General Guidelines1

(Version: 07/12/05)

Definition:  The general guidelines are those that apply in all coding situations.


Introduction:  These general guidelines are not meant to replace the College of American Pathologists (CAP) SNOMED CT Users Guide ‘rules’; they are meant to assist new veterinary users of SNOMED CT in order that data may be coded in a consistent and useable manner.


Aids and Guidelines:

1.     As they are developed, specific Guidelines will be posted to the VMDB website at  (SNOMED CT Usage Guidelines).  Guidelines will be reviewed and updated annually by the VMDB Taskforce on SNOMED Guidelines.   These are meant only as a guide to assist the user and are subject to change as technologies develop.

2.     Also available on the VMDB website are the Veterinary Coding Protocols (VCOPS).  Always search VCOPS first before choosing a SNOMED concept to ensure all veterinary users are coding common diagnoses/procedures in a consistent manner.   VCOPS  maintains a history of codes previously used; tracks codes requested on the Forum; assists the user in finding non-intuitive concepts (slang), provides guidance on post-coordinated concepts or incorrectly modeled ones, and assists the user in choosing a code with similar concepts.

3.      Discussion Forum:  Utilize the SNOMED Forum for discussions about SNOMED CT concepts, functionality, and syntax so that ‘everyone’ can learn from the questions and discussions.  The forum is maintained by Dr. Jeff Wilcke and may be found at the VMDB website or at  The Forum maintains a historical record of discussions that can be used as a searchable reference.   When posting a question, the topic should be the concept you are discussing.  Always include a definition or resource for a definition when requesting a new term or more complete modeling.

4.     The SNOMED CT Users Guide from CAP is a good resource for basic SNOMED CT principles.

5.     The CLUE browser from CAP is a good resource to see exactly how a term is modeled, as well is a resource for alternative terms and/or their hierarchy.


General Coding Comments:

A concept is a clinical term assigned a unique concept ID (a numeric value) in SNOMED CT.   By itself this numeric value has no meaning.  (The SNOMED ID is a legacy code that should not be used for data storage or transmission).

            Example:  53084003 = Bacterial pneumonia (disorder)

Concepts are related to one another by an ‘is-a’ type or an ‘attribute’.   See the Users Guide for explanations concerning relationships between concepts and their hierarchy.

Many concepts have descriptions in their hierarchy with Synonyms listed under these descriptions.   Do not rely on Synonyms to define a concept.  Many are inaccurate and are under review.   A synonym is helpful to locate the term but does not necessarily mean the same as the actual ‘Preferred Concept’.

·       Example:  Pinworm Disease takes the user to the concept Enterobiasis (266162007) which is a human pinworm species.  This does not meet veterinary needs even though the synonym would make it appear as though it does.


1.     Begin coding with a ‘root concept’ from the Clinical Finding hierarchy or the Procedure hierarchy.  NEVER begin with a morphology concept.

·       When in doubt as to the top level (or root concept) of the hierarchy, use the more generalized parent which is defined by the “is-a” term.

2.     Choose a concept from the appropriate hierarchy

·       an avulsion can be a ‘morphology’ or  ‘procedure’ or ‘disorder’

·       A ‘Procedure’ should be chosen from the ‘Procedure hierarchy’.

3.     Concept Status:  indicates the status of a concept such as ‘current’, ‘retired’, ‘limited’, ‘duplicate’, ‘ambiguous’,  ‘erroneous’, etc.    Do not use concepts that are retired, duplicated, erroneous, or ambiguous.   If a concept needed is listed as ‘limited’ it may be used but should be posted to the Forum to notify SNOMED that the term needs to be completely modeled or a new term created.  Use ‘limited’ concepts only if you cannot capture the information using a current concept or through post-coordination.  Note:  VMDB will reject records coded with ‘erroneous’ terms.

4.     Be careful using species-specific concepts unless necessary as future modeling may retire these concepts).

5.     Choose a ‘disorder’ concept for diagnoses whenever possible.    Example:  Vomiting is listed as both a disorder and finding in SNOMED.  If a disorder concept does not exist then use the finding unless otherwise directed in VCOPS.


Post-coordinating is taking an already existing concept and enhancing it using a combination of two or more codes.  (For example, Mycoplasma Bacterial Pneumonia does not exist as a pre-coordinated concept, thus, several concepts would need to be combined to build this term.)

1.     Attributes are used to post-coordinate concepts.   Each value must have its own attribute, thus multiple finding sites each need their own attribute of finding site. (Refer to the SNOMED Users Guide provided by CAP for those attributes allowed and further definitions on their use). 

Example:   Malignant Neoplasm of Shoulder and of Chest (disorder)*

·       Malignant Neoplastic Disease (disorder)

o      Finding Site (attribute)

§       Skin structure of shoulder (body structure)

§       Laterality (attribute)

·       Right (qualifier value)

o      Finding Site (attribute)

§       Skin of chest (body structure)

*Note:  Only use the laterality attribute with anatomy structures that are symmetrical.   And, link the attribute and value only to the finding site that you are lateralizing.

2.     Be cautious with the use of attributes so as Not to change the meaning of a diagnosis.    Example:   use ‘causative agent’ when it is known that a disorder is caused by the organism or substance.   See ‘attribute’ definitions and usage in the SNOMED CT Users Guide.

