Health Care Management Review
Issue: Volume 37(1), January/March 2012, p 4-13
Copyright: (C) 2012 Lippincott Williams & Wilkins, Inc.
Publication Type: [Special Section: Health Information Technology and
Management in the Era of Reform]
DOI: 10.1097/HMR.0b013e318239f2ff
ISSN: 0361-6274
Accession: 00004010-201201000-00002
Keywords: data envelopment analysis, efficiency, EMR, hospital EMR
adoption, strategy
[Special Section: Health Information Technology and Management in the Era of
Reform]
Hospital electronic medical record enterprise application strategies: Do they
matter?
Fareed, Naleef; Ozcan, Yasar A.; DeShazo, Jonathan P.
Author Information
Naleef Fareed, MBA, is Doctoral Student, Department of Health Administration, PO
Box 980203, Virginia Commonwealth University, Richmond. E-mail: [email protected].
Yasar A. Ozcan, PhD, is Professor, Department of Health Administration, PO Box
980203, Virginia Commonwealth University, Richmond. E-mail: [email protected].
Jonathan P. DeShazo, MPH, PhD, is Assistant Professor of Biomedical Informatics,
Department of Health Administration, PO Box 980203, Virginia Commonwealth
University, Richmond. E-mail: [email protected].
The authors have disclosed that they have no significant relationships with, or
financial interest in, any commercial companies pertaining to this article.
---------------------------------------------Outline
Abstract
Theory/Conceptual Framework
SV Strategy
BOB Strategy
BOS Strategy
Methods
Data and Data Sources
Efficiency Techniques and Statistical Approaches
Data envelopment analysis
Post hoc analysis
Variable Construction
DEA inputs and outputs
Efficiency
Hospital EMR enterprise application strategies
EMR implementation status—moderator
Hospital characteristics
Market characteristics
Results
Efficiency Results
Post hoc Analysis Results
Discussion
Practice Implications
References
Abstract
Background: Successful implementations and the ability to reap the benefits of
electronic medical record (EMR) systems may be correlated with the type of
enterprise application strategy that an administrator chooses when acquiring an
EMR system. Moreover, identifying the most optimal enterprise application
strategy is a task that may have important linkages with hospital performance.
Purpose: This study explored whether hospitals that have adopted differential
EMR enterprise application strategies concomitantly differ in their overall
efficiency. Specifically, the study examined whether hospitals with a single-vendor
strategy had a higher likelihood of being efficient than those with a best-of-breed
strategy and whether hospitals with a best-of-suite strategy had a higher
probability of being efficient than those with best-of-breed or single-vendor
strategies. A conceptual framework was used to formulate testable hypotheses.
Methodology: A retrospective cross-sectional approach using data envelopment
analysis was used to obtain efficiency scores of hospitals by EMR enterprise
application strategy. A Tobit regression analysis was then used to determine the
probability of a hospital being inefficient as related to its EMR enterprise
application strategy, while moderating for the hospital's EMR "implementation
status" and controlling for hospital and market characteristics.
Findings: The data envelopment analysis of hospitals suggested that only 32
hospitals were efficient in the study's sample of 2,171 hospitals. The results
from the post hoc analysis showed partial support for the hypothesis that
hospitals with a best-of-suite strategy were more likely to be efficient than
those with a single-vendor strategy.
Practice Implications: This study underscores the importance of understanding
the differences between the three strategies discussed in this article. On the
basis of the findings, hospital administrators should consider the efficiency
associations that a specific strategy may have compared with another prior to
moving toward an enterprise application strategy.
---------------------------------------------Successful implementation and the ability to reap the benefits of electronic
medical record (EMR) systems may be correlated with the type of enterprise
application strategy that an administrator chooses when acquiring an EMR system.
Enterprise application strategy is defined as the means through which an
organization designs information technology to facilitate the cooperation and
coordination of work across the organization (Gartner, 2011). An administrator's
ability to select the best enterprise application strategy for their EMR system
could help increase the availability of important information that improves the
speed and quality of decision making. This in turn may possibly also help
support a hospital's provision of more health care services with fewer resources
(i.e., efficiency).
Hospital administrators are faced with three different enterprise application
strategies when acquiring EMR systems: single vendor (SV), best of breed (BOB),
and best of suite (BOS). These strategies can be envisioned across a continuum,
where at one end is SV, which provides the most integrated system solution for a
hospital, whereas at the other end is BOB, which provides a hospital with the
optimal set of differentiated systems for specific functions (e.g., surgery,
nursing, and finance). In the middle of the continuum is BOS, which is a hybrid
of BOB and SV solutions.
When considering an EMR enterprise application strategy, hospital administrators
are faced with the conundrum of how much differentiation or integration of the
various information system (IS) functions is required for their hospital.
