This document organizes and presents the use cases and technical procedures for the development of a SNOMED CT to ICD-10-CM map by the National Library of Medicine for the US realm. This document provides prescriptive guidance on the purposes, conduct and outcomes of that project and is the authoritative source for project execution. A collaborative project in healthcare information interoperability with the National Library of Medicine and the National Center for Health Statistics
The integration of a clinical terminology such as SNOMED Clinical Terms (SNOMED CT) into computer based patient records systems provides a comprehensive and functional terminology of clinical terms, supporting interoperable transmission of patient-related data between information systems. The cross-maps from SNOMED CT to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (US Clinical Modification) are created to support the reimbursement and administrative reporting needs of the IHTSDO member countries which employ ICD-10-CM.
The business application principles of relevance to the deployment of this map include:
This document will reference the data sets, algorithms and intellectual products of the SNOMED CT to ICD-10-CM map as the MAP.
This map is intended to provide support within the specified use case, for IHTSDO members, WHO Collaborating Centers and interested countries which are deploying or have deployed SNOMED CT in clinical information systems and have ICD-10-CM employed in systems for purposes of reimbursement, statistical reporting, epidemiology, cancer registries, injury and other registries, quality reporting, safety reporting and research.
MAP with patient context management: Patient Jones is being discharged from the clinic after an encounter for clinical care. The attending physician has maintained an historical diagnosis and health-related problem list coded in SNOMED CT while caring for Jones and updates the problem list at discharge, selecting the problems which were addressed during today’s visit. The EHR vendor software employs the MAP, and uses a knowledge-based algorithm to evaluate sequential computable Map Rules. These rules evaluate context (data recorded about the patient e.g. age and gender) and co-morbidities (other problems recorded) in the electronic health record to identify the most appropriate candidate ICD-10-CM code list based on ICD-10-CM exclusion / inclusion guidance and other conventions. MAP Vendor software which cannot employ these rule-based features can employ the helpful Map Advice to provide a readable and understandable list of step-by-step instructions for the physician to support a choice of one or more ICD-10-CM codes. If necessary, the physician is prompted for additional information needed to decide between alternative codes or refine the output codes. See Appendix B for description of an algorithm to use the MAP interactively.
Alternatively, the MAP can be used by ICD-10-CM coding professional. The MAP can suggest candidate ICD-10-CM codes based on stored clinical problems encoded in SNOMED CT, which will be reviewed by the coding professional prior to submission for reimbursement by the insurance carrier. The Map Advice data further guides them with information regarding additional NCHS rubrics and requirements.
This mapping is a tabular, knowledge-based cross-link from SNOMED CT to ICD-10-CM in which the most accurate ICD target code or codes that best classify the SNOMED CT concept within the context of the remainder of the record are linked. The map is a link directed from the source SNOMED CT concept to the target ICD-10-CM classification.
The granularity and purpose of ICD-10-CM is different from that of SNOMED CT. SNOMED CT is a comprehensive reference terminology that supports both general and highly specific concepts. Each concept is defined by a set of attribute-value pairs (relationships) which make it distinct from all other concepts. SNOMED CT supports a model of meaning which specifies correct attributes and value sets for each domain of meaning.
ICD-10-CM is a classification of diseases and injuries with granularity of definition that has been chosen to provide utility for purposes of epidemiology, statistical reporting of mortality and morbidity and classification of care for reimbursement purposes. ICD-10-CM was created to classify a clinical concept by defining the classes (or ‘buckets’ of meaning) which contain the concept within the universe of ICD-10-CM classes.
Only domains of SNOMED CT which overlap in meaning with those of ICD-10-CM will be mapped. Due to differences in granularity, purpose and rubrics, assignment of a map equivalence between the SNOMED source and ICD target concepts is usually not appropriate. Instead, the map will link a SNOMED source concept to the ordered set of ICD-10-CM classes which contain the elements of meaning of the SNOMED concept as conceptualized by ICD-10-CM.
