Recommendations for the Submission of LOINC® Codes in Regulatory Applications to the U.S. Food and Drug Administration: Guidance for Industry

Administrative guidance, not enacted legislation. This captured publisher edition preserves the listed document and attachments; later developments are not independently consolidated.

Recommendations for the Submission of
LOINC® Codes in Regulatory Applications to
the U.S. Food and Drug Administration
LOINC Working Group:
U.S. Food and Drug Administration (FDA), U.S. National
Institutes of Health (NIH), Clinical Data Interchange Standards
Consortium (CDISC), and Regenstrief Institute
November 2017

i
Table of Contents
1 	LOINC Working Group (WG) Members and Affiliations ........................................................ 1
2 	Background......................................................................................................................... 1
2.1 	Purpose of this Recommendation Document ............................................................... 1
2.2 	Introduction .................................................................................................................. 2
2.3 	CDISC Member Survey Regarding LOINC Readiness ................................................. 3
2.4 	Pre-Requirement Submissions of LOINC Codes to FDA .............................................. 3
3 	Working Group Considerations ........................................................................................... 3
4 	Working Group Recommendations ..................................................................................... 6
5 	Potential Future Phases ...................................................................................................... 8
6 	References ......................................................................................................................... 8
7 	Appendices ......................................................................................................................... 9
7.1 	Abbreviations ............................................................................................................... 9
7.2 	Helpful Links ................................................................................................................ 9
7.3 	Educational Resources and Events ............................................................................10
7.3.1 	CDISC .................................................................................................................10
7.3.2 	LOINC..................................................................................................................10

1
1 	LOINC Working Group (WG) Members and Affiliations
Member Name 	Professional Affiliation
Swapna Abhyankar, M.D. 	Regenstrief Institute/LOINC
Lauren Becnel, Ph.D. 	CDISC
Boris Brodsky 	FDA
Ron Fitzmartin, Ph.D. 	FDA
Margaret Haber 	NCI
Virginia Hussong 	FDA
Clement McDonald, Ph.D. 	NLM
Erin Muhlbradt, Ph.D. 	NCI
Eileen Navarro, M.D. 	FDA
Vaishali Popat, M.D. 	FDA
Colleen Ratliffe 	FDA
Lilliam Rosario, Ph.D. 	FDA
Jessica Voqui, Pharm.D. 	FDA
Daniel Vreeman, D.P.T. 	Regenstrief Institute/LOINC
Alan Williams, Ph.D. 	FDA
Lawrence Wright, Ph.D. 	NCI
2 	Background
2.1 	Purpose of this Recommendation Document
Regulated clinical research is carried out and supported across multiple sectors including
academic medical centers, biopharmaceutical companies, biotechnology companies, clinical
research organizations, federal agencies, information technology vendors and others. The FDA
will soon begin requiring the inclusion of a medical laboratory and observation terminology coding
system, Logical Observation Identifiers Names and Codes (LOINC®). While this requirement will
help align semantics between healthcare and regulated clinical research, readiness for this
requirement and understanding of LOINC can vary dramatically across these sectors.
In January 2017, the FDA convened a Committee, including FDA, NIH, CDISC, and Regenstrief
Institute, to evaluate the use of terminology standards, with a particular focus on LOINC. 	The
Committee formed a LOINC Working Group to: (1) review the relationship between the CDISC
laboratory codes and the Regenstrief LOINC laboratory codes, (2) address the concerns raised
by the CDISC community in the 2017 survey, and (3) develop recommendations for the
submission of LOINC within the CDISC data exchange format in applications to the FDA.