Hospitals pursuing an SV strategy attempt to integrate administrative and
clinical applications over multiple locations under a single global software
application (Ford, Menachemi, Huerta, & Yu, 2010; Hermann, 2010). This software
usually does not require interfaces to ensure that all applications are
communicating with each other (Hermann, 2010). Those using a BOB strategy
typically seek to integrate different IS components developed by multiple
vendors, hopefully closely align the IS functions with the different requirements
of each specific hospital unit (Ford et al., 2010). Using interface engines
(e.g., Health Level Seven), applications are allowed to communicate between each
other (Hermann, 2010). The BOS strategy, the most recent of the three strategies,
helps hospitals take advantage of elements present in both of the previously
described strategies. Here, hospitals can utilize the enhanced functionality of
BOB products for some IS functions while also using an SV strategy to bundle
other IS functions (Hoehn, 2010). The rationale behind the SV strategy is that
integration among certain core IS functions could provide more value versus a
focus on their individual applications (Hoehn, 2010).
This study primarily questions whether hospitals that have adopted differential
EMR enterprise application strategies concomitantly differ in their overall
efficiency. For hospital administrators, strategy selection may be related to
efficiency gains linked to prudent decision making, which could also be
associated with negative financial performance, and potential quality risks. For
policy makers, strategy selection provides insights into what Health Information
Technology strategies might work and how hospitals may effectively achieve
"meaningful use" of their EMRs. Thus, this article attempts to help identify
which of the previously described EMR enterprise application strategies are
associated with hospitals' greater likelihood of being efficient.
Theory/Conceptual Framework
Organizational theorists have explored the nature, characteristics, and
performances of organizational integration and differentiation strategies.
Lawrence and Lorsch (1967a) defined integration as "the process of achieving
unity of effort among the various subsystems in the accomplishment of the
organization's task" (p. 4) and differentiation as "the state of segmentation of
the organizational system into subsystems, each of which tends to develop
particular attributes in relation to the requirements posed by its relevant
external environment" (p. 4). Some organizational theorists have posited that
managers, through the use of integration or differentiation strategies, are
capable of actively enacting tools and mechanisms (part of an organization's
structure), which in turn determine aspects of behavior in an organization
(Scott & Davis, 2007, p. 126). Similarly, this could be translated into the
context of this article's focus on a hospital administrator's rationale for
selecting a certain EMR enterprise application strategy. Here, an administrator's
choice of one EMR enterprise application strategy, as noted in the prior
discussion around the level of integration across the enterprise application
strategies, may range from BOB being the highly differentiated strategy to SV
being the highly integrated strategy. The BOS strategy would fall in the middle
as a hybrid. The selection of an EMR enterprise application strategy may also
have associations that are congruent with the previously noted arguments made by
the organizational theorists about the influence of organizational strategies on
behaviors within an organization. Hence, the EMR enterprise application
strategies could provide certain benefits and limitations relative to one
another that may be correlated with the efficient provision of health care
delivery. Possible associations from choosing each strategy and their relative
superiority are discussed in turn.
SV Strategy
Lawrence and Lorsch (1967a) noted that differences in group assumptions (i.e.,
work orientations), clusters of roles, and diverging perspectives on tasks and
goals lead to the inability of two or more groups to achieve effective
collaboration, integration, and cooperation. Pinsonneault and Kraemer (2002), in
their study of two organizations, found that managers of the organization with
the more integrated IS applications were able to streamline operations,
integrate functions, and operate more efficiently than the less integrated
organization.
In regard to SV, Ford et al. (2010) noted that hospitals that pursued this
strategy were able to streamline routine functions such as claims management,
reduce transaction costs through the use of a single contract with a vendor,
focus on competency building on just that product, and easily centralize the
process of maintaining the EMR system. Nevertheless, an SV strategy could entail
the use of systems that may not fit well with existing clinical processes;
require extensive transformations of policies, processes, and other activities
within a hospital; and build resistance among employees if they are forced to
adjust to a new standardized system that does not accommodate the unique aspects
of their tasks (Ehie & Madsen, 2005; Ford et al., 2010).
BOB Strategy
To accommodate multiple objectives, an organization may pursue a strategy of
differentiation to attend more closely to what customers want and what
competitors are doing (Scott & Davis, 2007). In an environment with high
uncertainty and complex tasks, Burns and Stalker (1961) proposed that organizations
would elect to pursue strategies of differentiation to accommodate diverse
interests, power differences, and work processes that were not capable of being
integrated across various individuals.