All pre-coordinated concepts issued by the IHTSDO within the current international release of SNOMED CT with active status, excluding concepts found in the IHTSDO non-human subset, within the following SNOMED CT domains may be mapped:
ICD-10-CM is a classification of diseases and findings for purposes of statistical reporting, epidemiology, reimbursement, cancer registries, injury and other registries, quality reporting, safety reporting and research. ICD-10-CM replaces ICD-9-CM volumes 1 and 2 for use in US healthcare settings in 2014.
The scope of ICD-10-CM is primarily designed for the classification of diseases and injuries designated as a formal diagnosis. ICD also classifies signs, symptoms, abnormal findings, complaints, and social circumstances that may appear in a health record. Thus ICD is used to classify data recorded under headings such as 'diagnosis', 'reason for admission', 'conditions treated', and 'reason for consultation'."
Some SNOMED CT concepts within the source domains may be normal findings or other concepts not intended for classification by ICD-10-CM. These are included in the MAP with records designating these concepts as “not classifiable” by ICD-10-CM
The MAP has cardinality of one SNOMED CT source concept to zero-to-many ICD-10-CM classification codes. Zero target codes are appropriate only if the source concept is not classifiable or is awaiting editorial review for guidance. In all other cases one or a combination of more than one ICD-10-CM codes will be assigned. Practically speaking, this is usually no more than three. More than one SNOMED CT concept can map to the same ICD-10-CM target.
When there is a need to choose between alternative ICD-10-CM codes, each possible target code is represented as a Map Rule (rule-based map). Related Map Rules are grouped together into one Map Group. Each Map Group will resolve to at most one target ICD-10-CM code at run-time.
The SNOMED CT mapping is constructed using the SNOMED CT, current version published by the International Health Terminology Standards Development Organisation; and the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Clinical Modification ICD-10-CM, published and maintained by the National Center for Health Statistics. The MAP will be reviewed and updated for each new release of SNOMED CT and ICD-10-CM. In planning for the development of this map, advice was sought from the Center for Medicare and Medicaid Services regarding ICD-10-CM code restrictions for reimbursement in the US. Due to regional variations in care profiles and carrier procedures for authorization of reimbursement, no authoritative resource could be identified to guide target code selection beyond those published by NCHS.
Release data structures for the MAP will be compliant with SNOMED CT Release Format 2 (RF2) datasets as defined in the IHTSDO document: SNOMED CT Release Format 2.0: Data Structures Specification. MAP data structures and definitions important to MAP deliverables defined below include:
Three additional Map Categories identify editorial process states but are published as a single Map Record for the source concept for completeness, auditability and transparency of the MAP:
Formatting of the Map Rule is specified in Augmented Bacchus Naur Form (ABNF) in Appendix C but generally is one of three forms:
a) a statement of the Map Rule in readable terms for the clinical user and
b) supplementary metadata guidance intended to clarify the map for the coding professional. Metadata advice supported in the MAP includes:
Assumptions that are central to the construction of the MAP include:
The SNOMED CT concept or statement taken from the health care record will be evaluated for meaning within the guidelines of the SNOMED Clinical Terms User Guide. No assumed context or modifying semantics will be inferred beyond the definition asserted by the fully specified name and the SNOMED CT defining relationships, excluding qualifiers. Identification of inconsistency between the fully specified name and the synonyms, or between the fully specified name and the defining relationships will constitute a case for ambiguity. This will cause a map record to be flagged for editorial review by the IHTSDO editorial staff. An understanding of the meaning (semantics) of the SNOMED CT concept is a necessary first step to an understandable, reproducible and useful map.
The organization, structure and conventions of the classification ICD-10-CM is different than SNOMED CT and meaning (semantics) within the classification is specified by the order and relationship of the chapters, section and categories. The position of a classification code within the axis, the title of the code and the associated conventions and guidance provided from the authoritative source further contribute to the specification of meaning of a classification code. Finally, since ICD-10-CM is a classification, the semantic space of a particular classification code depends upon the definition of sibling codes and others within the same category. Since ICD-10-CM is designed for statistical, epidemiological and reimbursement purposes, one ICD-10-CM classification code may include many SNOMED CT concepts within its semantic space.
SNOMED CT is a reference terminology that expresses the semantics of concepts within its domain by means of a controlled vocabulary and use of an extensive set of defining relationships. The relationships are employed in concept definition within a constrained and defined model of meaning applicable to each SNOMED semantic root. Understanding the meaning of a SNOMED CT concept requires evaluation of the vocabulary as well as the defining relationships.
