent) could all play a role and should
thus be examined. Also, rather than examining the aforementioned factors with singular
causation and linear associations, future studies could seek to capture the complex interactions
of the interdependencies among BA capabilities and other organizational elements, and
examine how different configurations create improved business value.
In conclusion, our primary research objective was to unravel the relationships among BA
capability, absorptive capacity and decision making effectiveness. With our focus on the role
of absorptive capacity, we found that BA systems may indeed reveal new opportunities for
transforming decision making process. Consequently, the findings of this study provide
interesting new insights into knowledge management, contributing to the BA literature by
proposing a BA-enabled decision making effectiveness model that takes into account the effect
of absorptive capacity.

Acknowledgements

A previous version of the paper was presented at the 2015 Pre-ICIS (International Conference
on Information Systems) SIG GlobDev 8th Annual Workshop, Fort Worth, Texas, December
13, 2015. The authors are very thankful to the attendants in this workshop for the valuable
feedback received. The authors also would like to acknowledge Ms. Muyin Lo from a southern
medical center in Taiwan and Dr. Chiahui Yu from National Chengchi University who assisted
in the data collection.

Appendix 1: Measurement Items

Effective use of data aggregation tools (Newly developed)
Please rate the effectiveness by which your organization uses the following business analytics
tools in the healthcare services.

16
 1. Collect data from external healthcare sources and from various health systems
throughout your organization.
2. Make patient records consistent, visible and easily accessible for further analysis.
3. Store patient data into appropriate databases.
Effective use of data analysis tools (Newly developed)
1. Identify important business insights and trends to improve healthcare services.
2. Predict patterns of care in response to patient needs.
3. Analyze data in near-real or real time that allows responses to unexpected clinical
events.
4. Analyze social media data to understand current trends from a large population.
Effective use of data interpretation tools (Newly developed)
1. Provide systemic and comprehensive reporting to help recognize feasible opportunities
for care improvement.
2. Support data visualization that enables users to easily interpret results.
3. Provide near-real or real time information on health care operations and services within
healthcare facilities and across health care systems.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

Absorptive capacity (Pavlou and El Sawy, 2010)

Please rate the effectiveness by which your organization can acquire, assimilate, transform, and
exploit knowledge with the aid of business analytics.
1. We have effective routines to identify value, and import new information and knowledge.
2. We have adequate routines to assimilate new information and knowledge.
3. We are effective in transforming existing information into new knowledge.
4. We are effective in utilizing knowledge into new services.

Decision making Effectiveness

1. As a result of business analytics systems, the quality of decisions has improved (Sanders
and Courtney, 1985).
2. As a result of business analytics systems, the speed at which we analyze decisions has
increased. (LaValle et al., 2011; Wixom et al., 2013)
3. As a result of business analytics systems, we have an increased understanding of our
customers. (Cao et al., 2015)

17
 References
Ashrafi, N., Kelleher, L. and Kuilboer, J. P. (2014), “The impact of business intelligence on
healthcare delivery in the USA”, Interdisciplinary Journal of Information, Knowledge,
and Management, Vol. 9, pp. 117-130.
Bardhan, I., Oh, J. H., Zheng, Z. and Kirksey, K. (2014), “Predictive analytics for readmission
of patients with congestive heart failure”, Information Systems Research, Vol. 26 No. 1,
pp. 19-39.
Barjis, J., Kolfschoten, G. and Maritz, J. (2013), “A sustainable and affordable support system
for rural healthcare delivery”, Decision Support Systems, 56, 223-233.
Barney, J. (1991), “Firm resources and sustained competitive advantage”, Journal of
Management, Vol. 17 No. 1, pp. 99-120.
Barney, J.B. (2001), “Resource-based theories of competitive advantage: a ten-year
retrospective on the resource-based view”, Journal of Management, Vol. 27 No. 6, pp.
643-650.
Bates, D.W., Saria, S., Ohno-Machado, L., Shah, A. and Escobar, G. (2014), “Big data in
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

