| |
| J Am Med Inform Assoc. 1998 Mar–Apr; 5(2): 203–213. | PMCID: PMC61291 |
Copyright © 1998, American Medical Informatics Association Evaluation of a “Lexically Assign, Logically Refine” Strategy
for Semi-automated Integration of Overlapping Terminologies Robert H. Dolin, MD, Stanley M. Huff, MD, Roberto A. Rocha, MD, PhD, Kent A. Spackman, MD, PhD, and Keith E. Campbell, MD, PhD Affiliations of the authors: Kaiser Permanente, Southern California, La
Palma, California (RHD); Intermountain Health Care, Salt Lake City, Utah
(SMH); Departamento de Cirurgia, Universidade Federal do
Paranà,
Paranà, Brazil (RAR); Department of Pathology,
Oregon Health Sciences University, Portland, Oregon (KAS); and the Section on
Medical Informatics, Stanford University School of Medicine, Stanford,
California (KEC). Received July 24, 1997; Accepted November 12, 1997. |
Abstract Abstract Objective: To evaluate a “lexically assign, logically
refine” (LALR) strategy for merging overlapping healthcare
terminologies. This strategy combines description logic classification with
lexical techniques that propose initial term definitions. The lexically
suggested initial definitions are manually refined by domain experts to yield
description logic definitions for each term in the overlapping terminologies
of interest. Logic-based techniques are then used to merge defined terms. Methods: A LALR strategy was applied to 7,763 LOINC and 2,050 SNOMED
procedure terms using a common set of defining relationships taken from the
LOINC data model. Candidate value restrictions were derived by lexically
comparing the procedure's name with other terms contained in the reference
SNOMED topography, living organism, function, and chemical axes. These
candidate restrictions were reviewed by a domain expert, transformed into
terminologic definitions for each of the terms, and then algorithmically
classified. Results: The authors successfully defined 5,724 (73%) LOINC and
1,151 (56%) SNOMED procedure terms using a LALR strategy. Algorithmic
classification of the defined concepts resulted in an organization mirroring
that of the reference hierarchies. The classification techniques appropriately
placed more detailed LOINC terms underneath the corresponding SNOMED terms,
thus forming a complementary relationship between the LOINC and SNOMED
terms. Discussion: LALR is a successful strategy for merging overlapping
terminologies in a test case where both terminologies can be defined using the
same defining relationships, and where value restrictions can be drawn from a
single reference hierarchy. Those concepts not having lexically suggested
value restrictions frequently indicate gaps in the reference hierarchy. |
Standardized definitions and standardized use of terminology in medical
record systems are prerejquisites for robust informatics applications such as
automated decision support and outcomes analysis. The proper way to attain
such standards and to formally define healthcare terminology is an area of
active discussion and
research. 1,2,3,4,5,6
Often, proposed approaches seem to be polarized into two paradigms: lexically
based, which relies on analysis of morphemic patterns within the terms being
defined to derive meaning, and logically based, which relies on axiomatic
definition of concepts and subsequent classification on those definitions to
derive meaning. We believe that pragmatic solutions to our terminology problems may be
found in hybrid solutions that leverage the best characteristics of each
approach. Here we describe an evaluation of a “lexically assign,
logically refine” (LALR) strategy that combines the efficiency of
linguistically based approaches with the precise semantics of description
logic. We leverage lexical algorithms and language's inherent structure to
propose relationships for terms in a terminology. These proposed definitions
are then reviewed and refined by domain experts, and subsequently loaded into
a description logic classifier. The LALR strategy allows us to efficiently and
formally define terms to the extent pragmatically possible with the
concomitant benefits of decreased ambiguity and increased precision available
to represent medical observations, diagnoses, and patient management
plans. Using our hybrid approach, we seek to demonstrate the practicality of a
LALR strategy that we believe can scale to answer the terminologic challenges
posed by the Computer-Based Patient Record
Institute7 and the
Institute of
Medicine.8 Although
our test case focuses on the semi-automated merging of overlapping
terminologies, we believe that the LALR approach can be generally applied to a
wide variety of terminologic challenges. We selected the SNOMED (Systematized Nomenclature of Human and Veterinary
Medicine) laboratory
procedures9 and the
LOINC (Logical Observation Identifier Names and Codes) laboratory
procedures10 as our
test case because the terminologies are individually important to our
respective organizations, and the two terminologies have complementary
strengths. SNOMED III is the foundation of Kaiser Permanente's Convergent Medical
Terminology
project,11 and has
been shown to have broad content coverage in many clinical
domains.12,13
Despite SNOMED's broad coverage of medical concepts, it is missing many of the
detailed laboratory terms necessary for standardizing reporting of laboratory
information, although it provides a hierarchical classification of laboratory
procedures. The LOINC database contains finely detailed test result identification
codes suitable for standardized reporting of laboratory information and has
been endorsed by the American Clinical Laboratory Association (ACLA). ACLA
recommends that all members report laboratory results using LOINC codes (ACLA
members account for approximately 70% of the volume of tests performed in the
United States). Additionally, the Health Care Financing Administration is
constructing ICD-10-PCS codes using LOINC
terms.14 Despite LOINC's finely detailed identification codes and consensus
regarding its adoption, LOINC lacks a hierarchical organization of those