The goal of the mapping process is to identify the meaning of a SNOMED CT concept, determine the best location of that concept in the ICD-10-CM semantic space as identified by one or more ICD-10-CM classification codes, and to create a link between the SNOMED CT concept identifier and the correct ICD-10-CM code(s). Since SNOMED CT is a reference terminology, this process cannot proceed reproducibly using only naming (terms or descriptions) conventions. A full understanding of both SNOMED CT and ICD-10-CM semantics, as specified in Section 11 Mapping Heuristics, is required for success.
As an example, the SNOMED CT concept 235991007 |Peritoneal eosinophilia (disorder)|, may be identified as a type of blood disorder by some lexical (terming) coding tools and mapped to D72.89 "Other specified disorders of white cells" in ICD-10-CM. However the concept 235991007 |Peritoneal eosinophila (disorder)| has defining relationships 213293008 |is a (attribute)| = 213293008 |Aseptic peritonitis (disorder)| with 116676008 |associated morphology (attribute)| = 23583003 |Inflammation (morphologic abnormality)| and 363698007 |finding site (attribute)| = 15425007 |Peritoneum structure|. From these relationships, the mapper should correctly identify that the SNOMED CT concept is an inflammatory gastrointestinal disorder, and appropriately map the concept to the ICD-10-CM semantic space K65.8 "Other peritonitis".
The Mapping Specialist will initiate the mapping process by evaluating the source SNOMED CT concept employing the Context-free assumption: SNOMED CT concepts to be mapped from the source domains will be presumed to “speak for themselves”. The concept definition as asserted in the fully specified name and the concept’s defining relationships (excluding qualifiers) will be presumed to encompass all information available for definition of the concept and interpreted as explained in the SNOMED Clinical Terms User Guide. This information alone will be employed by the map specialist in assessing the source meaning and researching target code(s) for the MAP. If the meaning of the SNOMED CT concept is judged to be ambiguous by the mapper, the concept will be managed as follows:
Cases for concern or question of ambiguity in the SNOMED CT source concept definition will include:
Discrepancies will be assessed relative to standard medical references and compared to guidance and definitions provided in the ICD-10-CM authoritative source.
Case #1 constitutes fundamental ambiguity in the meaning of the SNOMED CT concept and cannot be mapped pending clarification. These concepts will be assigned a map record with MAP CATEGORY of ambiguous (447640006 |Source SNOMED concept is ambiguous|). The editorial notes field will be populated with any information on the map accrued by the mappers during their research. The concept will be referred to an arbitration committee of SNOMED CT editors and the map lead. Once the ambiguity is resolved, the map will be completed and MAP CATEGORY will be updated with the appropriate assignment.
Case #2 represents a confusing issue for the mapper using only lexical tools to review SNOMED CT and ICD-10-CM. However the SNOMED CT definition is not truly ambiguous and the map for this concept will be completed as described herein. The SNOMED CT term which is the source of the confusion will be flagged by the mapping team in the editor notes for editorial review by the arbitration committee with the expectation that the confusing term will be marked for demotion as a non-synonymous lexical tag.
Examples of confusing and ambiguous mapping source concepts include:
The context-free assumption will require agreement regarding procedures for the map when certain elements of context are asserted in either the SNOMED CT source or the ICD-10-CM target reference. These context issues, along with mappings to multiple target codes, are detailed in the following. In all discussions, source always refers to the SNOMED CT concept and target refers to the ICD-10-CM classification.
Once the map specialist has reviewed and understands the SNOMED CT source concept to be mapped, she will employ the ICD-10-CM alphabetic index of diseases and nature of injury to research and select candidate ICD target codes for the map. She will usually use the Fully Specified Name from the SNOMED source concept to search the best textual references in the alphabetic reference that capture the meaning of the SNOMED CT concept. The alpha reference is organized hierarchically when multiple term modifiers may be considered as specifications of the primary term, and the map specialist will search through the specifications looking for the terms of interest.
