https://snomed.atlassian.net/browse/IHTSDO-89

Amendment History

Review Timetable

© International Health Terminology Standards Development Organisation 2012. All rights reserved.

SNOMED CT® was originally created by the College of American Pathologists.

The International Release of SNOMED CT® is distributed by the International Health Terminology Standards Development Organisation (IHTSDO), and is subject to the IHTSDO’s SNOMED CT® Affiliate Licence. Details of the SNOMED CT® Affiliate Licence may be found at http://www.ihtsdo.org/our-standards/licensing/.

No part of this document may be reproduced or transmitted in any form or by any means, or stored in any kind of retrieval system, except by an Affiliate of the IHTSDO in accordance with the SNOMED CT® Affiliate Licence. Any modification of this document (including without limitation the removal or modification of this notice) is prohibited without the express written permission of the IHTSDO.

Any copy of this document that is not obtained directly from the IHTSDO [or a Member of the IHTSDO] is not controlled by the IHTSDO, and may have been modified and may be out of date.  Any recipient of this document who has received it by other means is encouraged to obtain a copy directly from the IHTSDO [or a Member of the IHTSDO. Details of the Members of the IHTSDO may be found at http://www.ihtsdo.org/members/].

1 Glossary

1.1 Domain Terms

2 Introduction

2.1 Purpose

The purpose of this project is to inactivate the “morphologic abnormality” 107656002 – Congenital anomaly (morphologic abnormality) as well as the concept 21390004 - Developmental anomaly (morphologic abnormality) and its subtypes due to the inconsistency of modeling congenital disorders with either the OCCURRENCE attribute or the ASSOCIATED MORPHOLOGY attribute (or both) to represent congenital diseases.  It also proposes to revise existing “Congenital diseases” to a consistent modeling pattern using proximal primitive parents and an expansion of the PATHOLOGICAL PROCESS relationship range to provide more robust use of the classifier to infer parents.

SNOMED CT projects transition from Inception Phase → Elaboration Phase → Construction Phase → Transition Phase. This document describes the Inception Phase. The elaboration phase, in which one or more technical solutions may be developed and tested, may result in more than one document.

The purpose of the Inception Phase is to agree with stakeholders the detail of the problem to be addressed and its scope boundaries.  The resulting problem description must also be of sufficient detail such that the size and impact of any resolution might have on the terminology as a whole and its users can be understood.

The purpose of the Elaboration Phase is to develop, document and test one (or more) possible  technical solutions, and to reach a recommendation and provide a detailed specification of a preferred solution to be taken forward to the construction phase.

2.2 Audience

The audience for this document includes implementers who are involved in the use of Congenital and rare disease but especially SNOMED CT editorial staff who are responsible for consistent high quality modeling of new and existing SNOMED CT content.

2.2.1 Identification of stakeholders

SNOMED CT editorial staff

IHTSDO consultant terminologists with permission to edit in the International release

NRC editors of national extensions

Users of SNOMED CT focused on congenital and rare diseases

2.2.2 Input from stakeholders

No input as to the first phase of the project has been received.

2.2.3 Degree of consensus on the statement of problem

This is a recognized problem as it has been noted in the 20160731 edition of the Technical Implementation Guide:

6.6.2.2 Congenital anomaly

There is significant doubt about the usefulness of some of the congenital anomaly morphologies. Many disorders that involve congenital anomalies can be defined in terms of the OCCURRENCE attribute, leaving no need for a congenital version of the more general morphology concept. For example, 90293002 congenital stenosis (morphologic abnormality) could be removed because 415582006 stenosis (morphologic abnormality) can be combined with OCCURRENCE = congenital to define disorders that involve congenital stenosis. 

Because they are currently in the hierarchy, these congenital morphology concepts should be properly placed. Congenital anomaly morphology concepts should usually have non-congenital parents. For example, congenital stenosis needs to be a child of stenosis, in addition to being a child of congenital anomaly.

Consequently, changes to the editorial guide have been made which outlined the results of the first iteration of the elaboration and testing of this document:

6.1.3.12.1 Congenital

The attribute-value pair | occurrence |=| congenital | is applied to those disorders that are present at birth. Although the word congenital is often applied to genetic disorders, the term genetic is preferred for those disorders that arise from abnormalities of the genes.