In relation to BOB, Ford et al. (2010) indicated that a hospital administrator
may choose such a strategy in an effort to accommodate for the needs of
individual departments, in particular clinical practice preferences. Moreover, a
BOB strategy may be pursued to obtain the best technology in the industry for a
particular function. The strategy may also require relatively less investment,
face fewer resistance from staff, and be implemented much faster and thus help
improve the efficiency of care delivery (Hermann, 2010). Nonetheless, the
strategy faces important limitations. First and foremost, a BOB strategy is
typically linked with a fragmented IS system within a hospital (Ford et al.,
2010). Hospital administrators are faced with negotiating across several
vendors; administrators are confronted with increasing risks that entail from
multiple contracts with and agreeing to the varying objectives of the many
external vendors (Hermann, 2010). Second, interface engines are required to pass
information from one application to another, which may be troublesome if
crosswalks between systems do not occur properly. As noted by Leavitt (1962), as
organizations adopt a differentiation strategy, they are faced with more
problems of communication among subgroups. In the case of BOB systems, Hoehn
(2010) noted that "unfortunately, what we learned[horizontal ellipsis]was that
integration engines were not so easy to implement or maintain, and that data
within these islands of automation were not consistently defined and were a lot
harder to move across the enterprise" (p. 11). In light of the previously noted
major limitations, hospitals are faced with new factors that may increase their
inefficiencies; thus, it may be posited that,
Hypothesis 1: Hospitals pursuing an SV strategy are more likely to be efficient
than hospitals that have a BOB strategy.
BOS Strategy
Organizations may combine both integration and differentiation strategies to
create an inclusionary means through which they may deal with the varying
requirements of subsystems within an organization and also manage the interdependence
needs of the organization at the same time (Lawrence & Lorsch, 1967a). Lawrence
and Lorsch (1967a, 1967b) also noted that modern administrators were constantly
struggling with the need to reconcile the needs for standardization and
coordination; based on the possibility that organizations can achieve differentiation
and integration simultaneously, their research showed that these organizations
may in fact perform better. Hence, managers may actually be able to address the
needs for subsystems and total organizational performance concurrently.
In terms of BOS, managers are able to obtain the best value from integration and
specialization by assessing the relative merits of either option for a
particular application (Hoehn, 2010). This hybrid approach comes with less
transaction costs than a BOB strategy does; allows hospitals to outsource their
core, administrative functions and focus on the clinical applications that
increase their competitive advantage; and dampens potential disruptions of work
processes brought about by completely redesigning work tasks or staff resistance
as experienced through an SV strategy (Ford et al., 2010; Thouin, Hoffman, &
Ford, 2008). Hospital efficiency may be best optimized by this hybrid strategy.
The BOS strategy still contains limitations that are similar to those found in
BOB and SV strategies; for example, hospitals still have to deal with multiple
vendors. Despite the limitations, Ford et al. (2010) surmised that the BOS
strategy was, in all practicality, the strategy that had the best potential to
yield hospitals with the most benefits in relation to the other two strategies.
Hence, it may be hypothesized that,
Hypothesis 2: Hospitals pursuing a BOS strategy are more likely to be efficient
than hospitals that have an SV or a BOB strategy.
An empirical strategy is designed and utilized to test the previously listed
hypotheses. The description of this study's data sources, sampling strategy,
variables, and statistical techniques are presented in the following section.
Methods
Data and Data Sources
A retrospective cross-sectional design was used to examine the relationship of
EMR enterprise application strategy and efficiency. Data sources included the
American Hospital Association (AHA) 2008 Annual Survey of Hospitals, the 2008
Centers for Medicare and Medicaid Services (CMS) case mix index, the 2008 Area
Resource Files (ARF), and the 2008 Health Information Management Systems Society
(HIMSS) Analytics database. Nonfederal, general acute care hospitals were the
unit of analysis.
A total of 4,875 general medical hospitals from the AHA data set were merged
with 2,973 hospitals in the HIMSS data set. This data set was then merged with
the 2008 ARF and 2008 CMS data, respectively. The final study sample resulted in
2,171 hospitals. Table 1 provides a comparison of the hospitals in our sample
with all general acute care hospitals in the AHA data.
Briefly, the average hospital in our study sample had more beds than the average
AHA hospital does. A majority of the study hospitals were nonprofit and system
affiliated. There were fewer teaching hospitals among the study sample, but more
of the hospitals were located in urban markets. Table 1 also provides a
frequency distribution of the hospitals in the final sample by their EMR
enterprise application strategy.