This review may identify matching ICD terms, or may require searching through related cross references. Two types of cross-reference in the alphabetic listing need to be considered before assignment of a tentative target code(s). These are: ‘-see…’, and ‘-see also…’.
‘-see…’ is an explicit direction to look elsewhere in the index and no codes are found alongside this reference. The map specialist will review the directed descriptions in searching for target candidates.
Nodule(s), nodular - actinomycotic - see Actinomycosis |
‘-see also…’ instructs the user to look elsewhere if the detail they are looking for cannot be determined from the reference.
Anaplasia cervix (see also Dysplasia, cervix) N87.9 |
When a code is given alongside the term showing the ‘-see also…’ note, which appears in parentheses as in the above example, it may or may not be necessary to look under the alternative term. When no code is given, the instruction does not appear in parentheses as it is essential for the user to look under the alternative lead term. In this example with reference to the term “Angiofibroma”, the mapper must also evaluate the alpha references for the specific condition for relevant target classifications:
Angiofibroma —see also Neoplasm, benign, by site |
The map specialist will create preliminary map records employing the candidate target records resulting from these searches. If specializations are noted in targets selected from the alphabetic index and there exist sub-headings for the term selected, these will be considered for exclusion rules as documented in section 11.10 below. For source concepts that represent examples of poisonings caused by drugs or chemicals, the map specialist will use the index ICD-10-CM Table of Drugs and Chemicals to research the default maps and essential modifiers. An example of this is provided in section 11.6 below.
The map specialist will proceed from this source concept analysis to research the tabular guidance for the candidate targets and evaluate for issues of context which will alter the map records as explained in the following sections 11.3 through 11.11.
ICD-10-CM employs specializations of the International version of ICD-10 for coding of laterality within certain classification categories. For example,
H60.9 : Unspecified otitis externa H60.90 : Unspecified otitis externa, unspecified ear H60.91 : Unspecified otitis externa, right ear H60.92 : Unspecified otitis externa, left ear H60.93 : Unspecified otitis externa, bilateral
|
When the SNOMED CT source concept employs a specification of laterality (rare within the core international release) then the appropriate ICD-10-CM specialization will be chosen as the target code. As is more often true, if the source code does not specify laterality, then the unspecified target will be chosen –H60.90 in this case. The Map Advice CONSIDER LATERALITY SPECIFICATION will be added to the map record. (Exemplars: Laterality)
Classification categories such as M80 Osteoporosis with current pathologic fracture have guidance that 7th digit specialization for episode of care must be specified when choosing an ICD-10-CM target code. Category guidance includes a statement such as:
The appropriate 7th character is to be added to each code from category M80: A initial encounter for fracture D subsequent encounter for fracture with routine healing G subsequent encounter for fracture with delayed healing K subsequent encounter for fracture with nonunion P subsequent encounter for fracture with malunion S sequela |
When the SNOMED CT source code specifies a relationship to the episode of care, then the appropriate ICD-10-CM 7th digit specialization will be chosen for the target classification. When as often true, the SNOMED CT source does not specify such information, then a “wildcard” place-keeper of “?” will be included in the 7th digit of the target code created. For example, “203445009 | Osteoporosis of disuse with pathological fracture (disorder) |” maps to
“M80.00X? Age-related osteoporosis with current pathological fracture, unspecified site, episode of care unspecified”. The Map Advice EPISODE OF CARE INFORMATION NEEDED will be added to the map record indicating that additional episode of care information is required for full specification of the map target. (Exemplars: Episode)
If the source concept asserts a gender restriction, a target will be selected including the restriction. If no targets with restriction apply, a more general target will be chosen. The Map Category will indicate that the source is properly classified (447637006 |Map source concept is properly classified (foundation metadata concept)|) in ICD-10-CM, the Map Rule will default to TRUE and Map Advice will be NULL.
If the source concept does not assert gender yet only gender restricted target codes are found within ICD-10-CM, the map will be considered CONTEXT DEPENDENT. The mapping specialist will create up to three map records for this source concept, including rules for mapping to female and male record context as appropriate to the ICD-10-CM structures, and a default record providing Map Advice for the user should rule evaluation fail to specify an ICD-10-CM target code. Map rules for gender context will be of the form “IFA {248152002 Female (finding) OR 248153007 Male (finding)}”.