health care: using analytics to identify and manage high-risk and high-cost patients”,
Health Affairs, Vol. 33 No. 7, pp. 1123-1131.
Bharadwaj, A. S. (2000), “A resource-based perspective on information technology capability
and firm performance: an empirical investigation”, MIS Quarterly, Vol. 24 No. 1, pp.
169-196.
Bollen, K. and Lennox, R. (1991), “Conventional wisdom on measurement: a structural
equation perspective”, Psychological Bulletin, Vol. 110 No. 2, pp. 305-314.
Brewer, M.B., Campbell, D.T. and Crano, W.D. (1970), “Testing a single-factor model as an
alternative to the misuse of partial correlations in hypothesis-testing research”,
Sociometry, Vol. 33 No. 1, pp. 1-11.
Brislin, R.W. (1990). “Applied cross-cultural psychology: an introduction”, in Brislin RW
(Ed.), Applied Cross-cultural Psychology. Newbury Park, CA: Sage, pp. 9-33.
Cao, G., Duan, Y. and Li, G. (2015), “Linking business analytics to decision making
effectiveness: a path model analysis”, IEEE Transactions on Engineering Management,
Vol. 62 No. 3, pp. 384-395.
Chen, H., Chiang, R.H.L. and Storey, V.C. (2012), “Business intelligence and analytics: from
big data to big impact”, MIS Quarterly, Vol. 36 No. 4, pp. 1165-1188.
Cheung, G.W. and Lau, R.S. (2008), “Testing mediation and suppression effects of latent
variables: bootstrapping with structural equation models”, Organizational Research
Methods, Vol. 11 No. 2, pp. 296-325.
Cohen, W.M. and Levinthal, D.A. (1990), Absorptive capacity: a new perspective on learning
and innovation. Administrative Science Quarterly, 35(1), 128-152.
Cortada, J.W., Gordon, D. and Lenihan, B. (2012), The Value of Analytics in Healthcare: From
Insights to Outcomes. Somers, NY: IBM Global Business Services.
Dean, J.W. and Sharfman, M.P. (1996), “Does decision process matter? A study of strategic
decision-making effectiveness”, Academy of Management Journal, Vol. 39 No. 2, pp.
368-392.
Delen, D. (2014). Real-World Data Mining: Applied Business Analytics and Decision Making.
FT Press.
DeLone, W.H. and McLean, E.R. (1992), “Information systems success: the quest for the
dependent variable”, Information Systems Research, Vol. 3 No. 1, pp. 60-95.
Elbashir, M.Z., Collier, P.A. and Sutton, S.G. (2011), “The role of organizational absorptive
capacity in strategic use of business intelligence to support integrated management
control systems”, The Accounting Review, Vol. 86 No. 1, pp. 155-184.

18
 Erevelles, S., Fukawa, N. and Swayne, L. (2016), “Big data consumer analytics and the
transformation of marketing”, Journal of Business Research, Vol. 69 No. 2, pp. 897-904.
Fan, W. and Gordon, M.D. (2014), “The power of social media analytics”, Communications of
the ACM, Vol. 57 No. 6, pp. 74-81.
Ferranti, J.M., Langman, M.K., Tanaka, D., McCall, J. and Ahmad, A. (2010), “Bridging the
gap: leveraging business intelligence tools in support of patient safety and financial
effectiveness”, Journal of the American Medical Informatics Association, Vol. 17 No. 2,
pp. 136-143.
Fihn, S.D., Francis, J., Clancy, C., Nielson, C., Nelson, K., Rumsfeld, J., Cullen, T., Bates, J.
and Graham, G. L. (2014), “Insights from advanced analytics at the Veterans Health
Administration”, Health Affairs, Vol. 33 No. 7, pp. 1203-1211.
Fink, L., Yogev, N., and Even, A. (2016). “Business intelligence and organizational learning:
An empirical investigation of value creation processes”, Information & Management.
doi:10.1016/j.im.2016.03.009.
Flatten, T.C., Engelen, A., Zahra, S.A. and Brettel, M. (2011), “A measure of absorptive
capacity: scale development and validation”, European Management Journal, Vol. 29 No.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

