Keywords
data analysis - laboratory information systems - pathology - monitoring and surveillance
- order entry
Background and Significance
Background and Significance
Cost, quality, outcomes, and patient safety concerns continue to motivate efforts
to improve the utilization of health care services, including the laboratory. Health
care providers have been estimated to influence up to 80% of the costs of health care,[1] and it is widely believed that laboratory data are a major driver of medical decision
making.[2] In addition, studies have shown that the usage of medical services, including procedures
and laboratory tests, varies greatly based on region and geography.[3]
[4] This high level of variability suggests inappropriate utilization, including both
over- and underutilization of laboratory testing.[5] With the limited availability of medical resources, the need for judicious utilization
of laboratory tests becomes essential.
Health care systems have responded to the need to closely manage laboratory testing
with the formation of laboratory utilization management programs designed to monitor
new and existing test usage.[6]
[7] Data required for these efforts typically consists of reports from the laboratory
information system (LIS) to provide test level information including volume, result
values, and provider data for the tests to be monitored. However, while LIS data are
important in monitoring test usage, in our experience the information collected in
LIS reports is often insufficient for comprehensive and efficient utilization management.[8]
The inability of LIS data alone to provide insight into many utilization issues appears
to be largely due to the complexities introduced into the laboratory test ordering
process with the increased prevalence and capabilities of the electronic health record
(EHR).[9]
[10] An increasing percentage of health care systems worldwide now utilize an EHR that
typically includes many different pathways for laboratory ordering by providers, including
selecting from lists, searching, and using predefined order sets and panels.[11]
[12]
[13]
Well-designed EHR systems provide a platform to streamline workflow, standardize laboratory
ordering, promote adherence to guidelines, and deliver decision support alerts.[14]
[15]
[16] With several hundred to thousands of tests on a typical laboratory menu, search
functionality is found in most EHR implementations that permits providers to rapidly
search for a laboratory order in a given list of laboratory tests. With search, clinicians
enter search terms and the system returns a list of matching tests. In addition to
searching for laboratory tests, EHRs typically allow providers to create personalized
preference lists with tests they find particularly helpful or order most frequently.
Further, standardized order sets provided in most EHR implementations offer an opportunity
for the system to guide order selection by clinicians.[17]
[18]
[19]
[20] Order sets are groups of orders that are frequently ordered together. For example,
an anemia order set may include a complete blood count, ferritin, total iron binding
capacity, serum folate, vitamin B12, and/or other related tests.
It is well established that timely access to LIS data is useful to inform and improve
test utilization.[8] However, to our knowledge no reports have examined the usage of an EHR reporting
database to enhance a laboratory utilization program. Herein, we describe the construction
of an EHR laboratory orders database that we have used to generate targeted analyses
to advance a utilization management program.
Objectives
The aims of this study were (1) to develop a near real-time EHR laboratory orders
database with data relevant to laboratory utilization, (2) to examine the ability
of the EHR laboratory orders database to improve an established laboratory utilization
program, and (3) to document the key EHR fields with the highest value to laboratory
utilization management.
Methods
Setting
These initiatives were performed at the Massachusetts General Hospital (MGH), a 999-bed
tertiary care teaching hospital in Boston, Massachusetts, United States. The MGH is
a founding member of Partners HealthCare, a multi-institutional, collaborative, not-for-profit
health care system. These projects were completed as quality improvement initiatives
and, as such, were not formally reviewed by the institutional review board, per protocol.
The MGH Clinical Laboratories produce over 10 million test results per year with an
EHR orderable test menu of over 1,500 tests. Laboratory test utilization is managed
primarily by faculty and staff within the Department of Pathology, which contains
multiple interdisciplinary committees involved in test addition and removal, utilization,
and monitoring of utilization initiative effectiveness.[6] The Core Laboratory, under the umbrella of the Department of Pathology, is also
responsible for the budget for laboratory testing that is sent outside the system
to be performed by reference laboratories.
The Partners HealthCare system, including the MGH, uses Epic (Epic Clinical Systems,
Verona, Wisconsin, United States) as its EHR system. The Epic EHR is utilized in all
clinical areas including inpatient, outpatient, and emergency department settings.