The gender rule is usually the last rule in a group, just before the OTHERWISE TRUE rule.
If the source concept asserts an age or phase of life for onset of the disorder, a target will be selected first including the restriction or, if none is available, then a more general classification target (Map Category="Properly classified", Map Rule =TRUE, Map Advice=NULL) that is inclusive of the source concept. No Map Rule restrictions for age will be applied in cases where there is a properly classified ICD-10-CM Map target.
If the authoritative sources include specifications for employment of age, those will be used. Otherwise, these definitions for common phases of life will be employed when SNOMED CT or ICD-10-CM employ these descriptive terms and the age of onset restrictions in parentheses will be applied to the map rule:
Employing the map tooling, lower boundaries of age will be asserted with the operator ‘greater than or equal to’. Upper boundaries of age will be asserted with the operator ‘less than’ as identified in the definitions above.
If the source concept does not assert age or time of life and only age-restricted targets are available, the map will be considered CONTEXT DEPENDENT and the mapping specialist will assemble two or more Map Records including Map Rules to properly classify to each ICD-10-CM target classification. The map Rule will be constructed with reference to the SNOMED CT observable for “Age at onset of clinical finding”.
NOTES AND EXCEPTIONS:
The age rule is usually the last rule in a group, just before the OTHERWISE TRUE rule. In cases where there is both an age and gender rule, the age rule should be above the gender rule.
A source concept which identifies origination as a congenital or acquired condition will be mapped to a target of consistent or more general classification should one exist (Map Category="Properly classified", Map Rule=NULL). If a source concept is general and only specific targets exist, ICD-10-CM authoritative source will be searched for guidance of a default map record, either "Congenital" or "Acquired". When such advice is provided, this context will be employed to create the appropriate Map Record and Map Category will be (447637006 |Map source concept is properly classified|) Properly Classified, Map Rule=NULL, Map Advice = “MAPPED WITH NCHS GUIDANCE”. When the source concept is general and no guidance is provided, the map specialist will create map records employing Map Rules relevant for all appropriate targets.
Source concepts representing a poisoning from a drug or noxious substance may be mapped to one, two or three target classification codes depending upon the specificity of the source concept. The poisoning code for the drug or substance will be designated as the primary target code (MAP GROUP = 1), ICD chapter T36-T50. NCHS guidance dictates that the sole exception to this rule occurs when the manifestation is a neoplastic disorder. Neoplasms are always mapped as the first target code.
Should the source concept specify the symptoms or findings resulting from the poisoning, the appropriate target code will be mapped as the second map (MAP GROUP = 2) except in cases of neoplastic complications when the poisoning code will occur second.
When the SNOMED CT concept specifies the action intent/site of injury involved within the event, a specific ICD classification code from the range V00-Y99 External causes of morbidity will be employed as the second or third target code (MAP GROUP = 2 or 3). If the source concept does not specify intent, NCHS guidance will be reviewed for a default map which will be assigned a Map Category of "Properly classified” with Map advice of “MAPPED WITH NCHS GUIDANCE".
“T43.201? Poisoning by unspecified antidepressants, accidental (unintentional), episode of care unspecified” as the default target code. (Exemplars:Poisoning:#6)
Source concepts denoting a condition with an identifiable cause within scope of ICD-10-CM chapter XX (20) will be mapped to two target classification codes. The external cause classification will be assigned to the second target record (MAP GROUP=2)
If the external cause or location is not explicit in the SNOMED CT concept, the generic external cause code will not be listed as a target ICD-10-CM code.
Source concepts which are SNOMED CT 272379006 Events will be mapped as External causes or as Factors influencing health status when these are within the scope of ICD-10-CM
ICD-10-CM guidance does not include dagger and asterisk conventions found in the WHO publication of ICD-10. Mapping many concepts which follow those conventions internationally may NOT require two ICD-10-CM codes. NCHS guidance instead employs “USE ADDITIONAL” guidelines when necessary, otherwise more specific ICD-10-CM codes may have been generated to codes these cases.
Guidance provided by NCHS includes frequent examples of advice of the form:
“USE ADDITIONAL: code to....”