identification codes. We chose to test our LALR strategy using LOINC and
SNOMED because we felt that a combination of LOINC and SNOMED could produce a
comprehensive strategy for representing detailed laboratory terms as well as
appropriately classify those terms. |
Background Systematized Nomenclature of Medicine (SNOMED) SNOMED III is a multiaxial hierarchical coding
scheme. 9 Terms are
assigned to one of 11 independent systematized axes: Topography, Morphology,
Function, Living Organisms, Chemicals, Physical Agents, Occupations, Social
Context, Diseases, Procedures, and General Linkage/Modifiers. The code
assigned a term places it into its proper location in a monohierarchy, thereby
conveying contextual information for each term (e.g., Tuberculous pneumonia,
DE-14817, is a Respiratory tuberculosis, DE-14810, which is a Tuberculosis,
DE-14800, which is a Bacterial infectious disease, DE-10000, which is an
infectious disease, DE-00000). SNOMED concepts are further defined through the
use of cross-referencing to related concepts in other axes (e.g., Tuberculous
pneumonia, DE-14817, affects the topographic region Lung, T-28000, has a
morphologic feature of inflammation, M-40000, and an etiologic agent of
Mycobacterium tuberculosis hominis, L-21801). In the July 1996 release of
SNOMED, version 3.3, approximately 25% of the relevant clinical terms included
cross
references. 15Logical Observation Identifier Names and Codes (LOINC) The LOINC database provides a set of universal names and codes for
identifying laboratory test
results. 10,16
Initially developed in mid-1995, the August 1996 version contains
approximately 8,200 codes. Each result is defined as having a measured
component or analyte (e.g., potassium), a property being measured (e.g.,
mass), whether the measurement is a momentary observation at a point in time
or an observation integrated over time (e.g., 24-hour collection), the type of
system or sample (e.g., serum), the type of scale (e.g., quantitative), and,
where relevant, the method used to produce the result or other observation. In
addition, each concept is placed into a category or class (e.g., Chemistry,
Microbiology). Thus, each LOINC concept may have up to six defining
relationships: Analyte, Property, Time Aspect, System, Scale, and possibly a
Method. A non-defining attribute of a LOINC concept is its Class. The LOINC
data model is illustrated in Figure
1, using Coad and Yourdon's modeling
representation. 17 | Figure 1 The LOINC (Logical Observation Identifier Names and Codes) data model. (See
the Background section for details.) |
Merging and Mapping Techniques Prior efforts at automated mapping and merging of overlapping terminologies
have relied primarily on two different classes of techniques: lexical, which
base their comparisons on morphemic components of the terms themselves, and
logical, which base their comparisons on classification of the axiomatic
definition of terms. Lexical Techniques Lexical algorithms typically perform an initial normalization process on
the terms to be compared in an attempt to abstract away from lexical
(syntactic, morphologic, and orthographic) variation in
terms.2,5,6,18,19
The process may involve breaking each string into its constituent words and
morphemes; lower-casing each word; removing punctuation, stop words, and
duplicate words; and sorting the words in alphabetical order (e.g., the term
“Cell count of synovial fluid with differential count” would
become “cell count differential fluid synovial”). More sophisticated normalization techniques attempt to convert each word to
its canonical uninflected form (e.g., “treats” or
“treated” becomes
“treat”),5,18
expand
abbreviations,5
attempt to correct spelling
errors,20 and
substitute preferred terms for synonyms (e.g., “kidney” becomes
“renal”).2,5,21
Rocha et
al.18,19
lexically compared normalized word fragments (digrams) using a similarity
function. Digram comparisons break each word into fragments of two letters
each for comparison. For example, the word “morphemic” would be
broken into the fragments mo, or, rp, ph, he, em, mi, and ic for comparison
with the digrams from another word. This technique is less sensitive to minor
morphologic variations in words introduced by mis-spellings or other normal
variations in morphology. Digrams are created algorithmically and do not
require any knowledge about word formation rules, and do not rely upon the
existence of affix dictionaries. Following the normalization process, terms are compared using various
techniques such as exact normalized string matching, the Longest Common
Substring
algorithm,22 or a
similarity scoring
approach.18 Many of
these techniques are included in the National Library of Medicine's lexical
variant generation (lvg) tools, which are available to anyone who has signed
the UMLS license
agreement.23 Logic-based Techniques Logic-based techniques are used to formally define concepts and to map and
classify terms based on similarities of their
definitions.24,25,26,27
So, for example, if two distinct and lexically unrelated terms each define a
type of laboratory procedure that measures sodium, the two terms will be
appropriately classified together. There has been a general trend in
terminology research to make more explicit the definition of medical
concepts,4,19,27,28,29
both to disambiguate concepts and to facilitate automated terminology merging
efforts. The Canon group has argued for a medical-concept representation
language containing semantic types for each concept and a network of general
medical concepts in which implicit relations between elements are made
explicit.4 There has been an evolution over time in the formal representation of
medical concepts. An early effort to abstract away from specific terms to term
representations for use in modeling general medical language was developed by
Evans in the MedSORT
project.28 To
represent concepts, MedSORT required both a semantic classification scheme and
rules that determined how elements in the classification scheme might combine.