2, pp. 98-116.
Foshay, N. and Kuziemsky, C. (2014), “Towards an implementation framework for business
intelligence in healthcare”, International Journal of Information Management, Vol. 34
No. 1, pp. 20-27.
Francalanci, C. and Morabito, V. (2008), “IS integration and business performance: the
mediation effect of organizational absorptive capacity in SMEs”, Journal of Information
Technology, Vol. 23 No. 4, pp. 297-312.
Gefen, D., Rigdon, E.E. and Straub, D. (2011), “An update and extension to SEM guidelines
for administrative and social science research”, MIS Quarterly, Vol. 35 No. 2, pp. iii-xiv.
Ghosh, B. and Scott, J.E. (2011), “Antecedents and catalysts for developing a healthcare
analytic capability”, Communications of the Association for Information Systems, Vol. 29
No. 1, pp. 395-409.
Grewal, R. and Slotegraaf, R. J. (2007), “Embeddedness of organizational capabilities”.
Decision Sciences, Vol. 38 No. 3, pp. 451-488.
Groves, P., Kayyali, B., Knott, D. and Kuiken, S.V. (2013), The “Big Data” Revolution in
Healthcare: Accelerating Value and Innovation. McKinsey & Company.
Hair Jr.,J.F., Black, W., Babin, B. and Anderson, R. (2010), Multivariate data analysis (7th ed.).
Pearson Education Inc.
Halamka, J.D. (2014), “Early experiences with big data at an academic medical center”, Health
Affairs, Vol. 33 No. 7, pp. 1132-1138.
Hu, H., Wen, Y., Chua, T. S., and Li, X. (2014), “Toward scalable systems for big data
analytics: a technology tutorial”, Access, IEEE, Vol. 2, pp. 652-687.
Işık, Ö., Jones, M. C. and Sidorova, A. (2013), “Business intelligence success: the roles of BI
capabilities and decision environments”, Information & Management, Vol. 50 No. 1, pp.
13-23.
Jardine, A., Mullan, N., Gudes, O., Cosford, J., Moncrieff, S., West, G., Xiao, J., Yun, G. and
Someford, P. (2014), “Web-based geo-visualisation of spatial information to support
evidence-based health policy: A case study of the development process of 'Health
Tracks'”, Health Information Management Journal, Vol. 43 No. 2, pp. 7-16.
Jones, S.S., Heaton, P.S., Rudin, R.S. and Schneider, E.C. (2012), “Unraveling the IT
productivity paradox—lessons for health care”, New England Journal of Medicine, Vol.
366 No. 24, pp. 2243-2245.

19
 Joshi, K. D., Chi, L., Datta, A. and Han, S. (2010), “Changing the competitive landscape:
continuous innovation through IT-enabled knowledge capabilities”, Information Systems
Research, Vol. 21 No. 3, pp. 472-495.
Kash, B.A., Spaulding, A., Gamm, L. and Johnson, C.E. (2013). “Health care administrators’
perspectives on the role of absorptive capacity for strategic change initiatives: A
qualitative study”, Health Care Management Review, Vol. 38 No. 4, pp. 339-348.
Kowalczyk, D.W.I.M. and Buxmann, P. (2014), “Big data and information processing in
organizational decision processes”, Business & Information Systems Engineering, Vol. 6
No. 5, pp. 267-278.
Kwon, O., Lee, N. and Shin, B. (2014), “Data quality management, data usage experience and
acquisition intention of big data analytics”, International Journal of Information
Management, Vol. 34 No. 3, pp. 387-394.
LaValle, S., Lesser, E., Shockley, R., Hopkins, M.S. and Kruschwitz, N. (2011), “Big data,
analytics and the path from insights to value”, MIT Sloan Management Review, Vol. 52
No. 2, pp. 21-31.
Lev, S., Fiegenbaum, A., and Shoham, A. (2009). “Managing absorptive capacity stocks to
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

improve performance: Empirical evidence from the turbulent environment of Israeli