When encountered as advice during selection of a map target, the map specialist will consider this advice to be required for analysis relating to the possible addition of an additional map group. For example:
Apart from the “USE ADDITIONAL: code to....” case, multiple targets (multiple map groups) will generally be needed in the following two situations:
A second target code will not be used to capture the residual meaning of a SNOMED CT concept that is not explicitly covered by the ICD-10-CM code e.g. 301783004 |Abnormal perimenopausal bleeding (disorder) is mapped to N95.8 “Other specified menopausal and perimenopausal disorders” which broadly encompasses the meaning of the SNOMED CT concept. A second map group will not be added to capture the specific symptom of bleeding.
Managing context for ICD-10-CM exclusions proceeds with the assumption that mapping is occurring from SNOMED CT concepts in a clinical diagnosis list (problem list) to the ICD classification. It is possible that a SNOMED CT encoded problem can be further specified or that other concepts in that diagnosis list may be more specific statements of the same or similar diagnoses. More often than not, the more specific SNOMED CT diagnosis will have the same ICD-10-CM target. However, sometimes a more specific SNOMED CT concept will require a different map target for proper classification due to the exclusion and other rules in ICD-10-CM. In order to alert users of the MAP that the refinement of the SNOMED CT code can lead to a different ICD-10-CM code, exhaustive mapping of all descendants of relatively low level SNOMED CT source concepts is done. For all SNOMED CT concepts with 10 or less descendants (“low level concepts”), every descendant concept is evaluated independently to see if they map to a different ICD-10-CM target according to the exclusion and other rules in ICD-10-CM. Each descendant with different targets than their parent concepts will be represented by a separate Map Rule. Due to pragmatic considerations, and to avoid over-proliferation of Map Rules, SNOMED CT concepts with more than 10 descendants (“high level concepts”) are not exhaustively mapped. Those concepts will only have a default ICD-10-CM target and the Map Advice DESCENDANTS NOT EXHAUSTIVELY MAPPED.
Each descendant of a low level concept that has been exhaustively mapped will appear in the final publication file as an independently-mapped source concept.
Sometimes, a descendant concept will require an age or gender rule. This will result in a compound rule e.g. IFA (SNOMED CT concept1) AND age of onset < x years, map to target code T1.
Exclusion guidelines from NCHS coding publications will be evaluated to determine the appropriate target code. Generally, these exclusions are presented as essential modifiers in the alpha index or as “Excludes:” notes in the tabular list. The process of reviewing the exclusions is outlined below.
There are two types of modifier which appear in the ICD-10-CM Alphabetical Index, Volume 3. These are non-essential and essential modifiers.
Non-essential modifiers appear in parentheses following the terms they modify and do not affect the target code selection for a given condition, sign or symptom but are considered as alternatives to the expression of the term.
Polyuria (nocturnal) R35 |
Essential modifiers appear next to a lead term or as subterms indented below lead terms in the alphabetical index and do affect the selection of target code. They describe essential differences in site, etiology or type of disorder and must appear in the clinical statement for the code to be assigned. When an essential modifier denotes an alternative map target to the source term statement, the modifier will be considered as a possible exclusion to the map target selected. An exclusion rule will be created if that exclusion applies to a descendant of the source SNOMED CT concept.
Encephalopathia hyperbilirubinemica, newborn P57.9 - due to isoimmunization (conditions in P55.-) P57.0
Polyuria (nocturnal) R35 - psychogenic F45.3 |
Once target codes are selected employing the essential modifiers, the map specialist will examine “Excludes:” notes in the tabular list to see if any of them apply to descendants of the source SNOMED CT concept. If so, exclusion rules will be created.
All source concepts representing neoplastic disorders will be mapped. Map Groups will specify the ICD-10-CM code(s) from Chapter II for the concept. Morphology mapping with ICD-O is out of scope for the MAP.
Source concepts which specify birth findings by location will be mapped to a specific ICD-10-CM target when such exists. SNOMED CT birth findings which do not specify numbers of child born will be assumed to be singleton births.
For purposes of editorial clarity and error reduction, in the editing environment, the source MAP will be maintained in LONG editorial form which assigns a mapping row to each and every descendant of a low level SNOMED CT source concept that is exhaustively mapped.