The MGH Clinical Laboratories also use the Sunquest LIS (Sunquest Information Systems,
Tucson, Arizona, United States).
EHR Orders Database Creation
We modified an existing Structured Query Language Server (Microsoft, Redmond, Washington,
United States) database to accommodate the new EHR test ordering data. We had previously
used this database to import and store LIS data, including all laboratory test results.
In collaboration with our health system's Epic reporting team, we developed an EHR
report containing fields pertinent to all laboratory orders that were updated (e.g.,
ordered, released, resulted, or cancelled) the prior day. Fields of interest to our
utilization program are captured in the EHR report, examples of which are shown in
[Table 1]. Recommendations and comments on the content of the EHR report are presented in
the “Discussion” section.
Table 1
Selected key data fields imported from an EHR report into the EHR orders database
|
Data field
|
Information type
|
Availability in LIS
|
Comment
|
|
Enterprise medical record number
|
Patient information
|
No
|
Global medical record number binding medical record numbers at individual sites
|
|
Local medical record number (site 1)
|
Patient information
|
Yes
|
Medical record number for principal site
|
|
Local medical record number (site 2)
|
Patient information
|
No
|
Medical record number (if present) for sister academic site
|
|
Encounter number
|
Patient information
|
Yes
|
Unique identifier for patient encounter during which laboratory was ordered
|
|
Ordering department identifier
|
Provider
|
Yes
|
Department of ordering provider
|
|
Authorizing provider identifier
|
Provider
|
No
|
Department of authorizing provider
|
|
Ordering provider identifier
|
Provider
|
Yes
|
Numeric provider identification number
|
|
Ordering provider location
|
Provider
|
Yes
|
Provider's practice location
|
|
EHR order number
|
Order information
|
Yes
|
Provides link to LIS result, as EHR order number is stored with the result within
the LIS
|
|
Order date/Time
|
Order information
|
Yes
|
Date and time of order
|
|
Order source
|
Order information
|
No
|
Application within EHR where order was placed
|
|
Procedure code
|
Order information
|
No
|
Laboratory order code in EHR
|
|
Display name
|
Order information
|
No
|
Laboratory order display name in EHR
|
|
Cancellation reason
|
Order information
|
Yes
|
For cancelled tests, reason for cancellation
|
|
Priority information
|
Order information
|
Yes
|
Routine or STAT
|
|
Expected date
|
Order information
|
No
|
Date laboratory is expected to be collected
|
|
Expiration date
|
Order information
|
No
|
Date when order will expire
|
|
Release date/Time
|
Order information
|
Yes
|
Date/time that laboratory order was sent to laboratory information system
|
|
Accession number
|
Order information
|
Yes
|
Accession number within LIS, if laboratory is in process or resulted
|
|
Provider order entry comments
|
Patient-specific order details
|
No
|
Comments made by the provider in the EHR about this order during the ordering process
|
|
Order entry questions
|
Patient-specific order details
|
No
|
Prompts asked of provider when ordering test
|
|
Order set identifier
|
Order origination information
|
No
|
Order set from which order was selected from
|
|
Order set name
|
Order origination information
|
No
|
Name of order set displayed to provider in EHR
|
|
Laboratory preference list name
|
Order origination information
|
No
|
Preference list where order was placed
|
|
Resulting laboratory identifier
|
Order origination information
|
No
|
Resulting laboratory that was populated or selected during order entry
|
|
Resulting laboratory name
|
Order origination information
|
No
|
Name of laboratory where testing will be performed
|
Abbreviations: EHR, electronic health record; LIS, laboratory information system.
Note: The availability or lack of availability of these EHR fields in the laboratory
information system is noted.
The EHR laboratory orders report is generated daily via Epic's “Clarity” reporting
database and sent as a delimiter separated file to a shared file area. The EHR laboratory
orders database was configured to automatically import these files using a structured
process to parse and store the EHR data. All EHR data are stored in a single table.
Typical daily EHR report files are 60 to 70 megabytes in size, and contain 140 to
160,000 rows of data. The import process incorporates a variety of automated quality
assurance functions including file validation, data validation, and error notification.
The process automatically archives the original delimited report after importing it.