Masarie et al.25
defined concepts from several vocabularies using a frame-based system, where
the definitions were based on an analysis of the terms to be merged. Cimino et
al.27 have used a
semantic network for defining the concepts in the Columbia-Presbyterian MED
vocabulary. Their network is a notation for representing conceptual entities
and links between them, allowing the storage of factual knowledge that can be
intensional (describing the entities themselves) and extensional (describing
how entities are related to other entities), and includes a classification
hierarchy. Description logic, which defines concepts based on formal logic
theory, was developed to make explicit the semantics of frame-based
systems* and semantic
networks while retaining an emphasis on taxonomic structure as an organizing
principle.30 Campbell et al.3
have previously argued that formal logic can provide a framework for
formalizing the representation of medical concepts. It is well recognized that
there is a tradeoff in expressiveness and tractability of term classification
that has to be considered when using formal
logic.31,32
Rector et al.33
have argued for a more expressive formalism, claiming that worst-case
computational complexity is an inappropriate guide to the choice of formalism.
On the other hand, several description logics have been defined that support
less then full first-order logic, and offer complete and sound algorithms for
the classification of terms defined by the
logic.34,35
Such a subset, used in this study, is supported by K-Rep, a description logic
classification and terminology authoring system developed by researchers at
IBM.36
Expressiveness is deliberately constrained in the K-Rep system so that
algorithmic classification is guaranteed to be sound and complete. While this
constraint may pose a limitation in the definition of certain concepts
requiring, for example, a logical “or” or a logical
“not,” this was not an issue in the definition of the concepts in
this study, all of which could be fully defined. The explicit representation
of medical concepts in a description logic enables the formal determination of
the expressiveness and tractability of a particular representational
syntax. Logic-based classification algorithms typically organize a set of concepts
into a taxonomy based on axiomatic term definitions. As an initial step, the
defining characteristics for the set of concepts to be classified must be
determined. A LOINC concept, as shown in
Figure 1, is defined by its
analyte, measured property, measurement time aspect, specimen, scale, and
method. Cimino et
al.27 defined
laboratory specimens, laboratory tests, and medications based on relationships
taken from the UMLS Semantic
Network.37 Once the
defining relationships have been determined, appropriate value restrictions
for each defining relationship must be assigned. These values can be
determined by a combination of manual assignment, lexical suggestion, or
logical inheritance using the methods described above. Once concepts are defined, logic-based classification algorithms are
employed.27,32,34,35
Cimino et
al.27,38
have provided a description of their classification algorithms, which are used
to deepen a hierarchy and to identify the most appropriate location for a
given term in a given classification hierarchy. Description logic
classification algorithms are discussed in more detail by
Fitting.39 Prior experience with logic-based techniques suggests that these methods
may enhance results achievable with lexical techniques alone. Here we present
our evaluation of a specific strategy for combining logic-based and
linguistic-based techniques: Lexically Assign, Logically Refine (LALR). |
Methods Definitions This section defines the terms “defining relationship” and
“value restriction,” which are used throughout this report to
specify how we are defining healthcare terminology concepts. Defining
relationships (also known as “defining roles”) are those
relationships that link a concept to its defining characteristics. In
Figure 1, for instance, the
defining relationship HAS-ANALYTE indicates that a LOINC Observation is in
part defined by its analyte or measured component. A value restriction (also
known as a “role restriction” or “relationship value”)
is the target value that restricts the domain of a defining relationship. In
Figure 3, for instance, the
concept HLA-A-serotyping_P3-68530 has a value restriction of
HLA-A-Antigen-NOS_F-C4100 for the HAS-ANALYTE defining relationship. This