hospitals”, European Management Journal, Vol. 27 No. 1, pp. 13-25.
Lewis, B.R., Templeton, G.F. and Byrd, T.A. (2005), “A methodology for construct
development in MIS research”, European Journal of Information Systems, Vo.14, No.4,
pp. 388-400.
Lichtenthaler, U. (2009), “Absorptive capacity, environmental turbulence, and the
complementarity of organizational learning processes”, Academy of Management Journal,
Vol. 52 No. 4, pp. 822-846.
Lin, Y., Brown, R. A., Yang, H.J., Li, S., Lu, H. and Chen, H. 2011. “Data mining large-scale
electronic health records for clinical support”, IEEE Intelligent Systems, Vol. 26 No. 5, pp.
87-90.
Lin, Y., Wang, Y. and Kung, L. (2015), “Influences of cross-functional collaboration and
knowledge creation on technology commercialization: evidence from high-tech
industries”, Industrial Marketing Management, Vol. 49, pp. 128-138.
Liu, H., Ke, W., Wei, K. K. and Hua, Z. (2013), “The impact of IT capabilities on firm
performance: The mediating roles of absorptive capacity and supply chain agility”,
Decision Support Systems, Vol. 54 No. 3, pp. 1452-1462.
Mackenzie, S.B., Podsakoff, P.M. and Podsakoff, N.P. (2011), “Construct measurement and
validation procedures in MIS and behavioral research: integrating new and existing
techniques”, MIS Quarterly, Vo. 35 No. 2, pp. 293-334.
Malhotra, A., Gosain, S. and El Sawy, O.A. (2005), “Absorptive capacity configurations in
supply chains: gearing for partner-enabled market knowledge creation”, MIS Quarterly,
Vol. 29 No. 1, pp. 145-187.
Meador, C.L., Guyote, M.J. and Keen, P.G. (1984), “Setting priorities for DSS development”,
MIS Quarterly, Vol. 8 No. 2, pp. 117-129.
Nunnally, J.C. and Bernstein, I.H. (1994), Psychometric Theory, New York: McGraw-Hill.
Pavlou, P.A. and El Sawy, O.A. (2006), “From IT leveraging competence to competitive
advantage in turbulent environments: the case of new product development”, Information
Systems Research, Vol. 17 No. 3, pp. 198-227.
Pavlou, P.A. and El Sawy, O.A. (2010), “The “third hand”: IT-enabled competitive advantage
in turbulence through improvisational capabilities”, Information Systems Research, Vol.
21 No. 3, pp. 443-471.

20
 Pavlou, P.A., Liang, H. and Xue, Y. (2007), “Understanding and mitigating uncertainty in
online environments: a principal-agent perspective”, MIS Quarterly, Vol. 31 No. 1, pp.
105-136.
Phillips-Wren, G., Iyer, L.S., Kulkarni, U. and Ariyachandra, T. (2015), “Business analytics
in the context of big data: a roadmap for research”, Communications of the Association
for Information Systems, Vol. 37, pp. 448-472.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y. and Podsakoff, N.P. (2003), “Common method
biases in behavioral research: a critical review of the literature and recommended
remedies”, Journal of Applied Psychology, Vol. 88, pp. 879-903.
Popovič, A., Hackney, R., Coelho, P.S. and Jaklič, J. (2012), “Towards business intelligence
systems success: effects of maturity and culture on analytical decision making”, Decision
Support Systems, Vol. 54 No. 1, pp. 729-739.
Power, D.J. (2008), “Decision support systems: a historical overview”, in Handbook on
Decision Support Systems, Springer, Berlin Heidelberg.
Raghupathi, W. and Raghupathi, V. (2014), “Big data analytics in healthcare: promise and
potential”, Health Information Science and Systems, Vol. 2 No. 1, pp. 3.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

Riabacke, M., Danielson, M. and Ekenberg, L. (2012), “State-of-the-art prescriptive criteria

weight elicitation”, Advances in Decision Sciences, vol. 2012, pp. 1-24.
Roberts, N., Galluch, P.S., Dinger, M. and Grover, V. (2012), “Absorptive capacity and
information systems research: review, synthesis, and directions for future research”, MIS
Quarterly, Vol. 36 No. 2, pp.625-648.
Roski, J., Bo-Linn, G. and Andrews, T. (2014), “Creating value in health care through big data:
opportunities and policy implications”, Health Affairs, Vol. 33 No. 7, pp. 1115-1122.
Ross, J.W., Beath, C.M., and Quaadgras, A. (2013), “You may not need big data after all”,
Harvard Business Review, Vol. 91 No. 12, pp. 90-98.
Sanders, G.L. and Courtney, J.F. (1985), “A field study of organizational factors influencing
DSS success”, MIS Quarterly, Vol. 9 No. 1, pp. 77-93.
Santhanam, R. and Hartono, E. (2003), “Issues in linking information technology capability to
firm performance”, MIS Quarterly, Vol. 27 No. 1, pp. 125-153.
Schneeweiss, S. (2014). “Learning from big health care data”, The New England Journal of
Medicine, Vol. 370 No. 23, pp. 2161-2163.
Shanks, G. and Bekmamedova, N. (2012), “Achieving benefits with business analytics systems:
an evolutionary process perspective”, Journal of Decision Systems, Vol. 21 No. 3, pp.
231-244.
Sharma, R. and Yetton, P. (2003), “The contingent effects of management support and task
interdependence on successful information systems implementation”, MIS Quarterly, Vol.
27 No. 4, pp. 533-556.
Sharma, R., Mithas, S. and Kankanhalli, A. (2014), “Transforming decision-making processes:
a research agenda for understanding the impact of business analytics on organisations”,
European Journal of Information Systems, Vol. 23 No. 4, pp. 433-441.
Spruit, M., Vroon, R. and Batenburg, R. (2014), “Towards healthcare business intelligence in
long-term care: an explorative case study in the Netherlands”, Computers in Human
Behavior, Vol. 30, pp. 698-707.
Srinivasan, U. and Arunasalam, B. (2013), “Leveraging big data analytics to reduce healthcare
costs”, IT Professional, Vol. 15 No. 6, pp. 21-28.
Trkman, P., McCormack, K., De Oliveira, M.P.V. and Ladeira, M.B. (2010), “The impact of
business analytics on supply chain performance”, Decision Support Systems, Vol. 49 No.
3, pp. 318-327.