The preferred modeling pattern for congenital disorders requires consideration of the possible use of | associated morphology | with values that are congenital morphologic abnormalities, and the | occurrence | attribute with a value of "congenital". The currently accepted pattern is to use | occurrence |=| congenital |, and to discontinue the use of congenital morphologic abnormality concepts as values of | associated morphology |, replacing them with their non-congenital morphologic supertype. Once the congenital morphologies are no longer in use, they will be retired.
| Occurrence | should be in the same role group as | associated morphology | and | finding site |, because the morphology is located at the site, and the occurrence applies to the combined morphology / site pair.

For general congenital anomaly disorder grouper concepts such as | Congenital anomaly of cardiovascular system (disorder) |, the preferred value for | associated morphology | is | Developmental anomaly (morphologic abnormality) |.

When modeling congenital disorders, the following guidelines should be followed:

3 Statement of the problem or need

3.1 Summary of problem or need, as reported

There are currently 141 terms in the morphologic abnormality hierarchy that contain the string "congenital" either in the FSN or a description.  A review of the necessity of having both a congenital morphology and a congenital occurrence should be undertaken in order to help clean up the current congenital disease hierarchy. Additionally, then need to differentiate between those abnormal morphologies that arose through an abnormal developmental process vs. those that were acquired after birth is necessary.  Currently, this distinction is being made explicitly through the use of the concept 21390004 | Developmental anomaly (morphologic abnormality) and its subtypes, many of which represent the same morphology, only being differentiated by their presence at birth.

3.2 Summary of requested solution

There are two possible solutions:

1. Provide clear editorial guidance for the modeling of congenital disorders to eliminate the variability in the concepts.

2. Retire and replace relationships on concepts that use "Congenital anomlay" or "Developmental anomaly" and its subtype in favor of using the non-"congenital" form of thet morphology along with the OCCURRENCE relationship to provide a single way of representing congenital diseases.

3.3 Statement of problem as understood

When reviewing the child concepts of 66091009 – Congenital disease it is found that some concepts represent the congenital nature of the disorder by using the OCCURRENCE attribute with a value of 255399007 – Congenital (qualifier value), some by using the ASSOCIATED MORPHOLOGY attribute with a child of 21390004 | Developmental anomaly (morphologic abnormality) or 107656002 – Congenital anomaly (morphologic abnormality) and some with both of these attributes.  While initial guidance was provided to attempt to resolve this, inital modeling identified the specific need to explicitly distincguish those morpholgies that were developmental in nature from those that can be either acquired or developmental. It is clear that another approach to representing themechanismby which the morphology appears is needed.

3.4 Detailed analysis of reported problem, including background

The inconsistent approach to modeling congenital disease causes issues with classifier inferences as well as improper placement in some cases of new leaf concepts due to ancestor model inconsistencies. The classifier is not being leveraged in many cases to fill out the proper hierarchical structure and redundancies and crossover relationships exist.

There are approximately 6800 concepts currently listed as children of 66091009 – Congenital disease (disorder).  The various ways to represent the congenital nature of the disorder is shown in the following table:

Additionally, many of the concepts in this hierarchy have fully defined stated parents in the “congenital disease” hierarchy, yet also have a redundant OCCURRENCE attribute as well as a congenital morphology. These do not conform to the current modeling pattern of "zero-based" modeling using proximal primitive parents and defining relationships.