Users generally access the EHR orders database via Open Database Connectivity (ODBC)
and Microsoft Access (Microsoft). Data analysis is performed in Microsoft Excel, Microsoft
Access, R, or Python.
Results
We have utilized the EHR orders database as a central part of our laboratory utilization
program. Below, we provide three representative examples demonstrating how the EHR
orders database has been integrated into our program. The ability to have detailed
information on each EHR laboratory order, including the providers involved, the EHR
application module, and point of origin of the order, has allowed us to rapidly respond
to shifts in laboratory ordering and to make targeted changes to the EHR to respond
to the identified utilization issues. We provide examples below in the areas of volume
monitoring, search optimization, and miscellaneous test monitoring.
Volume Monitoring
One of the key tasks for utilization management programs is to monitor for unexpected
changes in test volumes. Unexpected increases in individual tests may reflect practice
changes, provider hiring, EHR menu changes, new EHR order set usage, or EHR errors.
These changes are important to track since, in addition to impacting the diagnostic
efficiency of laboratory workups, they may also directly impact the laboratory operating
budget. In the case of high-cost reference laboratory testing, even a small shift
in test ordering can quickly have a significant budgetary impact.
With the use of a routine LIS monitoring report that compares the most recent weekly
and monthly test volumes of all tests to historical test result volumes, a nearly
10-fold increase in red blood cell (RBC) folate ordering was noted over a period of
several weeks ([Fig. 1]). This was concerning as the serum folate test is typically preferred over RBC folate
for evaluating patients with suspected folate deficiency.[21] Further, while serum folate is an inexpensive test performed 24 hours a day with
rapid turnaround time in our laboratory, RBC folate is an expensive test performed
at a reference laboratory with a much longer turnaround time. Moreover, the Choosing
Wisely Collaborative, along with the American Society for Clinical Pathology, has
recently identified RBC folate as a test that should rarely be ordered.[22]
Fig. 1 Red blood cell (RBC) folate monthly volumes. The arrow indicates the month where
the anemia order set containing RBC folate was updated to include serum folate in
the place of RBC folate.
Our LIS monitoring report, although capable of identifying the volumes and providers
that ordered the testing, was unable to provide the context of the ordering process,
so the LIS report was of limited value in determining the reason behind the increase
in RBC folate ordering. Thus, we examined our EHR orders database to analyze the EHR
orders and attempt to determine the root cause for the increase in RBC folate ordering.
An understanding as to how the orders originated in the EHR is important to be able
to design an effective intervention. Key fields in our EHR orders database (see the
“Discussion” section for listing of database fields) allow us to determine if the
order originated from a provider selecting the order from the full list of all available
tests, a specialty preference list (e.g., Pediatrics test list), a provider's personal
preference list (i.e., their “favorites”), or an order set. In this case, most RBC
folate requests were shown to be ordered from a single “Anemia” order set (> 90% of
total orders, data not shown). With knowledge of the origin of the orders, we contacted
the clinical group overseeing the anemia order set to request removal of the RBC folate
test and its replacement with serum folate. Following the intervention, the frequency
of RBC folate orders quickly returned to its prior baseline ([Fig. 1]).
Menu/Search Optimization
A key consideration for every health care system is which tests are available on the
laboratory testing menu (i.e., the laboratory formulary).[23] The choices presented to the user when searching for a laboratory test are important
determinants for what eventually will be ordered by the clinician. Naming conventions,
synonyms, and sort order may all influence the provider's eventual choice of test.[24] Inappropriate provider test selections may have significant downstream workflow
impacts on the laboratory if the tests that are overutilized are highly manual compared
with the more appropriate testing.
For patient workups for monoclonal gammopathies, a typical screening test is the serum
protein electrophoresis (SPEP) panel. In our institution, we utilize an institutional
reflex protocol when a SPEP panel is ordered. We first perform a SPEP, serum immunoglobulins,
and total protein testing and the results of these tests are then reviewed by a laboratory
medical director. At that point, based on prior history, immunoglobulin levels, SPEP
findings, and EHR review, the director may determine the need to perform serum immunofixation
testing to further evaluate the sample and/or further characterize any identified
M components.