To reduce cluttering and improve readability, at publication, the map records for descendant concepts is evaluated algorithmically, and concepts whose descendants’ maps match exactly that of the source concept are not shown. This is the short form which is the only form in which the map is published. This means that all maps can be interpreted as applicable to a source concept and its descendants, unless when it is stated to be different by an IFA rule. High level concepts are an exception to this, since the descendants are not exhaustively mapped (as stated in the map advice).
The MAP is released as refsets according to the Release Format 2 (RF2) specification of the IHTSDO. A human-readable text-delimited file is also available to facilitate browsing of the MAP. The IHTSDO publications SNOMED CT Users Guide and Technical Reference Manual will provide documentation as to the MAP and definitions and description of the mapping data sets.
National Library of Medicine will identify publication mechanisms as they see fit.
The MAP shall be subject to review with each new version of ICD-10-CM and SNOMED CT.
Modifications and expansion of mapping heuristics and procedures may be made throughout the mapping project and will be documented by updating of this publication.
The MAP, MAP data structures and all documentation will be maintained and distributed in US English.
The I-MAGIC Algorithm (Interactive Map-Assisted Generation of ICD Codes)
This is an algorithm which utilizes the SNOMED CT to ICD-10-CM Rule Based Map in a real-time, interactive manner to generate ICD-10-CM codes from SNOMED CT encoded clinical problems.
Item | Description | Example
|
A | Whether a default ICD-10-CM target code is present in the Map | Yes: 6020002 Tinea pedis (disorder)→ B35.3 Tinea pedis
No: 410070006 Herniated urinary bladder (disorder)
|
B | With additional information from the EHR, whether there is sufficient information to arrive at a valid ICD-10-CM code | Yes: 410070006 Herniated urinary bladder (disorder) + gender information, e.g. if gender = Female → N81.10 Cystocele, unspecified
No: 371162008 Closed fracture of skull (disorder)
|
C | A valid ICD-10-CM code is identified for the SNOMED CT problem, but in some cases, it is possible to refine the code by adding extra information. The user decides whether the additional refinement is necessary | 67321002 Acquired deformity of hip (disorder) maps to M21.959 Unspecified acquired deformity of unspecified thigh by default. User can refine this concept by choosing one of the descendants (e.g., 12067001 Acquired coxa vara (disorder)) that will map to a different ICD-10-CM code. User can also specify laterality if desired
34842007 Antepartum hemorrhage maps to O46.90 Antepartum hemorrhage, unspecified, unspecified trimester by default. User can specify trimester if desired
|
1 | The application gathers patient context information (e.g. age, gender) from the EHR | In the case of 410070006 Herniated urinary bladder (disorder), the application looks for patient gender in the EHR, if Female → N81.10 Cystocele, unspecified, if Male → N32.89 Other specified disorders of bladder
|
2 | Get additional input from user | The method to interact with the user is implementation dependent. For example, for 371162008 Closed fracture of skull (disorder), the application can display the following list for the user to pick:
S02.91- Unspecified fracture of skull
|
3 | The application displays the applicable ICD-10-CM codes after evaluating the map rules based on patient context information, additional input from user and other problems on the problem list
|
|
4 | Get user input to refine ICD-10-CM code
| For 67321002 Acquired deformity of hip (disorder), the application displays the following choices
|
Notes:
Rule = TruthStatement / Clause
Clause = (ClauseFinding / ClauseObservable) [ ws ANDOP ws (ClauseFinding / ClauseObservable) ] ws
TruthStatement = ws 1*1("true" / "otherwise true") ws
;;; A Rule is either a truth statement or a clause
;;; A truth statement is either "true" or "otherwise true"
;;; A clause is either a clause with a finding or a clause with an observable
;;; and value optionally followed by the AND operator and a clause with a
;;; finding or observable and value.