While all subtypes of "Congenital anomaly (morphologic abnormality)” have been moved under the "Developmental anomaly (morphologic abnomality)" hierarchy, there still exists a large number of concepts that use "Congenital anomaly" as the ASSOCIATED MORPHOLOGY attribute value.  This morphology, as well as the "developmental anomaly: subhierarchy are problematic as they conflate the aspects of both timing of appearance and structure into a single concept, which due to the lack of a concept model for Morphologic abnormality, cannot reflect the full semantics of these aspects.  This requires additional stated relationships to force both of these to be represented in the modeling of a concept.  Since nearly all of the “Developmental anomaly” morphologies have a non-congenital counterpart.  Since the Morphologic abnormality hierarchy is a subtree of the body structure, it was initially unclear why it had been necessary to represent the timing of appearance of a morphologic feature as an integral part of a single concept when the OCCURRENCE attribute provides this role for other timings of appearance of a particular morphology (e.g. fetal, adult, etc.).  It has since been determined that the more important aspect of these developmental morphologies was to be explicit about the process by which they came to be due to a lack of an appropriate, allowed PATHOLOGICAL PROCESS value

However, because both timing (Occurrence) and morphology can be represented as separate relationships, it is clear from the existing content that editors find it difficult to consistently represent them.  It would be desirable to have a single guideline to consistently represent congenital disorders as well as remove the apparent duplicate morphologies that are there solely to represent a developmental process.

3.5 Subsidiary and interrelated problems

The lack of an approved value for the PATHOLOGICAL PROCESS relationship that would allow the distinction between developmental and acquired morphologies prevents the inactiviation of specific congenital morpholgoies.

4 Risks / Benefits

4.1.1 Risks of not addressing the problem

The risk of not addressing the problem is that congenital content will still be modeled/overmodeled in inconsistent ways that may not allow concepts to be fully defined, not take full advantage of the classifier.

4.1.2 Risks of addressing the problem

The biggest risk in addressing this is that the proposed change is focused on one specific subhierarchy of the terminology.  Clear guidance on where the proposed changes may be applied is required to avoid more extreme variability in outside of Congenital diseases.  It is not expected that the users of SNOMED CT will see any demonstrable effect on their use of the terminology. 

One potential risk would be analysts that have designed their SNOMED CT analytical tools to take this inconsistency into account, although it is anticipated that changes to the inferred relationships would only be an improvement over the current content.

5 Requirements: criteria for success and completion

5.1 Criteria for success/completion

1. Where possible, “Congenital disease” would have no fully defined stated children. (i.e. modeled to proximal primitive parent)

2. All fully defined concepts under 66091009 – Congenital disease would have a stated OCCURRENCE relationship. 

3. Any ASSOCIATED MORPHOLOGY relationship would not use existing concepts under Developmental anomaly.

4. Once 1-3 are complete, the concepts under Developmental anomaly (morphologic abnormality) would be inactivated if they have a non-congenital supertype or moved under a more appropriate ancestor.

5. All concepts with “congenital” in the FSN will be classified under “Congenital disease (disorder).

5.2 Strategic and/or specific operational use cases

Overall quality improvement of SNOMED CT has become increasingly important.  This involves primarily a more standardized approach to the modeling of concepts in a particular space of SNOMED CT.  This project addresses one small area and can be used as a guide for how to attempt a more comprehensive application of these principles through out the terminology.

5.2.1 Use case 1

Provide all SNOMED CT concepts for Congential disorders of the X system.  Consistent application of modeling patterns will allow high levels of recall and precision.

5.2.1.1 Fit with IHTSDO strategy

Problem List

5.2.2 Use case 2…

5.2.2.1 Fit with IHTSDO strategy

From IHTSDO Content Product Development Plan

Identify relevant test cases

Choose a subhierarchy under Congenital disease and apply these criteria to those concepts

6 Outline Possible Technical Approaches and Concept Model

Explain what the problem solution is likely to entail

This approach will require a systematic review of each concept under the top level concept "Congenital disease".  It will require evaluation of the modeling style for adherence to the current modeling guidelines and the appropriate assignment of values for defining relationships to ensure appropriate classification 

Will concept model change be needed?

This proposal expands on the initial proposal to simply reassign morphologies and addition of the OCCURRENCE attribute where needed.  It requires the expansion or the range of the PATHOLOGICAL PROCESS attribute to allow for the use of 

308490002 | Pathological developmental process (qualifier value)

Will terming guidance and rules be needed?