When we migrated our institution to a new EHR (Epic), we included two tests related
to monoclonal gammopathy evaluation on the EHR laboratory menu, namely, “SPEP panel”
and “SPEP panel with immunofixation.” When the former test was ordered, the laboratory
would perform the reflex protocol outlined above and only perform immunofixations
when indicated by the algorithm. In contrast, when “SPEP panel with immunofixation”
was ordered, the laboratory would always perform both a SPEP and an immunofixation
on the sample.
Following implementation of the new EHR, the immunology section of our laboratory
reported a significant increase in immunofixation testing. This was substantiated
by our LIS monitoring report demonstrating a steady increase in serum immunofixation
orders ([Fig. 2]). In our prior order entry systems, before implementation of the new EHR, the “SPEP
panel with immunofixation” test was ordered at a much lower frequency. Detailed review
of LIS volume reports for the two tests did not demonstrate any clusters of providers,
locations, or other clues to the basis for the increased ordering of serum immunofixation.
Fig. 2 Serum immunofixation monthly volumes. The arrow indicates where the electronic health
record (EHR) order for serum protein electrophoresis (SPEP) with immunofixation was
removed from the EHR menu.
Thus, to provide further insight to how the two tests were being ordered, we reviewed
our EHR orders database. In contrast to the RBC folate example above, review of the
EHR orders database for these tests indicated that most “SPEP with immunofixation”
tests were not originating from order sets or provider favorites. Rather, most SPEP
with immunofixation orders were selected following a search of the full outpatient
test menu (72% of total SPEP with immunofixation orders, data not shown). This information
that was obtained from the EHR orders report and not available in the LIS was important
in planning our intervention.
In many situations, limiting EHR options is an efficient approach to managing test
utilization.[23] In this case, as determined by the EHR data, a key decision point occurred when
clinicians performed a search of the outpatient test menu and they were presented
with two options “SPEP panel” and “SPEP panel with immunofixation.” At this point
in the ordering process, many users appear to have selected the choice that appeared
to be more comprehensive (“SPEP panel with immunofixation”), not appreciating that
the standard “SPEP panel” has an associated reflex protocol that would order the immunofixation
when indicated.
After conferring with providers, we confirmed that we could eliminate the SPEP with
immunofixation test from the EHR menu. One consideration with removal of a test from
the menu is that providers may have added that test to their personal laboratory preference
list (their “favorites”). With our EHR orders database, we also noted that a significant
number of the orders originated from provider personal preference lists (22% of total
SPEP with immunofixation orders, data not shown). To transition these users to using
the standard SPEP panel, we requested a change to over 50 personal preference lists
involving substitution of the more desirable SPEP panel for the less desirable test,
SPEP panel with immunofixation. Following these interventions, the serum immunofixation
order frequency returned to its prior baseline ([Fig. 2]).
Miscellaneous Test Monitoring
The EHR laboratory testing menu for our hospital contains over 1,500 individual tests.
These tests were selected on the basis of test volumes, clinical assessment of their
utility, and certain other workflow considerations. However, as a matter of practicality,
infrequently needed tests, including many esoteric and highly specialized genetic
and microbiologic tests used to diagnose rare diseases, were excluded from the test
menu. Indeed, there were likely several thousand tests that we could have put on our
test menu, most of which would be rarely (e.g., several times per year or less) if
ever ordered. We chose not to include these tests on the EHR testing menu for several
reasons. First, many of these tests should not be ordered except in highly selected
conditions and adding them to the general test menu could lead to their use in situations
where they are not indicated, resulting in inappropriate utilization and increased
cost. In addition, adding each test to the EHR menu requires mapping, testing, and
maintenance to ensure the orders can be electronically interfaced to the LIS and that
the results will be faithfully transmitted from the LIS to the EHR.
There are several options for managing miscellaneous test requests in EHR implementations.
One solution is to not permit miscellaneous test ordering in the EHR. This option
generally requires providers to fill out a paper requisition form when requesting
miscellaneous tests. While paper requisitions could presumably limit the number of
test orders by making it inconvenient to order miscellaneous tests, this option presents
several problems. First, paper requisitions make it difficult to track the orders
being placed and do not permit the EHR and LIS to produce barcoded labels for the
orders to facilitate “closed loop” order to result tracking. In addition, paper requisitions
also introduce multiple opportunities for transcription error and add labor to each
order as these test requests must be ordered in the LIS upon reaching the laboratory.