Efficiency Techniques and Statistical Approaches
Data envelopment analysis
Data envelopment analysis (DEA) was used to measure hospital efficiency in this
study. Using this nonparametric method, a "best practice" frontier is identified
through which a decision-making unit (DMU) is compared among its peers. An
efficiency score of 1 indicates that a DMU is efficient and is on the frontier,
whereas scores between 0 and 1 are indicative of DMUs that are inefficient and
that fall outside the frontier. Input models focus on the extent to which input
quantities can be reduced without changing output quantities, whereas output
models focus on an organization's attempt to maximize outputs without altering
input quantities (Mark, Jones, Lindley, & Ozcan, 2009). The assumption is that
with input-oriented models, managers have control over their inputs and not
their outputs; the converse is assumed with output-oriented models. Another
important distinction involves whether researchers assume constant returns to
scale or variable returns to scale (VRS). The constant returns to scale model
assumes that "there is a linear, proportional change in outputs for changes in
inputs whereas the VRS assumes that returns are dependent on changes in volume"
(Mark et al., 2009, p. 183). In this analysis, hospitals were the DMUs. An
input-oriented model is selected for the same reasons as described previously
and the study used a VRS model because hospitals of different size (which are
pooled together in this study) cannot be assumed to have similar economies of
scale (Ozcan, 2008, p. 43).
Post hoc analysis
An optimal technique for this study's analyses is the maximum likelihood
regression method: Tobit, which adjusts for the bounded dependent efficiency
variable by censoring data that are equal to 1 (Chilingerian, 1995). As
recommended by Chilingerian (1995), the DEA scores were transformed with the
following formula:
Inefficiency score = (1/DEA score) - 1
Thus, setting the score of 1 as the lower bound limit for the Tobit analysis,
the equation for the study was as follows:
Inefficiency score = f(EMR enterprise application strategies, EMR enterprise
application strategy x EMR implementation status, hospital characteristics, and
market characteristics)
It is essential to note that the signs of all coefficients are reversed: where a
positive sign means an association with inefficiency and a negative sign means
an association with higher levels of efficiency (Chilingerian, 1995).
Variable Construction
DEA inputs and outputs
For the DEA, a robust model is used to capture the performance that can be
attributed primarily to hospital management (Ozcan, 2008, p. 108). Input
measures included hospital bed size, service mix, full-time equivalent (FTE)
labor force, and other operational expenses. Ozcan and Luke (1993) demonstrated
that bed size is a valid proxy for hospital assets and, thus, capital investments
as well. Another important capital input, service mix, was used to measure the
technological complexity of hospitals and their ability to perform complex
procedures (Kazley & Ozcan, 2009; O'Neill, Rauner, Heidenberger, & Kraus, 2008).
In the United States, physicians are generally not hospital employees; hence,
this study prudently attributes labor as nonphysician labor or their FTEs-this
includes nursing, diagnostic, therapy, and technical personnel (Ozcan, 2008, p.
107). Other operational expenses accounted for medical supplies, utilities, and
related expenses minus the expenses for the labor force (Ozcan, 2008, p. 107).
The FTE labor force consisted of all nonphysician full-time employees plus the
weighted (i.e., 0.5) number of part-time personnel employed. Service mix is the
weighted sum of 139 potential services that might be offered by a hospital. The
weights for each service were based on the assessment of three independent
experts in the field of health services research (Ozcan & Luke, 2011).
Case-mix adjusted admissions and outpatient visits were used as the output
measures. The former can be debated as being a better measure of inpatient
activity than "inpatient days" because the introduction of the Prospective
Payment System has led to hospitals shifting their focus away from inpatient
days to the actual "case" as their primary means of collecting reimbursement
(O'Neill et al., 2008). The latter measure has allowed for increased differentiation
of visits, which is important due to the variation in reimbursement levels.
Case-mix adjusted admissions were calculated as the number of inpatient
admissions multiplied by the hospital's average Medicare case mix for that year.
Unfortunately, lack of data availability that provides the case mix of
outpatient visits precludes such adjustments for this measure. The aforementioned
input and output variables have been commonly used in previous studies
(Hollingsworth, 2008; Ozcan, 2008, p. 109). Data from the AHA and CMS were the
primary sources for the DEA inputs and output variables.
Efficiency
Efficiency scores are computed using a DEA-based software application (Saitech
Inc., 2011). Like in Hsieh, Clement, and Bazzoli (2010), hospital efficiency is
defined "as the minimum feasible reduction of inputs while holding the amount of
outputs constant" (p. 78), and this definition was selected "because hospitals
are more able to control their inputs (i.e., costs, labors, capital assets) than
their outputs (i.e., outpatient visits, inpatient discharges)" (p. 78).