2
This document, prepared by CDISC with inputs from the Working Group members, serves to
provide the community with information about LOINC and to make recommendations to the FDA
on the initial required scope.
2.2 	Introduction
Standards for biomedical research data exchange and terminology support data retrieval, system
interoperability and research reproducibility by providing common structures and shared meaning
for the collection, aggregation, analysis and exchange of clinical and non-clinical research data
between or among electronic systems. Per the FDA website’s Study Data Resource page,
“Having standard, uniform study data enables FDA scientists to explore many new research
questions by combining data from multiple studies. Data standards also help FDA receive,
process, review, and archive submissions more efficiently and effectively.”
The FDA Data Standards Catalog explicitly states which standards the FDA’s Center for Biologics
Evaluation and Research (CBER), and, Center for Drug Evaluation and Research (CDER), accept
and/or require for regulatory submissions, alongside effective dates for each data exchange or
terminology standard. Standards referenced in the FDA catalog include the CDISC Study Data
Tabulation Model (SDTM) data exchange/tabulation standard and LOINC, a terminology standard
maintained by the Regenstrief Institute. LOINC is a freely available international standard for
identifying health measurements, observations, and documents that is ubiquitous in health data
systems worldwide and is an essential ingredient of system interoperability. Presently, LOINC is
utilized in more than 172 countries and has been adopted as a national standard in nearly 30
countries, including the United States.
The FDA Data Standards Catalog currently specifies CDISC data exchange standards for the
submission of clinical study datasets to CBER and CDER. It also includes the requirement for
LOINC codes to populate the “LBLOINC” variable in the laboratory (LB) domain specification
when study data are submitted to the FDA using CDISC data exchange standards. This
requirement applies to new drug applications (NDAs), abbreviated new drug application for
generics 	(ANDAs), 	biologics 	license 	applications 	(BLAs), 	and 	e-submissions 	of 	some
investigational new drugs (INDs) in accordance with section II.A of the FDA guidance document
“Providing Regulatory Submissions in Electronic Format – Standardized Study Data” 1
. The latest
version of the FDA Data Standards Catalog at the time of writing this recommendations document
is v4.5.2 (04-13-2017). The requirement for LOINC codes begins for studies that start after March
15, 2020 for NDAs, ANDAs and certain BLAs, and on March 15, 2021 for certain INDs.

3
2.3 	CDISC Member Survey Regarding LOINC Readiness
When the FDA requirement for the use of LOINC was published, some CDISC members
expressed concern about inclusion of LOINC in electronic submissions. CDISC launched a
member survey in the first quarter of 2017 to better understand these concerns, their prevalence
and member readiness. This survey is informational only, as it does not necessarily represent the
full breadth of global clinical researchers’ knowledge and use of LOINC.
Data were collected from 180 respondents (Figure 1),
which showed that 16% were currently collecting
LOINC codes in at least some studies, whereas 84%
were not. Nearly 10% of respondents reported never
having heard of LOINC. For those that were collecting
LOINC codes, they reported doing so for 15-50% of
their total volume of trials and noted that LOINC codes
were 	useful, 	when they were 	received, 	as they
provided a more precise description of the collected
assay. They also mentioned that local and academic
laboratories were unable to provide LOINC codes
in many cases.
Among the 85% of respondents who did not collect
LOINC codes, some indicated that they had been
requesting LOINC codes from central laboratories,
but the codes had not been provided. Other concerns included realm specificity for sponsors
engaged in global trials, availability of LOINC codes for complex new tests and from smaller
laboratories, and inconsistency among different laboratories with regard to the application of
LOINC codes. In general, the respondents were concerned about the lack of clear guidance with
regard to FDA requirements for use of LOINC and the potential for each organization
implementing bespoke solutions that may not truly meet regulatory need.
The FDA/Clinical Data Interchange Standards Consortium (CDISC)/United States National
Institutes of Health (NIH)/Regenstrief Institute LOINC Working Group (“Working Group”) was
created to address some of the respondents’ concerns and provide specific recommendations for
the use of LOINC in clinical trial data submissions to the FDA.
2.4 	Pre-Requirement Submissions of LOINC Codes to FDA
Currently, FDA supports, and in the near future will require, the use of LOINC codes in clinical
trials. Until required, the Working Group encourages sponsors to voluntarily follow these
recommendations and any others that the FDA may issue when submitting laboratory data to
FDA.
3 	Working Group Considerations
It is important to recognize that LOINC includes two major divisions, Laboratory and Clinical, and
that the FDA’s current support of LOINC for submitting clinical trial data to CBER and CDER, as
identified in the FDA Data Standards Catalog, is limited to the laboratory division. The LOINC
Figure 1. Organizations engaged in
regulated research that are currently
collecting LOINC codes. N = 180.