To prevent the challenges associated with paper requisitions, we chose to offer a
“Miscellaneous laboratory test” order in our EHR that can be utilized to place an
electronic request for “write-in” tests. When ordering the Miscellaneous test request,
the provider is prompted to provide the name of the tests they are ordering as well
as any other details they may be aware of such as the tube type or the preferred performing
laboratory. Having the “Miscellaneous laboratory test” as an orderable test within
our EHR menu permits an electronic order to be sent to the LIS and a barcoded label
to be generated to facilitate sample collection and tracking.
The use of an EHR-orderable “Miscellaneous test” strategy requires a commitment to
monitoring these tests. If used to circumvent the standard test menu, the inappropriate
use of the “Miscellaneous test” order can lead to incorrect tube types, insufficient
specimen, and delays in specimen processing. We determined that the ability to track
these orders was essential to our continued efforts to optimize test utilization.
Having an electronic order permits the utilization of our combined EHR orders and
LIS results database to analyze each miscellaneous order request and determine if
it is appropriate. In our utilization database, we employ the EHR order number to
link the EHR order with the LIS result, as the EHR order number is sent to the LIS
and associated with the test result. At our institution in 2017 approximately 0.4%
of orders placed (40 of 9,500 laboratory orders placed per day) were miscellaneous
test requests.
With the use of the utilization database to analyze both the miscellaneous test orders
being placed as well as the tests that were eventually resulted from these orders,
we have been able to use the miscellaneous test order and its monitoring to continuously
improve laboratory quality. Using the EHR laboratory orders database, we analyze the
context of each miscellaneous test order to understand the EHR module that the order
originated, the responses in the EHR to required questions, and examine the other
orders that were ordered during the same ordering session.
The analysis of each miscellaneous test order has several possible outcomes as outlined
in [Table 2]. Outcomes of the analysis include the following: (1) If a clinically useful test
that is not on the current EHR menu is being ordered frequently as a miscellaneous
test, it would likely be of value to add the test to the EHR test menu. (2) Tests
that are uncommon but deemed appropriate in the ordering context of the patient can
remain as miscellaneous tests (i.e., remain off the structured test menu). (3) When
tests that are currently on the existing EHR test menu were ordered via the miscellaneous
order code instead of the structured order, one possibility is that the clinician
was unable to find the test after searching the EHR test menu.[24]
[25] In these cases, reviewing the synonyms defined in the EHR for the test, verifying
the inclusion of the test on relevant facility and specialty preference lists, and/or
educating clinicians would be appropriate next steps. In this scenario, we also analyze
the EHR laboratory orders database for those providers that have correctly found the
structured order of interest to understand if order availability issues, including
the orders' presence or absence on order sets, facility lists, or specialty preference
lists, may contribute to a given provider's ability to find the correct test. (4)
In some circumstances, orders are placed for miscellaneous tests that are unhelpful,
outdated, or inappropriate for the clinical context. In this case, an appropriate
action would be to educate clinicians rather than make any changes to the EHR test
menu.
Table 2
Possible outcomes to EHR miscellaneous test monitoring analysis
|
Conclusion of EHR order review
|
Next steps
|
Comment
|
|
High volume test that is not currently available to be ordered in the EHR and will
be used in future
|
Add test to the EHR facility list for future ordering
|
Add to EHR facility list
|
|
“One time” or high cost esoteric test not currently available to order in the EHR
|
Appropriate as miscellaneous. Do not add to EHR facility list
|
No further action needed
|
|
Test already built in the EHR and on the current EHR menu
|
Review EHR synonyms, test display name, order sets, and preference lists to ensure
test can be easily found with search and that test is on appropriate lists and order
sets. Make EHR changes as needed
|
Educate clinician if recurrent issue and no EHR systematic issue is found
|
|
Test not on menu and is not appropriate for clinical context
|
No change to EHR menu
|
Educate clinician
|
Abbreviation: EHR, electronic health record.