ClauseObservable = "IFA" ws ( AttributeObservable ws NumericOperator ws Value )
;;; Observable clause has a mandatory value
ClauseFinding = "IFA" ws ( AttributeFinding )
;;;
AttributeObservable = ConceptObservable
;;; This could be removed and AttributeObservable changed to
;;; ConceptObservable without affecting the grammar
AttributeFinding = ConceptFinding
;;; This could be removed and AttributeFinding changed to
;;; ConceptFinding without affecting the grammar
NumericOperator = (“<” / “>=”)
;;;Age at onset rules use greaterthanorequals for lower bounds and lessthan for upper bound
Value = ConceptAny / Numeric / OtherText
;;; Added in ConceptAny, which allows any Concept to be used as a Value,
;;; although clearly not all concepts are suitable
ConceptObservable = SctId ws pipe ws FullySpecifiedNameObservable ws pipe
ConceptFinding = SctId ws pipe ws FullySpecifiedNameFinding ws pipe
ConceptAny = SctId ws pipe ws FullySpecifiedName ws pipe
SctId = 6*18( digit )
FullySpecifiedNameObservable = 1*nonwsnonpipe *( ( 1*SP "(" *SP 1*nonwsnonparennonpipe *SP ")" !( ws pipe ) ) / ( 1*SP 1*nonwsnonparennonpipe !( ws pipe ) ) ) *SP 1*1( "(observable entity)" )
;;; The FSN of an observable must have a semantic tag = "observable entity"
;;; and may contain other embedded parenthesised strings. The ! (NOT)
;;; look-ahead operator serves to prevent the parser consuming the
;;; semantic tag when it's looking for words before the tag.
FullySpecifiedNameFinding = 1*nonwsnonpipe *( ( 1*SP "(" *SP 1*nonwsnonparennonpipe *SP ")" !( ws pipe ) ) / ( 1*SP 1*nonwsnonparennonpipe !( ws pipe ) ) ) *SP 1*1( "(finding)" / "(disorder)" )
;;; The FSN of a finding must have a semantic tag = "disorder" or
;;; "finding" and may contain other embedded parenthesised strings.
;;; The ! (NOT) look-ahead operator serves to prevent the parser consuming
;;; the semantic tag when it's looking for words before the tag.
FullySpecifiedName = 1*nonwsnonpipe *( ( 1*SP "(" *SP 1*nonwsnonparennonpipe *SP ")" !( ws pipe ) ) / ( 1*SP 1*nonwsnonparennonpipe !( ws pipe ) ) ) *SP 1*1( "(" SemanticTag ")" )
;;; Any FSN must have a semantic tag and may contain other embedded
;;; parenthesised strings. The ! (NOT) look-ahead operator serves to
;;; prevent the parser consuming the semantic tag when it's looking for
;;; words before the tag.
Numeric = 1*(digit) !( *ws /OtherText )
OtherText = 1*(nonwsnonsemicolonnonpipe) *( 1*SP 1*nonwsnonsemicolonnonpipe )
;;; OtherText is used in Value and may not contain a semicolon because
;;; semicolon is the AND operator and follows a Value. Note that FSNs in
;;; Values may contain semicolons.
SemanticTag = 1*(nonwsnonparennonpipe) *( 1*SP 1*nonwsnonparennonpipe )
;;; A Semantic Tag may consist of words separated by whitespace, but may
;;; contain whitespace.
digit = %x30-39
ws = *( SP / HTAB / CR / LF )
SP = %x20
HTAB = %x09
CR = %x0D
LF = %x0A
pipe = %x7C
nonwsnonparennonpipe = %x21-27 / %x2A-7B / %x7D-7E / UTF8-2 / UTF8-3 / UTF8-4 ; no parentheses
nonwsnonsemicolonnonpipe = %x21-3A / %x3C-7B / %x7D-7E / UTF8-2 / UTF8-3 / UTF8-4 ; no parentheses
nonwsnonpipe = %x28-29 / nonwsnonparennonpipe
ANDOP = “AND” ;
UTF8-2 = %xC2-DF UTF8-tail
UTF8-3 = %xE0 %xA0-BF UTF8-tail / %xE1-EC 2( UTF8-tail ) / %xED %x80-9F UTF8-tail / %xEE-EF 2( UTF8-tail )
UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / %xF1-F3 3( UTF8-tail ) / %xF4 %x80-8F 2( UTF8-tail )
UTF8-tail = %x80-BF
Approvals
|
Draft Amendment History
|