value restriction in effect states that any concept in the terminology
database subsumed by HLA-A-Antigen-NOS_F-C4100 is a valid analyte for those
concepts subsumed by HLA-A-serotyping-NOS_P3-68530. The value restrictions on
a concept are the logical intersection of the value restrictions stated in the
definition of that concept and in the definitions of the transitive closure of
its parent concepts. | Figure 3Algorithmic classification based on measured component. Given the analyte
hierarchy for the SNOMED Function Axis and the concept definitions shown, the
resulting SNOMED Procedure hierarchy is automatically generated. (Concept
HLA-A1-Measurement_L-4718-3 (more ...) |
Test Data Set Of the 8,182 concepts in the August 1996 version 1.0h LOINC database, 419
(those in LOINC Classes CLIN, MISC, and DRUGDOSE) were excluded due to lack of
overlap with any SNOMED Procedure categories, leaving a total of 7,763 LOINC
concepts for study. 2,317 concepts are present in the SNOMED version 3.3
Procedure Axis, Chapter P3 Laboratory Procedures. Of these, 267 (section P3-0
General Laboratory Procedures and Services, subsection P3-53 Microbial
Identification Kit Methods, and subsection P3-70 Chemistry Methods) were
excluded because they contain concepts not readily adaptable to definition
using the LOINC model, for a total of 2,050 SNOMED concepts for study. Defining Relationships Are Taken from the LOINC Model As noted above and shown in Figure
1, a LOINC concept is defined by its analyte, property, time
aspect, system, scale, and method. From this model, the corresponding defining
relationships became HAS-ANALYTE, HAS-PROPERTY, HAS-TIME-ASPECT, HAS-SYSTEM,
HAS-SCALE, and HAS-METHOD. Value Restrictions Are Drawn from SNOMED All LOINC concepts are assigned an analyte, property, time aspect,
specimen, scale type, and method by the LOINC committee. In a prior study,
these components of the 7,763 LOINC concepts in our test data set were
lexically mapped to SNOMED
concepts, 19 using
the methods described by Rocha et
al. 18 These SNOMED
mappings formed the value restrictions for the LOINC concepts in this
study. Lexical techniques were used to suggest value restrictions for the SNOMED
test data set. The set of SNOMED concepts was cross-referenced with each of
the other SNOMED axes looking for potential lexical matches that might provide
an appropriate restriction for any of the defining relationships. Prior to
comparison, term strings in the SNOMED test data set and in the target SNOMED
axis were normalized (including all SNOMED synonyms) using the norm program
available from the National Library of
Medicine.23 A
detailed description of this process is provided by McCray et
al.,5 and includes
breaking each string into its constituent words, lower-casing each word,
removing punctuation and stop words, sorting the words in alphabetical order,
and conversion of each word to its uninflected form. In addition, certain
patterns, such as “oscope/oscopy” were identified as high-yield
for detecting matches. Comparison of the normalized concepts included partial
phrase matching (e.g., “Alkaline phosphatase isoenzymes
measurement” would pair with “Alkaline phosphatase
isoenzyme”). Potential matches were placed into a report for manual review. A single
reviewer (RHD) examined the report to determine whether or not the lexically
suggested target SNOMED term would be accepted as a value restriction for the
corresponding test SNOMED concept, and to indicate the defining relationship
for the restriction. Our primary focus was on defining each concept at least with respect to its
measured component (i.e., determining the concept filling the HAS-ANALYTE
value restriction) because procedure names frequently contain some lexical
variant of the substance being analyzed, thereby greatly enabling a lexical
approach, and because a classification based on measured components results in
a clinically sensible hierarchical structure. Terminologic Definitions Are Expressed in a Description Logic
Syntax We used a Knowledge Representation Syntax Specification
(KRSS) 40 -based
description logic syntax to represent terminologic definitions. Once these
definitions were created, they were classified by the K-Rep
system. 36 The
syntax allows us to express a concept's parents (i.e., specify
“IsA” relationships), defining relationships and their
restrictions, and non-defining attributes of a concept (such as its SNOMED
code). Additional expressiveness supported by the syntax, such as cardinality
constraints, was not required to be used.