Thus, by capturing detailed information in our EHR database regarding miscellaneous
test requests we have been able to continuously adapt our test menu according to the
needs of our clinicians as well as monitor the appropriateness of the orders being
placed. Importantly, much of the key information that is used to make decisions during
the analysis of miscellaneous test orders is derived solely from the EHR orders database
and is not found in the LIS.
Discussion
As the cost of health care continues to escalate, the need for effective utilization
management programs becomes ever more important.[26] The increasingly ubiquitous role of the EHR in facilitating electronic laboratory
test ordering has made an understanding of EHR workflows and ordering essential to
clinical laboratory test utilization management programs. In addition, while the EHR
contains literally thousands of data points on every patient, there is a relatively
constrained subset of data that is useful for understanding the details and context
of an individual laboratory order. In this report, we identified a subset of information
that could be readily extracted from the EHR and then examined whether this information
provides value to our laboratory utilization management program. We demonstrate that
knowledge of the EHR details of the order provides valuable insights into the test
ordering process and may be useful in the formulation of EHR-based solutions to a
given utilization issue. We recommend that all laboratories develop and maintain near
real-time data from the EHR to assist with utilization management initiatives.
The approach described here should be applicable to most health care settings where
the EHR is a dominant method for laboratory test ordering. With the increased use
of the EHR throughout health care systems, more and more systems have now implemented
electronic laboratory ordering within the EHR.[9] In addition, most EHR systems have been designed with reporting needs considered
and many types of reports can typically be generated from the EHR. It is important
that these reports be generated in a near real-time manner (e.g., next day) since
laboratory utilization patterns can change rapidly. This may be especially the case
in the event of an EHR error in mapping or menu updating that could inadvertently
have a significant and immediate impact on laboratory test ordering.
The aggregation and importing of daily EHR reports into a queryable EHR orders database
offers numerous advantages compared with the individual EHR reports alone. First,
it allows custom queries to be written whenever a utilization issue arises and to
be able to run these queries across any time period of interest to observe trends
and trajectories of testing. In addition, the EHR orders database rapidly generates
a historical record of EHR activity that can serve as a baseline for utilization monitoring
efforts. This baseline allows the determination of volume trends and the ability to
observe shifts in the process of how tests are ordered.
Having these data accessible to a wide range of individuals limits the number of custom
EHR reports that need to be requested, decreasing the demands on resource-constrained
EHR reporting teams and reducing the time required for analysis. Thus, in addition
to data extraction and storage processes it is also necessary for organizations to
develop user-friendly interfaces for laboratory staff to visualize and interact with
the data to facilitate end-user engagement. This interaction can take a variety of
forms, including direct access to the database via ODBC connections, visualizations
of the database built into programs such as spreadsheet software or programming languages
like R or Python, or even dashboards directly accessing the EHR orders database. The
means of user interaction with the database depends on the needs and experience of
the end user, and therefore should be customized to suit the needs of each individual
organization. We have created our database in such a way that it can be securely and
directly accessed by data scientists via direct connections, but also have developed
simpler interfaces, including Microsoft Access reports, for laboratory directors and
other laboratory staff to access the information for utilization and quality efforts
such as this study.
Our costs for the EHR database creation were modest, as we simply added a single table
to an existing database and modified an existing procedure to extract and store the
EHR report data in the database. As opposed to highly clinical EHR data, such as medications
and problem list information, which tend to be complex, the EHR orders data does not
require high levels of processing or analysis to make the data useful for utilization
management. We primarily use simple Microsoft Access-based queries via the EHR reporting
database for our utilization initiatives.
Awareness of the EHR details of test ordering, including the frequency, involved providers,
and context, allows the clinical laboratory to identify trends in test usage, consider
additions and removals to the test menu, consider modifications to existing order
sets, and identify targets for more advanced decision support. In our experience,
certain EHR fields have been demonstrated to be useful to include in reports used
for test utilization monitoring. These fields fall into three major groupings: patient
data, provider data, and order data. [Table 1] provides a listing of some of the key fields included in our EHR orders report along
with some comments regarding the individual fields. It further delineates which of
the EHR order report fields could potentially be obtained from a LIS report.