Hospital EMR enterprise application strategies
In the post hoc analysis, the independent variables of interest concern the
hospital's EMR enterprise application strategy. The HIMSS (2011) analytics
database records a hospital's EMR enterprise application strategy based on three
philosophies: SV, BOB, and BOS. In estimating the relationship of a hospital's
EMR enterprise application strategy and probability of efficiency, potential
confounders may exist. An important moderator in the aforementioned relationship
is EMR implementation status. Hospital-level (i.e., internal factors) and
market-level (i.e., environmental factors) confounders may also be associated
with the selection of an EMR enterprise application strategy used by a hospital
or their efficiency and thus need to be accounted for. A discussion of these
variables follows.
EMR implementation status-moderator
Leatt and Schneck (1984) argued that organizational outcomes are determined
primarily by the "fit" between key aspects of an organization's structure and
its operating environment. Pfeffer and Salancik (1978) noted that variations in
the ability to match between an organization's information processing requirements
and its information processing capabilities could affect its performance.
Likewise, efficiency gains (or losses) from a certain vendor selection strategy
could be achieved (or lost) even before a hospital achieves a fully automated
EMR system. Several categories of EMR implementation status are recorded in the
HIMSS analytics database. These include hospitals that (a) possess a fully
automated EMR, (b) have not automated their EMR, (c) are in the process of
installing an EMR, (d) have an EMR that is to be replaced, (e) have contracted
with a vendor or not yet installed their EMR, or (f) have not yet contracted
with a specific vendor for an EMR but have enacted processes to evaluate
purchasing an application (HIMSS, 2011). Because EMR implementation status could
either enhance or dampen the effect size of a hospital's efficiency that could
be linked to its EMR enterprise application strategy, multiplicative interaction
terms (Pennings, 1987) are created to account for this phenomenon. That is, each
EMR enterprise application strategy is multiplied by a hospital's respective EMR
implementation status (while using three, different interaction terms, which all
contain fully automated EMR, as the reference categories).
Hospital characteristics
This study's post hoc analysis control for a hospital's ownership type, system
affiliation, teaching status, bed size, whether it had Joint Commission on
Accreditation of Healthcare Organizations (JCAHO) accreditation, and payer mix.
All the variables in this group were obtained from the AHA.
Hsieh et al. (2010) and Lee, Chun, and Lee (2008) argued that ownership type
influenced a hospital's level of efficiency due to their varying objectives and
missions regarding the delivery of health care. In addition, Rosko (1999) found
that for-profit hospitals tend to be less efficient than nonprofits, where the
latter also provides more uncompensated care. This study used three separate
indicators for ownership type: for profit, nonprofit, and public.
Rosko, Proenca, Zinn, and Bazzoli (2007), in their study, found that system-affiliated
hospitals tended to offer more efficient care than did non-system-affiliated
hospitals. The authors reasoned that hospitals with system membership are able
to eliminate duplication of operational functions, take advantage of economies
of scale and scope, and also better coordinate services across facilities (Rosko
et al., 2007).
Lee et al. (2008) also suggested that a hospital's teaching status may affect
the level of efficiency. In this case, the authors argued that these types of
hospitals tended to provide a broader range of services and thus use more input
services, which may in turn decrease their level of efficiency versus nonteaching
hospitals (Lee et al., 2008).
Ozcan (1992) and Lee et al. (2008) noted that a hospital's efficiency may be
significantly affected by its size (as measured by number of beds). One possible
reason for this is that small hospitals have different roles in the delivery of
care and that they simply do not have the economies of scope to provide some of
the wide-ranging services that larger hospitals offer (Lee et al., 2008). Like
most health services studies, the log of the bed size measurement was used in
this study because there was an extreme skew in the distribution of the data.
Chen, Rathore, Radford, and Krumholz (2003) noted that hospitals accredited by
JCAHO were expected to strive toward high standards of quality and efficiency.
By complying with JCAHO's standards, hospitals are able to avoid adverse events
and medical errors and hence require fewer and less resources to discharge
patients with the best possible outcomes (Chen et al., 2003).
Hospitals with higher proportions of Medicare or Medicaid patients may be forced
to operate more efficiently because they are faced with higher uncertainty of
receiving payments and due to the relatively low rates of reimbursement from
these programs (Hsieh et al., 2010; Lee et al., 2008). In reaction to these
factors, hospitals with higher proportions of Medicare or Medicaid patients may
attempt to operate more efficiently by controlling the types of services offered
and/or the amount of uncompensated care provided (Hsieh et al., 2010; Lee et
al., 2008).
Market characteristics
This study's post hoc analysis also controlled for hospital services area (HSA)
competition, county population (a proxy for market size), county unemployment
rate (a proxy for the amount of uninsured), county percentage of the population
over 65 years of age, the county percentage of African Americans in the
population, and geographical location. Data for the variables in this group were
obtained from the ARF and AHA.