Figure 2 shows how a sample
LOINC concept might be represented in conceptual graph notation, and a similar
KRSS-like representation. | Figure 2A sample concept defined according to the LOINC model (of
Figure 1). The middle section
shows the concept defined in conceptual graph notation, while the bottom
section shows a similar definition in description logic syntax. |
All target SNOMED axes (i.e., all SNOMED axes that contain value
restrictions for concepts in the SNOMED and LOINC test data sets), are also
defined in the description logic and included in the terminology database. In
general, these axes are comprised of “primitive”
concepts—concepts that cannot be fully defined, and are instead placed
into a manually constructed hierarchy. The hierarchical structure of these
axes exist within the SNOMED database, and was reflected in the description
logic via the specification of “IsA” relationships. For example,
Figure 3 shows a representation
of the SNOMED Function Axis, in which HLA-A-Antigen-NOS_F-C4100 IsA
HLA-Antigen-NOS_F-C4000. Defined Concepts Are Conceptually Merged The initial merge placed LOINC concepts into corresponding SNOMED Procedure
categories (e.g., Chemistry-Procedure_P3-70000,
Microbiology-Procedure_P3-50000), based on the LOINC Class. (This choice was
arbitrary, and subsequent techniques would have been equally applicable had we
instead placed SNOMED concepts into the LOINC classes.) The mapping between
the 18 LOINC classes and the nine corresponding SNOMED Procedure categories
was performed manually. As a result, the IsA relationship for each concept was
set equal to the name of its corresponding SNOMED Procedure category. All
subsequent classification within a category was algorithmically determined,
based on the logical definition of each concept. A detailed description of classification algorithms similar to those used
by K-Rep for this report have previously been
published.32 The
initial step involves parsing the textual concept definitions for syntactic
correctness. Next comes a process of normalization that converts the
definitions into a standard form. This is followed by completion, during which
value restrictions are inherited from the transitive closure of all parent
concepts. These inherited restrictions are logically intersected with local
restrictions contained in the concept definition to form a completely explicit
concept definition. From there, a concept is automatically integrated into the
taxonomy by comparing its definition with definitions of concepts already in
the taxonomy. The process of determining a new concept's location in the
taxonomy is called classification. When one concept is more general than
another, the first concept properly subsumes the second. This can arise if the
restriction of a defining relationship of one concept subsumes the restriction
of the same relationship of another concept, as represented in the target
SNOMED axis, or when the second concept's definition expresses additional
constraints beyond those of the first concept. When two concepts are
equivalent, they subsume one another, but neither properly subsumes the other.
For one concept to subsume another, every value restriction of the first
concept must also be true of the second concept. |
Results Lexically Suggested Value-restriction Determination As noted above, all LOINC concepts are assigned an analyte by the LOINC
committee. In a prior study, the analytes of 5,724 (74%) of the 7,763 LOINC
concepts in our test data set were lexically mapped to SNOMED
concepts, 19 using
the methods described by Rocha et
al. 18
( Table 1). Examining the
LOINC concepts in another way, there are a total of 4,191 unique analytes
among the 7,763 LOINC concepts studied, of which 2,241 (53%) could be mapped
to a SNOMED concept. There are 53 unique properties, 12 (23%) of which could
be mapped to SNOMED; eight unique time aspects, none of which mapped; 83
unique system or specimen types, 61 (73%) of which mapped; four unique scale
types, two (50%) of which mapped; and 140 unique method types, 73 (52%) of
which mapped to a SNOMED concept. | Table 1 Lexical Determination of the Measured Components or Analytes of LOINC and
SNOMED Concepts |
Previously, SNOMED concepts had not been cross-referenced to their
corresponding analytes. Lexical algorithms in this study suggested 3,852
restrictions for the HAS-ANALYTE relationship, of which 1,151 (30%) were
accepted as valid, resulting in 56% of the SNOMED concepts being assigned a
HAS-ANALYTE value restriction (Table
1). These restrictions were found in several SNOMED axes,
including Topography (e.g., Eosinophil, T-C1340), Function (e.g.,
11-Deoxycortisol, F-B2480), Living Organisms (e.g., Bordetella pertussis,
L-12801), and Chemicals (e.g., Acetaminophen, C-60130). 472 restrictions for
the HAS-SYSTEM role and 45 restrictions for the HAS-SCALE-TYPE role were also
determined. No restrictions for the HAS-PROPERTY, HAS-TIME-ASPECT, or
HAS-METHOD-TYPE roles were found using our lexical approach. Overall, of the 9,813 concepts in the combined SNOMED and LOINC test set,
6,875 (70%) had a restriction for the HAS-ANALYTE relationship determined by
lexical techniques. Terminology Merging and Algorithmic Classification As mentioned above, the initial merge placed LOINC concepts into
corresponding SNOMED Procedure categories (e.g., Chemistry-Procedure_P3-70000,
Microbiology-Procedure_P3-50000), based on the LOINC Class. All subsequent
classification within a category was algorithmically determined, based on the
logical definition of each concept. Of the 9,813 concepts in the test data
set, the 6,875 whose measured component was able to be lexically determined
were then merged and algorithmically classified. Because each concept is
defined with respect to its measured component, and because each measured
component is a SNOMED concept, the resulting algorithmic classification
hierarchy assumed the structure of the corresponding analyte hierarchies. This
is illustrated in Figure 3. Faulty classification can result from errors in the reference hierarchy.