21
 Wamba, S.F., Akter, S., Edwards, A., Chopin, G. and Gnanzou, D. (2015), “How ‘big data’ can
make big impact: findings from a systematic review and a longitudinal case study”,
International Journal of Production Economics, Vol. 165, pp.234-246.
Wang, Y., Kung, L., and Byrd, T.A. (2016), “Big data analytics: Understanding its capabilities
and potential benefits for healthcare organizations”, Technological Forecasting and
Social Change, doi:10.1016/j.techfore.2015.12.019.
Wang, Y., Kung, L., Ting, C. and Byrd, T.A. (2015), “Beyond a technical perspective:
understanding big data capabilities in health care”, in 48th Hawaii International
Conference on System Sciences (HICSS) in Kauai, Hawaii, IEEE, pp. 3044-3053.
Wang, Y., Kung, L., Wang, W.Y.C. and Cegielski, C. (2014), “Developing a big data-enabled
transformation model in healthcare: A practice based view’, in 34th International
Conference on Information Systems (ICIS) in Milan, Italy.
Ward, M. J., Marsolo, K. A. and Froehle, C.M. (2014), “Applications of business analytics in
healthcare”, Business Horizons, Vol. 57 No. 5, pp. 571-582.
Watson, H.J. (2014), “Tutorial: big data analytics: concepts, technologies, and applications”,
Communications of the Association for Information Systems, Vol. 34, pp. 1247-1268.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

Wixom, B., Yen, B. and Relich, M. (2013), “Maximizing value from business analytics”, MIS
Quarterly Executive, Vol. 12 No. 2, pp. 111-123.
Wu, L. and Hu, Y.P. (2012), “Examining knowledge management enabled performance for
hospital professionals: A dynamic capability view and the mediating role of process
capability”, Journal of the Association for Information Systems, Vol. 13 No. 12, pp.
976-999.
Zahra, S.A. and George, G. (2002), “The net-enabled business innovation cycle and the
evolution of dynamic capabilities”, Information Systems Research, Vol. 13 No. 2, pp.
147-150.

Author biographies

Yichuan Wang is an Assistant Professor at Newcastle University London. He received his PhD
degree in Management Information Systems from Auburn University. His current
research interests are healthcare information technology, big data analytics, and social
media marketing. His articles have appeared in Journal of Business Research,
International Journal of Production Economics, Industrial Marketing Management,
Technological Forecasting and Social Change, International Journal of Information
Management, and Journal of Computer Information Systems, among others.

Terry Anthony Byrd is Bray Distinguished Professor of Management Information Systems

(MIS) at the Raymond J. Harbert College of Business, Auburn University. He holds a BS
in Electrical Engineering from the University of Massachusetts at Amherst and a PhD in
MIS from the University of South Carolina. His research has appeared in MIS Quarterly,
Journal of Management Information Systems, European Journal of Information Systems,
Decision Sciences, Information Systems Journal, Journal of Association of Information
Systems and other leading journals. His current research interests focus on the design,
development, implementation, diffusion, and infusion of information technology in
facilitating a variety of individual, group organizational and societal behaviors and
initiatives to achieve positive results, especially in the healthcare domain.