Chang and Troyer (2009) argued that hospitals in less concentrated (i.e., more
competitive) markets may feel more pressure to operate efficiently than
hospitals in monopolistic markets in an effort to maintain ideal allocations of
service provisions while also maximizing profits. To measure hospital competition,
the Herfindahl-Hirschman Index is constructed using the sum of the squared
market shares for all hospitals in their respective HSA codes (Dartmouth Atlas,
2011), where market shares are computed using the number of inpatient days for
each hospital in the HSA code.
In their longitudinal assessment of the effect of market factors on hospital
efficiency, Wang, Ozcan, Wan, and Harrison (1999) found that hospitals with
larger markets had significant excesses in input resources that resulted in high
inefficiencies. The log of the population measurement was used in this study
because there was an extreme skew in the distribution of the data.
Rosko (1999) noted that hospitals typically reallocated their capacities to
become more efficient in an effort to adjust for increased needs of indigent
care. This study uses a popular proxy variable to measure the rate of uninsured
in a hospital's market: unemployment rate.
The demographics within a hospital's market can also affect the demand and
supply of health services and, thus, its efficiency. Two important factors
included in this study were a County's proportion of African Americans and the
proportion of individuals over the age of 65. Both groups have been found to
require more capital-intensive services and have higher rates of hospitalizations
(Pappas, Wilbur, Kozak, & Fisher, 1997).
Hospital geographical location can be viewed from two aspects: the level of
urbaness and the region where the hospital is located. Lindrooth, Bazzoli,
Needleman, and Hasnain-Wynia (2006) suggested that substantial nursing shortages
were present in the past decade particularly in urban areas. On the basis of
this notion, Hsieh et al. (2010) reasoned that urban hospitals may have become
more efficient in their use of labor inputs for the production of hospital
services.
The demand and supply of services may also vary by the region within which a
hospital is located. To account for these variations that may have particularly
strong implications for the inputs in a hospital's production function, this
study used four indicators for region: Northeast, West, Midwest, and South.
Table 2 provides descriptive statistics of the DEA input and output variables
for the hospital sample. Table 3 provides descriptive statistics of the
independent variables used in the Tobit analysis.
Results
Efficiency Results
The results from the DEA demonstrated that only 32 hospitals were found to be
efficient based on the VRS efficiency scores. On average, the hospitals had an
efficiency score of 0.55.
Post hoc Analysis Results
Table 4 provides the results from the Tobit analysis. On the basis of the
results, the study's first hypothesis is rejected: Hospitals with an SV strategy
were not more likely to be efficient than hospitals with a BOB strategy.
However, the results provided partial support for the second hypothesis. Here,
it appears that hospitals with a BOS EMR enterprise application strategy may be
significantly (p
Other significant findings across the model showed that hospitals that were
teaching, had system membership, or were in larger markets were more likely to
be efficient. Conversely, hospitals that were nonprofit, had more beds, were
JCAHO accredited, had a greater share of Medicare or Medicaid patients, were in
an urban market, or were situated in the Northeast or Midwest (versus the South)
were found to be less likely to be efficient across both the models.
In regard to the interaction terms in the model, hospitals with an SV strategy
and an EMR status of "not automated" or have "contracted with a vendor or not
yet installed their EMR" were likely to be more efficient than hospitals with an
SV strategy and an EMR status of "fully automated." In contrast, hospitals with
an SV strategy and an EMR status of "not yet contracted" or those with a BOS
strategy and an EMR status of "not automated" were likely to be less efficient
than their respective, fully automated counterparts.
Discussion
This study examined whether hospitals with an SV strategy were more likely to be
efficient than those with a BOB strategy and whether those with a BOS strategy
would be more likely to be efficient than those with an SV or a BOB strategy. A
Tobit model was designed to test the hypotheses using efficiency scores obtained
from DEA techniques. The DEA of hospitals in 2008 suggested that only 32
hospitals in the study's sample were efficient. The results from the post hoc
analysis showed partial support for the second hypothesis; that is, hospitals
with a BOS strategy were more likely to be efficient than those with an SV
strategy.
Based on this study's analysis, the moderate association of the BOS strategy
with efficiency, in a complex organizational setting (i.e., hospitals), is an
important phenomenon that needs to be acknowledged. A BOS strategy may be an
optimal balance of consistency and flexibility for hospitals to pursue to
quickly implement their EMR systems and effectively manage their complex
processes. With the many intricacies and hurdles present in achieving meaningful
use of EMR systems, a BOS strategy could tactically help hospitals rapidly meet
meaningful use requirements and obtain reimbursements from the pay-for-performance
program initiated by the Office of the National Coordinator for Health
Information Technology (Ford et al., 2010). Moreover, the strategy could also
increase satisfaction among clinicians and patients due to the strategically
effective adoption of EMR systems.