4
Laboratory division generally aligns with the CDISC SDTM LB (laboratory) domain.
One of the primary differences between LOINC and the CDISC LB domain is that LOINC terms
are pre-coordinated and CDISC LB concepts are post-coordinated. 	According to Stevens and
Sattler, “A pre-coordinated [set of information or terminologies] has all the terms and relationships
between them needed by an application; it is static; ‘what you see is what you get’. A post-
coordinated [set of information or terminologies] has the building blocks for the terms needed in
an application such that they can be built as required, and so that their relationship can be
determined as required; it is dynamic; the ontology is much more than ‘what you see’, since you
can compose new expressions from the given building blocks. 3
” This means that in LOINC, a
single concept represents the combination of several different aspects of a laboratory test, such
as the analyte being measured, the timing (e.g., point in time versus 24-hour collection), and the
specimen type, while the CDISC LB domain contains three separate concepts for the analyte,
timing, and specimen type.
Pre-coordination and post-coordination both offer benefits. Pre-coordinated data support more
rapid data entry and reuse. For example, it is far easier in busy clinical environments to enter a
single 	piece
of 	data 	that
contains
multiple data
elements,
rather 	than
spend
additional
time entering
each 	data
element
individually.
LOINC 	fully
specified
names 	are
pre-
coordinated
using 	a 	six
part 	model
that
distinguishes
among 	tests
with clinically
different
meanings.
The six parts,
also 	called
Dimensions, are the Component, Property, Timing, System, Scale, and Method. All axes other
Figure 2. Comparing pre-coordinated LOINC and post-coordinated CDISC
LB Variables for a laboratory test measuring Albumin in serum or plasma. Two
pre- coordinated LOINC codes (blue boxes, left) are shown. Each LOINC code is a
single entity comprised of six primary Parts. For Post-Coordinated CDISC LB
variables (green, right) and associated controlled terminologies (grey, right) for the
albumin test are shown with “C-code” identifiers from the National Cancer
Institute’s Enterprise Vocabulary Services.

5
than the Method have a value for every LOINC term. The Method may or may not have a value.
LOINC’s pre-coordinated names match the level of specificity of common laboratory test ordering
and reporting paradigms where there are separate fields for indicating the test name, the test
result value, units of measure, reference range, etc. As of version 2.61 (June 2017), the LOINC
release now contains an artifact linking the pre-coordinated LOINC terms to the set of LOINC
parts of which it is composed 3
. 	Having these codes and relationships facilitates secondary
aggregation and analysis of primary data coded with LOINC.
On the other hand, CDISC is post-coordinated for ease of granular data analysis. Each aspect of
the test (e.g., analyte, specimen, etc.) has a different identifier, which makes it easier to aggregate
and analyze data based on the specific combination of concepts that are relevant to the research
question. See Figure 2 for a comparison of pre-coordination and post-coordination, where multiple
CDISC SDTM LB domain variables map to a single LOINC code. Related LOINC codes may
require representation by different variables and CDISC controlled terms. Figure 2 includes two
LOINC codes: one with data in all six major parts, 2862-7, and a less specified, “method-less”
LOINC code, 1751-7. The related code, 2862-7, specifies albumin measurements were made by
electrophoresis, which maps to CDISC’s LB.METHOD variable. For the method-less LOINC,
LB.METHOD cannot be mapped. Similarly, some LOINCs do not specify a specific System, which
maps to LB.SPEC, and instead indicate that the System is ‘XXX’ to denote that it can be one of
several locations or specimen types. End users should utilize the most specific LOINC possible,
where known information matches the LOINC part values for that code, without over specifying
by selecting a LOINC code that has one or more part values that are more granular than what is
known for the result data (Figure 3). End users should not over specify LOINCs. If the method is
not known, for example for a laboratory test as shown in Figure 3, mapping to LOINC 2862-7
would be over specifying because it includes a defined method, and LOINC 1751-7 would be a
more appropriate match because it is method-less.
A mapping between the six major LOINC parts and CDISC LB variables is provided in Table 1.
Harmonizing the CDISC lab domain variable values with the LOINC Part values represents a
substantial volume of work. In the absence of clear use cases for such harmonization, the Working
Figure 3. Selecting the most
specific LOINC available
without over specifying. For an
albumin test from either serum or
plasma where the analyte was
measured by the electrophoresis
method and is reported in g/dL,
LOINC 2862-7 is appropriate as
indicated by a green arrow.