This is illustrated in Figure
4. In this example, Androstenedione_F-B2820 is modeled as a child
of Androgen-NOS_F-B2800, but should also be modeled as a child of
17-Ketogenic-steroids_F-B2420. Making this manual correction in the Function
Axis will automatically correct the classification of concepts in the
Procedure Axis. | Figure 4Faulty classification due to reference hierarchy. Classification errors in
the reference Function Axis are mirrored in the automatically generated
Procedure Axis. In this case, Androstenedione_F-B2820 also IsA
17-Ketogenic-steroids_F-B2420. Making (more ...) |
When two concepts measure the same component (i.e., have the same
restriction for the HAS-ANALYTE relationship), the determination of the
parent-child relationship is determined by the other defining relationships
and value restrictions. More formally, if one concept is more general than
another, the first concept is said to properly subsume the second. In such a
case, the second concept's definition expresses additional constraints beyond
those of the first concept. In the hierarchy, the second concept will appear
as a child of the first concept. This is shown in
Figure 5. | Figure 5Classification of laboratory procedures measuring the same substance. When
one concept is more general than another, the first concept will properly
subsume the second. Given the concept definitions shown, the resulting SNOMED
Procedure hierarchy is (more ...) |
Of the SNOMED procedures in the test database, 80% share some measured
component with a LOINC concept, and as a result will acquire a LOINC concept
as a child. Because the initial merge combines LOINC and SNOMED concepts into
SNOMED Procedure categories, all LOINC concepts acquire a SNOMED parent. Of
the LOINC concepts, 50% share some measured component with a SNOMED concept,
and as a result will also be classified under a more granular SNOMED concept
(as in Figure 5). The remaining
LOINC concepts potentially can form a flat hierarchy and an algorithm similar
to that used by Cimino et
al.27,38
can be used to deepen these areas. A flat hierarchy can also result where
there are a number of specific tests without a more general test defined. This
is shown in Figure 4, where the
concept Hormone-measurement-NOS_P3-XFB has been created based on the structure
of the reference hierarchy to deepen the Procedure hierarchy. |
Discussion We found the LALR strategy works well for automatically merging overlapping
terminologies if the terminologies to be merged can be logically defined using
the same model (i.e., both terminologies can be defined using the same
defining relationships), and if their value restrictions can be drawn from the
same reference hierarchy. Key points in our approach are that concept
definitions are expressed in a formal description logic, defining
relationships are taken from a pre-established model (LOINC), lexical
techniques are used in the determination of value restrictions, and value
restrictions for the terminologies to be merged are drawn from the same
reference terminology (SNOMED). Generalizing this approach outside the realm
of laboratory procedures will rely on the presence of defining models for
concepts in other domains, and a rich reference terminology that is likely to
provide lexically suggested value restrictions for concepts in the
terminologies to be merged. Blois41 has
described a hierarchical schema of the sciences, ranging from the low-level
sciences (e.g., mathematics, chemistry, and physics), which lend themselves to
formalization over a relatively small domain of possible values and can
therefore be more precisely defined, to the more abstract sciences (e.g.,
psychiatry and social science), which in part are built up from the underlying
sciences, and are less able to be precisely defined. This would suggest that
the application of description logic and classification might inherently be
easier to implement for certain domains (where concepts can be defined) than
for other domains. Consistent with Blois's assertions are the emergence of
models defining surgical procedures, human anatomy, and
medications,1,14,27,42
although models defining mental health, rehabilitation, and human disease
states are also being
developed.14,43
The emergence of these models will support the extension of our methods
outside the scope of laboratory procedures. Within the LOINC database, concepts filling the HAS-PROPERTY relationship
are drawn from the International Union of Pure and Applied Chemistry
(IUPAC),44 and
those filling the HAS-SYSTEM relationship are drawn from ASTM
E1238-9445 and HL7
version 2.2.46
Lexical techniques were then used to map these concepts to synonymous SNOMED
concepts. If evolving terminologies were to be defined from the outset based
on atomic SNOMED building blocks, subsequent merging efforts could be
substantially decreased, not to mention the beneficial impact this might have
on increasing the overall content coverage of SNOMED. Our test database
contained 2,039 LOINC concepts that could not be lexically mapped to a SNOMED
analyte. To allow these LOINC concepts to be fully defined, 590 new concepts
were added to SNOMED 3.4 (Kent Spackman, personal communication). Such
enhancements to the overall content coverage of SNOMED are likely to benefit
subsequent applications of our approach. Our methodology suggests the need for different manual tasks when modeling
concepts that can be logically defined than when modeling those that cannot.