It is also evident from this study's findings that the focus of the discussions
around hospital EMR enterprise application strategies should shift away from the
more prevailing and dichotomous comparison between SV and BOB strategies and
refocus on the nuanced and understudied comparison between SV and BOS strategies.
Nonetheless, findings suggest that hospitals using a BOB strategy may be as
efficient as those using a BOS or SV strategy. The greater flexibility of a BOB
strategy does provide advantages over the other strategies. Conversely, given
the realities that these systems require hospitals to go to great lengths to
ensure crosswalks occur smoothly between IS applications, hospital administrators
will have to carefully consider their ability to meet these important requirements
early on in the purchasing phase.
Using EMR implementation status as a moderator in the study provided some
interesting results. Although two of the four significant findings were
unsurprising (i.e., their fully automated counterparts were more efficient), the
remaining results were contradictory to popular beliefs in the industry. It is
unclear as to why hospitals with an SV strategy and an EMR status of "not
automated" or "contracted with a vendor or not yet installed their EMR" were
likely to be more efficient than hospitals with an SV strategy and an EMR status
of "fully automated." It is possible that this phenomenon may be associated with
the nature of EMR implementation for hospitals that adopt an SV strategy, which
may be very different from hospitals adopting the other EMR enterprise
application strategies. More research in this area should be considered.
There are several limitations of our study worth noting. First, EMR systems
contain a diverse set of features that may or may not be the same across two
systems. Moreover, it is important to be aware that an SV EMR system may contain
applications from different vendors, which could possibly be "bundled" solutions
than truly integrated systems. Second, although HIMSS allows a convenient
categorization of a hospital's EMR enterprise application strategy, a hospital's
current stated strategy may not reflect its current application portfolio and
might even be simply dictated by previous practices. Third, the level at which a
hospital's administration effectively carries out the stated strategy is another
dimension that could not be accounted for in this study as well. It is vital to
be considerate of the aforementioned phenomenon because there may be scenarios
where, for example, a hospital with a BOB strategy may be very successful at
integrating all its applications and thus possibly outperform another, similar
hospital that has an SV strategy. Fourth, the DEA methodology reflects a
relative assessment of a hospital's efficiency versus a truly absolute level of
efficiency. Moreover, the selection of inputs and outputs can strongly influence
the overall outcomes in the DEA. Nonetheless, the variables used in this study
were determined based on their relevance to the study's objectives, their common
use in prior literature, and the availability of data. Fifth, the cross-sectional
nature of the study's design prevents us from presenting a causal link among the
study's relationships. Hence, a longitudinal assessment is recommended as the
next step in research related to EMR enterprise application strategies and
hospital efficiency. Despite these limitations, the DEA methodology and post hoc
analysis provide a good starting point for researchers to further delve into the
important subject matter of hospital EMR strategies.
Practice Implications
Although this study's findings do not identify clear relationships between each
enterprise application strategy, they do underscore the importance of understanding
the differences between the three strategies. Hospital administrators should
consider the potential correlation hospital efficiency may have with one
enterprise application strategy over another. Large capital requirements and the
potential ability to garner generous government incentives should be enough
motivation to convince chief executive officers and chief information officers
the need to pay more attention to what enterprise application strategy their
hospital pursues.
References
Burns T., Stalker B. (1961). The management of innovations. London, England:
Tavistock.
Chang C., Troyer J. (2009). The impact of TennCare on hospital efficiency.
Health Care Management Science, 12, 201-216. Bibliographic Links
Chen J., Rathore S., Radford J., Krumholz H. (2003). JCAHO accreditation and
quality of care for acute myocardial infarction. Health Affairs, 22, 243-254.
Bibliographic Links
Chilingerian J. A. (1995). Evaluating physician efficiency in hospitals: A
multivariate analysis of best practices. European Journal of Operational
Research, 8, 548-574.
Dartmouth Atlas. (2011). The Dartmouth atlas of health care. Retrieved August
20, 2011, from http://www.dartmouthatlas.org/
Ehie I., Madsen M. (2005). Identifying critical issues in enterprise resource
planning (ERP) implementation. Computers in Industry, 56, 545-557.
Ford E., Menachemi N., Huerta T., Yu F. (2010). Hospital IT adoption strategies
associated with implementation success: Implications for achieving meaningful
use. Journal of Healthcare Management, 55, 175-189. Bibliographic Links
Gartner. (2011). IT definitions and glossary. Retrieved August 20, 2011, from
http://www.gartner.com/technology/it-glossary/
Health Information and Management Systems Society. (2011). Health Information
and Management Systems Society: Data handbook. Retrieved from http://www.himss.org
Hermann S. (2010). Best-of-breed versus integrated systems. American Journal of
Health-System Pharmacy, 67, 1406-1410.