6
Group declared it to be out of scope for this initial set of recommendations, but it may be
considered in future work.
Table 1. A comparison of LOINC Parts to CDISC Laboratory (LB) Domain Variables
LOINC Part
(entire column represents a single LOINC
code)
CDISC Standard Equivalent
(each row represents a single CDISC
code)
Component 	LBTEST/CD + other variables
Property 	-
Time 	MULTIPLE: Various Timing Variables
System 	SPEC+LOC
Scale 	-
Method 	METHOD + other variables
4 	Working Group Recommendations
In general, the Working Group recommends that where a valid LOINC code exists for a given lab
value, the community should provide the code in LBLOINC. Many laboratories have mapped their
internal codes to LOINC codes and will make such mappings available in electronic messages or
as standalone mapping files. Increasingly, in vitro Diagnostic (IVD) vendors are making mappings 4
of their internal test codes to LOINC codes available for their products, which will increase the
efficiency and consistency of identifying appropriate LOINC codes. If such LOINC codes are not
available from the laboratory, the Working Group recommends that sponsors not attempt to derive
a mapping to LOINC as they create the SDTM files. Lab test names are notoriously ambiguous,
and making assumptions in the mapping process without confirming with the test performer can
lead to serious data quality issues.
Based on CDISC 2017 LOINC survey data, the Working Group also feels it is important that the
requirement for LOINC code submissions start with a subset of common diagnostic lab tests that
are also used within clinical research in order to ease researchers’ adoption of the FDA
requirement. To create this list, the CDISC Lab Working Group assessed the most commonly
utilized lab tests in clinical trials as defined by three major central laboratories, the LOINC 2,000+
and data summarizing LOINC codes that have been received by the FDA. These LOINC codes
are predicted to represent ~80-90% of the routine care labs also utilized within clinical trials. A
mapping of this subset of common LOINC codes to specific combinations of CDISC LB domain
variables and controlled terminologies has been created by the CDISC Lab Working Group. At
the time of writing this recommendation document, this mapping document was undergoing

7
review by the CDISC Lab Working Group and key members from the CDISC Controlled
Terminologies 	Team 	and 	this 	LOINC 	Working 	Group. 	The 	mapping 	document 	and 	an
implementation guide will be made available from the CDISC Lab Working Group on the CDISC
and Regenstrief websites.
The Working Group specifically recommends that:
• 	Regardless of whether a LOINC code is available or not for a given lab test record,
submitters must still submit lab data in CDISC format, including the use of CDISC
controlled terminology for lab variables.
• 	Sponsors should include LOINC codes for any test where a valid LOINC code exists for
new trials’ electronic submissions per the FDA Data Standards Catalog requirement dates.
o 	LOINC codes that contain specified methods should be utilized over similar LOINC
codes that do not have a method specified in the Method part as appropriate. See
Figures 2 and 3 for detail.
o 	When the specimen type for a lab test is known and/or where the reference range
differs by specimen types, a LOINC code with a specific System (such as Ser/Plas,
Urine, CSF, etc.) should be used instead of a similar LOINC code that has “XXX”
as the System.
o 	Sponsors may submit LOINC codes for lab tests outside of the common subset
described above when it becomes available. The Working Group recommends that
sponsors be encouraged to do so in order to allow the community and regulatory
officials to adapt to the new requirement. When provided, LOINC codes for other
lab tests should be accepted by FDA.
o 	The most specific LOINC code should always be used without over-specifying
details that are not known by the sponsor, such as method or scale.
o 	LOINC codes with a status of deprecated or discouraged should not be submitted.
 	Appropriate status for a LOINC code is defined in the LOINC Manual
Section 11.2 Classification of LOINC Term Status.
• 	For any lab test where a LOINC code is not submitted, the reason for its omission should
be noted in the clinical Study Data Reviewers Guide.
o 	The Working Group proposes that a starter set of reasons be predetermined
(perhaps as CDISC terms) for consistency of reporting, including:
 	Performing laboratory unable to determine if appropriate LOINC code
exists
 	Performing laboratory indicates that no appropriate LOINC code currently
exists
• 	LOINC code specified in LB.LOINC applies to LB.ORRES (original result) rather than to
LB.STRES(C/N) 	(standardized 	result). 	Such 	an 	approach 	is 	consistent 	with 	the
recommendation that LB.LOINC should be taken from original laboratory result (i.e., the
data transfer itself), and not be derived/converted into another form.
• 	LOINC codes should be submitted for human clinical studies only.