22
 Table 1. Key constructs of BA capability
BA systems Tools Key functionalities Effective use of BA systems
• Collect data from external
• Middleware sources and from various
• Data warehouse • Extracting data from systems throughout the
Data • Extract-transform-load large amounts of data healthcare units
aggregation (ELT) tools • Transforming data • Make data consistent, visible
tools • Hadoop distributed into standard formats and easily accessible for
file system (HDFS) • Data storage analysis
• NoSQL database • Store data into appropriate
databases
• Processing large • Identify important business
amounts of insights to improve costly
unstructured and healthcare services such as
semi-structured data unnecessary diagnostic tests
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

across a massively and treatments

• Apache Hadoop/Map parallel cluster of • Predict pattern of care to
Reduce servers using Hadoop quickly response patient
• Statistical analysis Map/Reduce needs
Data analysis • OLAP • Real-time analysis by • Analyze data in near-real or
tools • Predictive modeling utilizing stream real time that allows to
• Social media analytics computing quickly respond to
• Machine learning • In-database analytics unexpected events
• Text mining/NLP for analyzing the • Analyze social media data
structure of patient such as patient subjective
records opinions, medicine
• Social media recommendations and ratings
analytics for to understand current trends
analyzing web data in a large population
• Provide systemic and
comprehensive reporting
mechanisms to help
recognize feasible
opportunities for
• General summary of improvement
• Visual
Data data • Support data visualization
dashboards/systems
interpretation • Visualization that enables users to easily
tools • Reporting
reporting interpret results
systems/interfaces
• Real-time reporting • Provide near-real or real time
information on health care
operations and services
within health care facilities
and across health care
systems
 Table 2. Descriptive Statistics and Correlations
Variable Mean S.D. α CR 1 2 3 4 5
Effective use of data aggregation 4.40 1.42 0.91 0.92 0.78
Effective use of data analysis 4.65 1.33 0.84 0.85 0.05 0.59
Effective use of data interpretation 3.97 1.20 0.91 0.91 0.19* 0.05 0.78
Absorptive capacity 3.66 1.10 0.85 0.86 0.21** 0.19* 0.50** 0.60
Decision-making effectiveness 4.32 1.14 0.80 0.80 0.11 0.17* 0.47** 0.47** 0.57
Note: N=152; AVEs on diagonal
CR: Composite reliability; *p<0.05, **p<0.01, ***p<0.001
Downloaded by Fudan University At 01:32 24 March 2017 (PT)
 Table 3. Model fit summary and nested model comparisons
Model Chi-square df p-value ∆ χ2 CFI SRMR RMSEA (90C.I.)
A 150.248 112 .009 - 0.973 0.0785 0.048 (0.025, 0.066)
B 245.963 119 .000 95.715 0.911 0.1962 0.840 (0.069, 0.099)
C 209.907 116 .000 59.659 0.935 0.1719 0.073 (0.057, 0.089)
D 210.907 118 .000 60.659 0.935 0.1730 0.072 (0.056, 0.088)
E 162.321 115 .002 12.073 0.967 0.0880 0.052 (0.032, 0.070)
Notes: SRMR = standard root mean square residual; CFI = comparative fit index; RMSEA = root mean square
error of approximation.
The proposed model served as the baseline for chi-square difference testing
Model A: the proposed model, no path coefficients among the five latent variables were constrained to zero.
Model B: all path coefficients among the five latent variables were constrained to zero.
Model C: all path coefficients to and from absorptive capacity were constrained to zero.
Model D: all path coefficients from BA capabilities were constrained to zero.
Model E: all path coefficients from the BA capabilities to decision-making effectiveness were constrained to
zero.
Downloaded by Fudan University At 01:32 24 March 2017 (PT)
 H5b

Effective use of
data aggregation
tools H5a
H2

Effective use of H3 Absorptive H1 Decision-making

data analysis tools capacity effectiveness

H4
Effective use of H5c
data interpretation
tools
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

Figure 1. The proposed research model

 .106

Effective use of
data aggregation
tools
-.014
.068

Effective use of .163* Absorptive Decision-making

capacity .441*** Effectiveness
data analysis tools
R2=.301 R2=.404
.163***

Effective use of .304***

data interpretation
Downloaded by Fudan University At 01:32 24 March 2017 (PT)

tools

Figure 2. Path diagram and standardized estimates

Note: Summary of standardized path coefficients for the hypothesized model with the full
sample (N = 152). Solid lines represent significant coefficients, and dotted lines represent
non-significant coefficients, * p < .05; ** p < .01; *** p < .001.