The EHR provider data elements that have proven to be useful for laboratory utilization
efforts include the entire team who may be involved in the order. A limitation of
LIS provider data are that typically only a single provider is electronically sent
to the LIS in the electronic ordering message. In a training environment, however,
there can be several providers involved in the decision to order a laboratory test
(e.g., attending physician, fellow, resident, nurse practitioner, medical assistant,
etc.). Knowledge of the relevant provider is essential to targeting solutions to many
utilization and quality issues. Our EHR orders report captures both the ordering provider
as well as the provider responsible for follow-up and billing.
The details of the order that have been shown to be especially useful to laboratory
utilization efforts include the ordering location and “order origination” information.
In particular, the details of order origination, including the EHR module, encounter
type, preference list, and order set, are not typically sent to the LIS and are thus
unavailable to laboratories when doing utilization analysis based on LIS reports alone.
In the EHR there are many potential modules and pathways that a given laboratory order
can originate from, including orders selected after searching the full facility list
of tests, orders selected from personal preference lists, orders from departmental
preference lists, and orders selected from within an order set. For each laboratory
order, the specific originating location is specified in our EHR orders database.
Knowledge of the origin of the order within the EHR provides our utilization management
team with a clear target for interventions that cannot be determined from simple LIS
volume data. In many cases based on this EHR orders data, we have dramatically improved
utilization practices by editing a single order set or departmental preference list.
Without this knowledge of where the inappropriate orders were originating, we would
not have been able to customize our solutions to fit the issue or would have had to
spend time reviewing EHR patient charts to attempt to determine the context of the
orders.
The inclusion of additional EHR information in the database, including patient medications,
problem lists, and other clinical data that may impact test ordering and resulting,
is a future direction of our work. Ready access to these additional EHR data elements
may improve the ability of the laboratory to perform institution-specific reflex protocols,
provide clinically relevant interpretive comments, and improve the efficiency and
quality of laboratory interpretive services.
Conclusion
Herein, we have observed the utility of an EHR orders database to provide insight
into a variety of utilization challenges. EHR orders data, particularly when combined
with LIS data, can provide novel insights and identify ordering patterns by clinician,
department, EHR application module, EHR preference list, EHR order set, patient encounter
type, and patient diagnosis. We demonstrate that incorporating this information into
a database framework facilitates analysis and utilization optimization.
Clinical Relevance Statement
Clinical Relevance Statement
This article demonstrates the value of aggregated electronic health record (EHR) data
to improve a laboratory test utilization program. While the EHR contains thousands
of distinct data fields, a small subset of EHR laboratory orders data was identified
in the study and shown to be valuable for test utilization initiatives. As the EHR
increasingly becomes the primary means of laboratory ordering, the context and ordering
pathways within the EHR will become increasingly important for health care systems
to be able to analyze and monitor.
Multiple Choice Questions
Multiple Choice Questions
-
Which of these data fields are often lacking in laboratory information system reports?
Correct Answer: The correct answer is option c. The laboratory order number, test name, and test
cost are generally associated with the test code in the laboratory information system
(LIS). The order origination information, such as which order set the order originated,
are not sent to the LIS. Order origination information is valuable information for
utilization management programs to be able to target interventions within the EHR.
-
What is the value of a laboratory order EHR database versus individual daily EHR reports
of laboratory ordering activity?
-
Lower cost to create and maintain a database than generating reports
-
More data fields can be extracted and stored in a database than in a daily report
-
EHR orders database information cannot easily be combined with other data whereas
reports can be easily combined
-
Longitudinal queries and analysis can be performed on an orders database to assess
for trends.
Correct Answer: The correct answer is option d. Although database construction and maintenance may
incur costs not associated with creating simple reports, the costs can be modest if
a constrained subset of data is stored. A key advantage of having laboratory order
information in a database is that longitudinal analysis can be performed to identify
trends and ordering patterns. Moreover, databases typically provide a simple interface
to link data stored in one database with other key data sources. In the case of laboratory
utilization management, we routinely associate EHR orders data with laboratory results
data to provide a useful window into our combined EHR and laboratory operations.