The reference SNOMED hierarchies that furnished value restrictions (such as
Chemicals, Living Organisms) are generally composed of “primitive”
concepts, meaning concepts that are unable to be fully defined with defining
relationships. Manual tasks may include hierarchy construction and refinement
as well as the addition of new concepts. For those concepts that can be
logically defined (such as laboratory procedures), manual tasks may be better
directed at refining concept definitions, including the review of lexically
suggested restrictions, and adding new reference terms where necessary to
allow for expression of value restrictions. Emerging environments such as
Gálapagos11
will enable the comparison of manually and automatically generated hierarchies
to cross validate both the term definitions and the structure of the reference
hierarchies. Cimino et al.37
have previously described a set of terminology design criteria. While the use
of description logic does add formalism to terminology, allowing for the
greater fulfillment of these criteria, it also will require a consideration of
the optimal balance between expressiveness and tractability (of
classification), similar to such considerations in other areas of medical
informatics and computer science in
general.32,47
Criteria for a healthcare terminology include, among others, the ideas of
nonvagueness and nonredundancy. Nonvagueness is achieved when concepts in the
terminology are complete in meaning. Redundancy exists when multiple terms for
the same concept are added to the terminology as unique concepts. If the
achievement of nonvagueness requires the expressive power of full first-order
logic, then the automatic determination of redundancy can become intractable
or undecidable. The emergence of a workable terminology that is applied to
real clinical situations and to pooled clinical data should help illustrate
these tradeoffs. |
Acknowledgments The authors thank Simon P. Cohn, MD, MPH (Kaiser Permanente Healthcare),
and John E. Mattison, MD (Kaiser Permanente, Southern California Region), for
their support of this work; Stephanie Lipow, of Lexical Technology,
Incorporated, for her help with the lexical processing; the JAMIA reviewers
for their detailed insightful comments; and the members of the Kaiser
Permanente Interregional Convergent Medical Terminology group for their
constant reminder that nothing is without controversy, and that through
controversy comes convergence. |
Notes |
|
References 1. Rector AL, Glowinski AJ, Nowlan WA, Rossi-Mori A. Medical-concept
models and medical records: an approach based on GALEN and PEN&PAD.
J Am Med Informatics Assoc.
1995;2:
19-35. 2. Barrows RC Jr, Cimino JJ, Clayton PD. Mapping clinically useful
terminology to a controlled medical vocabulary. Proc SCAMC.
1994: 211-5. 3. Campbell KE, Das AK, Musen MA. A logical foundation for
representation of clinical data. J Am Med Informatics Assoc.
1994;1:
218-32. 4. Evans DA, Cimino JJ, Hersh WR, Huff SM, Bell DS. Toward a
medical-concept representation language. J Am Med Informatics
Assoc. 1994;1:
207-17. 5. McCray AT, Srinivasan S, Browne AC. Lexical methods for managing
variation in biomedical terminologies. Proc SCAMC.
1994; 235-9. 6. Sheretz DD, Tuttle MS, Olson NE, Erlbaum MS, Nelson SJ. Lexical
mapping in the UMLS Metathesaurus. Proc SCAMC.
1989: 494-9. 7. Computer-Based Patient Record Institute. Proposal to
accelerate standards for computer-based patient record systems.
Version 3.0. CPRI, April 29, 1994. 8. Institute of Medicine, Committee on Regional Health Data Networks.
Health data in the information age: use, disclosure, and
privacy. Washington, DC: National Academy Press,
1994. 9. Cote RA, Rothwell DJ, Beckett RS, Palotay JL (eds). SNOMED
International—The systematized nomenclature of human and veterinary
medicine. Northfield, IL: College of American Pathologists,
1993. 10. Forrey AW, McDonald CJ, DeMoor G, et al. Logical observation
identifier names and codes (LOINC) database: a public use set of codes and
names for electronic reporting of clinical laboratory test results.
Clin Chem. 1996; 42:
81-90. [PubMed]. 11. Campbell KE, Cohn SP, Chute CG, Gennels G, Shortliffe EH.
Gálapagos: computer-based support for
evolution of a convergent medical terminology. J Am Med Informatics
Assoc. Symposium supplement 1996:
269-73. 12. Board of Directors of the American Medical Informatics Association.
Standards for medical identifiers, codes, and messages needed to create an
efficient computer-stored medical record. J Am Med Informatics
Assoc. 1994;1:
1-7. 13. Chute CG, Cohn SP, Campbell KE, Oliver DE, Campbell JR. The content
coverage of clinical classifications. J Am Med Informatics
Assoc. 1996;3:
224-33. 14. Averill RF, Mullin RL, Steinbeck BA, Goldfield NI, Grant T.
The development on the ICD-10 procedure coding system
(ICD-10-PCS). Draft, June 5, 1996. 3M Health Information
Systems, Wallingford, CT. 15. Rothwell DJ, Cote RA. Managing information with SNOMED:
understanding the model. J Am Med Informatics Assoc. Symposium
supplement 1996: 80-3. 17. Coad P, Yourdon E. Object-oriented Analysis. 2nd ed.