8
5 	Potential Future Phases
The Working Group may reconvene in the future to make further recommendations. The Working
Group considered topics such as the inclusion of additional LOINC codes for tests that are not
routinely performed, LOINC codes for labs performed on non-human subjects, and codes for more
clinically-focused variables. Though it is highly desirable for regulators to easily and routinely
analyze standard of care lab tests that have assigned LOINC codes, challenges exist to this
approach for research lab tests, most of which will not have an assigned LOINC code.
Regarding 	the 	pre-coordinated 	LOINC 	model 	and 	post-coordinated 	CDISC 	model, 	a
harmonization of CDISC LB domains with LOINC represents a substantial volume of work. In the
absence of clear use cases for such harmonization, the Working Group determined that it was
out of scope for this initial set of recommendations, but it may be considered in future work.
6 	References
1. 	Providing Regulatory Submissions In Electronic Format – Standardized Study Data.
United States Food and Drug Administration
https://www.fda.gov/downloads/Drugs/Guidances/UCM292334.pdf. Accessed May 15,
2017.
2. 	Robert Stevens and Uli Sattler (2013) Post-coordination: Making things up as you go
along. Ontogenesis. http://ontogenesis.knowledgeblog.org/1305. Accessed June 23,
2017.
3. 	Dan 	Vreeman. 	(2017) 	LOINC 	Version 2.61 and 	RELMA 	Version 	6.20 Available.
https://loinc.org/news/loinc-version-2-61-and-relma-version-6-20-available/. 	Accessed
June 23, 2017.
4. 	IICC News. (2017) IVD Industry Connectivety announces LIVC specification for digital
publication 	of 	LOINC 	to 	vendor 	IVD 	test 	results 	in 	clinical 	laboratories.
http://ivdconnectivity.org/iicc_announces_livd_specification/. Accessed June 26, 2017.

9
7 	Appendices
7.1 	Abbreviations
Abbreviation 	Description
ANDA 	Abbreviated New Drug Application
BLA 	Biologics License Applications
CBER 	Center for Biologics Evaluation and Research
CDER 	Center for Drug Evaluation and Research
CDISC 	Clinical Data Interchange Standards Consortium
CRO 	Contract Research Organization
FDA 	United States Food and Drug Administration
IND 	Investigational New Drug
LB 	Laboratory
LOINC 	Logical Observation Identifiers Names and Codes
MB 	Microbiology
NCI 	United States National Cancer Institute
NDA 	New Drug Application
NIH 	United States National Institutes of Health
NLM 	United States National Library of Medicine
SDRG 	Study Data Reviewer’s Guide
SEND 	Standard for Exchange of Non- Clinical Data
SDTM 	Study Data Tabulation Model
7.2 	Helpful Links
CDISC Homepage: http://www.cdisc.org
CDISC SDTM webpage: https://www.cdisc.org/standards/foundational/sdtm
CDISC SDTM Implementation Guide webpage:
https://www.cdisc.org/standards/foundational/sdtmig
LOINC® Homepage: https://loinc.org/
FDA Study Data Standards Resources, including the Data Standards Catalog:
https://www.fda.gov/ForIndustry/DataStandards/StudyDataStandards/default.htm

10
7.3 	Educational Resources and Events
7.3.1 	CDISC
CDISC – SDTM002A – SDTM Basics, CDISC – SDTM003A – Basics of the SDTM
Implementation Guide, and CDISC – SDTM010A – Findings Domains:
https://secure.trainingcampus.net/uas/modules/trees/store/wcatalog.aspx?cat=1029&ci=
10&vtype=2&pi=0
CDISC Public Courses: https://www.cdisc.org/education/public-courses
SDTM Standard in PDF: https://www.cdisc.org/standards/foundational/sdtm
SDTM IG in PDF: https://www.cdisc.org/standards/foundational/sdtmig
SDTM in Other Machine-Readable Formats (CDISC Members Only and Individual
Academic Researchers Upon Request): https://www.cdisc.org/members-only/share-
exports
7.3.2 	LOINC
LOINC Getting Started Guide: https://loinc.org/get-started/ LOINC overview video:
https://youtu.be/OJYQ2xHNIUU LOINC and RELMA Workshops:
https://loinc.org/meetings
LOINC Essentials eBook, which covers the basics of LOINC and specific mapping
strategies: https://danielvreeman.com/loinc-essentials/