No

6.1 Initial design

6.1.1 Outline of initial design

The general approach for this project would be to replace all existing ASSOCIATED MORPHOLOGY relationships that use concepts under 107656002 – Congenital anomaly (morphologic abnormaity) and 21390004 | Developmental anomaly (morphologic abnormality) with appropriate concepts that do not include the “congenital” timing aspect.  For each term where this is done, it will be verified that the OCCURRENCE =  Congenital relationship exists.

Include the OCCURRENCE attribute in the same role group as ASSOCIATED MORPHOLOGY and FINDING SITE.  The appearance of the morphology is tied to the actual structure, so these two attributes should be linked in the same role group.  For example:

Stated:                                                                                                                                           Inferred:

Open

Open

Following this initial cleanup, each term in the “Congenital disease” hierarchy will be evaluated and potentially remodeled to its proximal primitive parents and appropriate relationships to allow for maximal use of the classifier to assign inferred parentage.

6.1.2 Significant design or implementation decisions / compromises

6.1.3 Evaluation of Design

6.1.3.1 Exceptions and Problems

it was found that in many cases, the use of the non-congenital morphology required the addition of a second role group that explicitly stated that this was a developmental pathology by adding the ASSOCIATED MORPHOLOGY = Developmental anomaly in order for the classifier to position it correctly within the hierarchy.  Thus the need in many cases for two role groups to specify the process that resulted in the disorder.

6.1.3.2 Design Strengths

A consistent modeling representation and the ability to inactivate a small number of "Congenital X (morphologic abnormality)" concepts

6.1.3.3 Design Weakness

The need for two role groups in some cases to allow for proper classification can lead to inconsistencies when the incorrect inferences are not detected by the editor.  It also adds an additional modleing burden to make the determination as to when the additional relationship group is needed.

6.1.3.4 Design Risks

Initial testing of this approach on a substantial number of terms has found that there is no change to the inferred parents for concepts when the “non-congenital” morphology is assigned where the term is exclusively a morphology that occurs during development (i.e there is no "acquired" equivalent).  However, the inability to inactivate a number of the "Congenital X" morphologies without the addition of a second role group leads to additional editing and maintenance burdens.

6.2 Iteration One

6.2.1 Outline of revised design

Redesign the Solution Identify objectives of iteration, and the major changes to previous design

The objective of this iteration is to remove the need for separate subhierarchies of morphologic abnormalities that are either present atr birth or may be acquired.  This will enable the accoplishment of the original objective, to retire "Congenital X (morphologic abnormality)" concepts as well as resolve the need for multiple relationship groups to represent developmental structural anomalies.

6.2.2 Significant design or implementation changes

1.  Extend the range of PATHOLOGICAL PROCESS  to include 308490002 | Pathological developmental process (qualifier value) |

2.  Replace all occurrences of ASSOCIATED MORPHOLOGY values containing "107656002 | Congenital anomaly (morphologic abnormality) |" or "21390004 | Developmental anomaly (morphologic abnormality) |" with "49755003 | Morphologically abnormal structure (morphologic abnormality) |"

3. Where possible, reassign the ASSOCIATED MORPHOLOGY values using "Congenital X (morphologic abnomality)" with the non-congenital supertype.  In some cases this is not possible as there is no appropriate (i.e. non-congenital) supertype.

4. Ensure that, where present the FINDING SITE, OCCURRENCE, PATHOLOGICAL PROCESS and ASSOCIATED MORPHOLOGY relationships are in a single relationship group.

5. Remove redundant relationship groups that were added in the first iteration to ensure proper classification of congenital diseases.

6. Revise, where necessary current modeling structures to the "Closest proximal primitive" modeling style adopted by editorial policy as well as reviewing the specificity of existing relationships to ensure that all concepts that can be sufficiently defined have been. 

6.2.3 Evaluation of Revised Design

6.2.3.1 Exceptions and Problems

This iteration of the design only addresses revisions to concepts that are subtypes of 276654001 | Congenital malformation (disorder) | 

This use of the PATHOLOGICAL PROCESS relationship design is ONLY applicable to concepts falling under 276654001 | Congenital malformation (disorder) |, which is an immediate subtype of "Congenital disease".  For other types of congenital diseases, such as "Congenital infectious disease", "Congenital metabolic disease", "Congenital traumatic disease", etc. Alternative modeling patterns may need to be applied.  This will be developed in a future iteration of this project due to the size of the project as a whole.