Englewood Cliffs, NJ: Prentice-Hall, 1991. 18. Rocha RA, Huff SM. Using diagrams to map controlled medical
vocabularies. Proc SCAMC. 1994:
172-6. 19. Rocha RA, Huff SM. Coupling vocabularies and data structures:
lessons from LOINC™. J Am Med Informatics Assoc.
Symposium supplement 1996:
90-4. 20. Peterson JL. Computer programs for detecting and correcting
spelling errors. Communications of the ACM.
1980;23:
676-87. 21. Zeng Q, Cimino JJ. Mapping medical vocabularies to the Unified
Medical Language System. J Am Med Informatics Assoc. Symposium
supplement 1996: 105-9. 22. Friedman C, Sidelli R. Tolerating spelling errors during patient
validation. Comput Biomed Res.
1992; 25:
486-509. [PubMed]. 24. Rocha RA, Rocha BHSC, Huff SM. Automated translation between
medical vocabularies using a frame-based interlingua. Proc
SCAMC. 1993: 690-4. 25. Masarie FE Jr, Miller RA, Bouhaddou O, Giuse NB, Warner HR. An
interlingua for electronic interchange of medical information: using frames to
map between clinical vocabularies. Comput Biomed Res.
1991; 24:
379-400. [PubMed]. 26. Cimino JJ, Barnett GO. Automated translation between medical
terminologies using semantic definitions. MD Comput.
1990; 7:
104-9. [PubMed]. 27. Cimino JJ, Clayton PD, Hripcsak G, Johnson SB. Knowledge-based
approaches to the maintenance of a large controlled medical terminology.
J Am Med Informatics Assoc.
1994;1:
35-50. 28. Evans DA. Pragmatically-structured, lexical-semantic knowledge
bases for unified medical language systems. Proc SCAMC.
1988: 169-73. 29. Rector AL, Nowlan WA, Glowinski A. Goals for concept representation
in the GALEN Project. Proc SCAMC. 1993:
414-8. 30. Russel SJ. Artificial Intelligence: A Modern
Approach. Englewood Cliffs, NJ: Prentice-Hall,
1995. 31. Doyle J, Patil RS. Two theses of knowledge representation: language
restrictions, taxonomic classification, and the utility of representation
services. Artificial Intelligence.
1991;48:
261-97. 32. Levesque HJ, Brachman RJ. Expressiveness and tractability in
knowledge representation and reasoning. Comput Intel.
1987;3:
78-93. 33. Rector AL, Nowlan WA, Kay S. Conceptual knowledge: the core of
medical information systems. MEDINFO. 1992:
1420-6. 34. Borgida A, Brachman RJ, McGuinness DL, Alperin Resnick L. CLASSIC:
a structural data model for objects. SIGMOD Record.
1989;18:
58-67. 35. Brachman RJ, Schmolze JG. An overview of the KL-ONE knowledge
representation system. Cogn Sci.
1985;9:
171-202. 36. Mays E, Weida R, Dionne R, et al. Scaleable and expressive medical
terminologies. J Am Med Informatics Assoc. Symposium supplement
1996: 259-63. 37. Lindberg DAB, Humphreys BL, McCray AT. The Unified Medical Language
System. Meth Inform Med.
1993; 32:
281-91. [PubMed]. 38. Cimino JJ, Hripcsak G, Johnson SB, Friedman C, Fink DJ, Clayton PD.
UMLS as knowledge base—a rule-based expert system approach to controlled
medical vocabulary management. Proc SCAMC. 1990:
175-9. 39. Fitting M. First-order Logic and Automated Theorem
Proving. New York: Springer-Verlag, 1990. (Gries D, ed.
Texts and Monographs in Computer Science). 41. Blois MS. Medicine and the nature of vertical reasoning. N
Engl J Med. 1988; 318:
847-51. [PubMed]. 43. Brown PJB, O'Neil M, Price C. Semantic representation of disorders
in Version 3 of the Read Codes. Proceedings of the International
Medical Informatics Association, Working Group 6. January
1997, Jacksonville, FL:
209-14. 44. International Union of Pure and Applied Chemistry/International
Federation of Clinical Chemistry. The Silver Book: Compendium of
Terminology and Nomenclature of Properties in Clinical Laboratory
Sciences. Oxford, England: Blackwell Scientific Publishers,
1995. 45. ASTM E1238-94. Standard Specification for Transferring
Clinical Observations between Independent Computer Systems.
Philadelphia, PA: American Society for Testing Materials,
1994. 46. Health Level Seven. An application protocol for electronic
data exchange in healthcare environments. Version 2.2. Ann Arbor, MI:
Health Level Seven, 1994. 47. Dolin RH. Expressiveness and query complexity in an electronic
health record data model. J Am Med Informatics Assoc. Symposium
supplement 1996: 522-6. |
| |