Hoehn B. (2010). Integration issues: What is the best clinical information
systems strategy? Journal of Health Information Management, 24, 10-12.
Hollingsworth B. (2008). The measurement of efficiency and productivity of
health care delivery. Health Economics, 17, 1107-1128. Bibliographic Links
Hsieh H., Clement D., Bazzoli G. (2010). Impacts of market and organizational
characteristics on hospital efficiency and uncompensated care. Health Care
Management Review, 35, 77-87. Ovid Full Text Bibliographic Links
Kazley A., Ozcan Y. (2009). Electronic medical record use and efficiency: A DEA
and windows analysis of hospitals. Socio-Economic Planning Sciences, 43,
209-216.
Lawrence P., Lorsch J. (1967a). Differentiation and integration in complex
organizations. Administrative Science Quarterly, 12, 1-47.
Lawrence P., Lorsch J. (1967b). Organization and environment: Managing
differentiation and integration. Cambridge, MA: Harvard Business School Press.
Leatt P., Schneck R. (1984). Criteria for grouping nursing subunits in
hospitals. Academy of Management Journal, 27, 150-165. Bibliographic Links
Leavitt H. J. (1962). Management according to task: Organizational differentiation.
Management International, 2, 13-22.
Lee K., Chun L., Lee J. (2008). Reforming the hospital service structure to
improve efficiency: Urban hospital specialization. Health Policy, 87, 41-49.
Bibliographic Links
Lindrooth C., Bazzoli G., Needleman J., Hasnain-Wynia R. (2006). The effect of
changes in hospital reimbursement on nurse staffing decisions at safety net and
nonsafety net hospitals. Health Service Research, 41, 701-720.
Mark B., Jones C., Lindley L., Ozcan Y. (2009). An examination of technical
efficiency, quality, and patient safety in acute care nursing units. Policy,
Politics, & Nursing Practice, 10, 180-186.
O'Neill L., Rauner M., Heidenberger K., Kraus M. (2008). A cross-national
comparison and taxonomy of DEA-based hospital efficiency studies. Socio-Economic
Planning Sciences, 42, 158-189.
Ozcan Y. (1992). Sensitivity analysis of hospital efficiency under alternative
output/input and peer groups: A review. Knowledge and Politics, 5, 1-29.
Ozcan Y. (2008). Health care benchmarking and performance evaluation: An
assessment using data envelopment analysis (DEA). Norwell, MA: Springer.
Ozcan Y., Luke R. (1993). A national study of the efficiency of hospitals in
urban markets. Health Services Research, 28, 719-739. Bibliographic Links
Ozcan Y., Luke R. (2011). Health care delivery restricting and productivity
change: Assessing the Veterans Integrated Service Networks (VISNs) using the
Malmquist approach. Medical Care Research and Review, 68, 20S-35S. Bibliographic
Links
Pappas G., Wilbur C., Kozak L., Fisher G. (1997). Potentially avoidable
hospitalizations: Inequalities in rates between US socioeconomic groups.
American Journal of Public Health, 87, 811-816. Buy Now Bibliographic Links
Pennings M. (1987). Structural contingency theory: A multivariate test.
Organizational Studies, 8, 223-240.
Pfeffer J., Salancik R. (1978). The external control of organizations: A
resource dependence perspective. New York, NY: Harper & Row.
Pinsonneault A., Kraemer K. (2002). Exploring the role of information
technology in organizational downsizing: A tale of two American cities.
Organization Science, 13, 191-208.
Rosko M. (1999). Impact of internal and external environmental pressures on
hospital inefficiency. Health Care Management Science, 2, 63-74. Bibliographic
Links
Rosko M., Proenca J., Zinn J., Bazzoli G. (2007). Hospital inefficiency: What
is the impact of membership in different types of systems? Inquiry, 44, 335-349.
Saitech Inc. (2011). DEA-Solver Pro5.0 [Computer software]. Holmdel, NJ: Saitech
Inc.
Scott R., Davis G. (2007). Organizations and organizing: Rational, natural, and
open system perspectives. Upper Saddle River, NJ: Prentice Hall.
Thouin M., Hoffman J., Ford E. (2008). The effect of information technology
(IT) investments on firm-level performance in the healthcare industry. Health
Care Management Review, 33, 60-69. Ovid Full Text Bibliographic Links
Wang B. B., Ozcan Y. A., Wan T. T., Harrison J. (1999). Trends in hospital
efficiency among metropolitan markets. Journal of Medical Systems, 23, 83-97.
Bibliographic Links
Key words: data envelopment analysis; efficiency; EMR; hospital EMR adoption;
strategy
----------------------------------------------