6.2.3.2 Design Strengths

Since disorders may be congenital, but not developmental anomalies (e.g. congenital infectious disease), or developmental anomalies, but not congenital (e.g. permant tooth development), or both, this separation of process and morphology allows for better representation of all classes of congenital or developmental disorders. 

6.2.3.3 Design Weakness

This current design only addresses one subhierarchy of the Congenital disease hierachy and does not explicitly address the overlap between Genetic and hereditary diseases and congenital diseases.

6.2.3.4 Design Risks

7 Recommendation

1. Implement the solution as recommended in Iteration one.  Request extension of the PATHOLOGICAL PROCESS attribute.  This has been reviewed by the IHTSDO Editorial Advisory Group and approved in principle.  

2. Develop editorial guidance for the use of the new PATHOLOGICAL PROCESS relationship ensuring that it is only currently allowed for use on subtypes of "Congenital malformation"

7.1 Detailed design final specification

See details listed in section 6.2.2

7.2 Iteration plan

8 Quality program criteria
8.1 Quality metrics

8.1.1 Quality metric 1

8.1.2 Quality metric 2

8.2 Use case scenarios

Create Test Cases: Review the requirements to be tested as set out in the Inception Phase document. Identify and outline relevant Test Cases. Identify test data needs. Share and evaluate the Test Cases

8.2.1 Scenario One

8.2.1.1 Expected Setting

8.2.1.2 Data capture requirement

8.2.1.3 Data retrieval requirement

8.2.2 Scenario Two

8.2.2.1 Expected Setting

8.2.2.2 Data capture requirement

8.2.2.3 Data retrieval requirement

8.2.3 Scenario …

8.3 Test cases

9 Project Resource Estimates

Estimate project size; Forecast project velocity and duration\

Project size for this iteration (subtypes of Congenital malformation) is approximately 5500 concepts.  A skilled editor would be able to revise 20-60 concepts per day. The estimate for the amount of effort would therefore be 3 -6 months of effort for a dedicated editor without the applicaiton of batch changes.  The ability to do batch changes to faciliate the groupings and mass replacement is being investigated.

Evaluate risks; Establish costs and articulate value; Plan deployment; Outline project lifecycle

This project will be undertaken by the Head of Terminology as a continuation of the Consultant Terminologist project begun in 2012.  Deployment will be as a complete revision from a separate branch of the terminology authoring environment.  

9.1 Scope of construction phase

Optionally:

Skills required

Solution Specification (Elaboration)

Implementation

Outline of work packages

Preventing recurrence of problem

Division of project into stages

9.2 Projection of remaining overall project resource requirements

9.2.1 Expected project resource requirement category

Is it “fast track” or does it require project management?  It is neither a fast track nor a project management related project.  It is part of the continuous quality improvement efforts underway for the terminology as a whole.

9.2.2 Expected project impact and benefit

Allows for removal of redundancy in the Morphologic abnormality hierarchy, which will improve consistency in future modeling of Congenital diseases as well as improving the internal consistency and quality of a subhierachy of SNOMED CT.

9.2.3 Indicative resource estimates for construction, transition and maintenance:

Construction and transition phase:     <100 new concepts to be authored

Maintenance phase:    <100 new ‘frequent usage’ concept requests in 1^st 3 years

Estimate project size; Forecast project velocity and duration

Project size for phase 1 is approximately 5500 existing concepts.  Velocity of the project is dependent on the ability to use batch processing for explicitly defined changes

10 Appendices

10.1 Appendix One : Related user requests

Related to a number of IHTSDO content tracker items related to Congenital disease.

artf227329 : Meaning of Congenital rubella infection (disorder)

artf222646 : Review concept: Congenital vascular anomaly

artf222527 : Add Parents to: Congenital trigger thumb

artf222479 : Concept model: Congenital trigger thumb

artf222431 : Concept additions to Congenital anomaly of peripheral blood vessel

artf6288 : definition and clarification of familial, genetic, and inherited, plus congenital / acquired and related