3.     AT ALL TIMES, the ‘rule of odds’ apply.    The entry of codes should always result in an odd number (e.g. if post-coordinating beyond a single term, attributes must be used to tie codes together).

Example:  Mycoplasma bacterial pneumonia is defined as

·       Bacterial pneumonia (disorder)

o      Causative agent (attribute)

o      Mycoplasma (organism)


4.     Order is very important in post-coordinating.   The root concept is always followed by the attribute and its value.  Thus, one would NOT say:

·       Mycoplasma (the organism)

a.      Has ‘causative agent’ Bacterial Pneumonia

·       NOR, Lung structure (body structure)

a.      Has ‘finding site’ Bacterial Pneumonia

Structurally these makes no sense.

5.     When post-coordinating, it is helpful to locate an already existing defined concept that is similar to the diagnosis you are attempting to capture to use as a model and mimic it.  Consistency in data entry will increase the chances of pulling all the correct data during retrievals.

o      Example:  No pre-coordinated term exists for “Pneumonia due to bacilli”.  Following the rules it would be modeled similarly to ‘Pneumonia due to staphylococcus’ by using  Bacterial pneumonia (disorder) using the attribute ‘causative agent’and the organism bacilli.

6.     Be cautious when post-coordinating so that redundancy does not occur.  If a concept is already built into the definition of the root concept then it would be redundant to post-coordinate:

Example:  Open comminuted fracture femur is diagnosed.


·       Open fracture of femur (disorder)

o      Associated morphology (attribute)

o      Open comminuted fracture (morphologic abnormality)*

*because open fracture is already part of the disorder


·       Open fracture of femur (disorder)

o      Associated morphology (attribute)

o      Comminuted fracture (morphologic abnormality)

However, some redundancy may be unavoidable.  With the disorder ‘Fracture of femur’ if you wanted to say that it was part of the distal femur you would need to add a ‘finding site’ of ‘Distal femur’ to make it more complete.



7.     Be cautious when building post-coordinated terms that the meaning can not be misconstrued.  See the following example using multiple finding sites:

Example 1*:

·       Open wound (disorder)

o      Finding site (attribute)

§       Back structure (body structure)

o      Finding site (attribute)

§       Lower back structure (body structure)

*Note:  Does this mean there are 2 wounds or is this an attempt to be more specific?

Example 2**: 

·       Open wound (disorder)

o      Finding site (attribute)

§       Multiple topographic sites (body structure)

o      Finding site (attribute)

§       Hindlimb (body structure)

**Note:  Does this mean multiple open wounds and one on the hindlimb?

8.     In general, use of morphology concepts are limited to diagnoses taken from pathology reports (necropsy, biopsy, cytology).    Otherwise use a pre-coordinated concept.

9.     Be cautious when coding from culture reports.    Not all organisms found on culture are necessarily the causative agent of a disorder and may be ‘normal flora’.  Each institution needs to follow their own internal policy in coding from these.


 Topography (Body Structure):

1.     Topography concepts are chosen from the Body Structure Hierarchy and post-coordinated using the attribute ‘finding site’ for Clinical Findings and the attribute ‘procedure site’ for Procedures.

2.     Be cautious when choosing body structure concepts for veterinary usage.   Veterinary anatomy is continually being developed and there are some inaccuracies.

3.     Use ‘non-human’ body structures when possible. 

4.     Many concepts are modeled to human body structures, so when choosing one of these avoid species-specific such as ‘4th rib’.

5.     Although hindlimb and forelimb are available as body structures, they are rarely available when choosing an already pre-coordinated term, e.g., Malignant neoplasm of upper limb would need to be used for forelimb.

6.     Use body part structure but NOT entire unless it affects the entire body structure (e.g., entire liver structure would include all lobes, vessels attached to the liver, etc.)

7.     Specificity (granularity) in coding a finding site should be decided by the individual institution. 

8.     It is preferred to not use the ‘Multiple topographic sites’ concept code.  Rather, code each finding site separately in order to make the data retrievable and useful.  Use the multiple topographic sites concept only when sites are not defined in the record.



1.     Stay away from the general species codes under the ‘domestic mammal’ hierarchy, e.g. equine species (26570006) and bovine species (79058000).   These concepts have no children and will most likely be developed as ‘occupational usage’codes at some later date.     Use the actual genus species concepts such as Equus caballus (35354009) and Bos taurus (34618005).

2.     Notify the Forum if a breed or species needed is not available.  Include a taxonomy reference in posting when possible.  

3.     SNOMED is creating some ‘superbreed’ concepts for those breeds where it is not always clear in the medical record as to the specificity of the breed.    For example, ‘Dachshund’ is a superbreed of Miniature Dachshund, Longhaired Standard Dachshund, etc.

4.     VMDB is one resource for breed and species codes,  and breed synonyms.  

5.     Another resource used by the SNOMED Forum team for modeling taxonomy is the USDA taxonomic site at


VMDB Submissions (also see VMDB Submission Rules at

1.     Submit only Diagnoses and Procedure terms to VMDB at this time.    Do not submit exam findings from the physical examination that are not diagnosed.

2.     Submit only SNOMED CT sanctioned codes.  Records submitted with non-sanctioned codes will be rejected and sent back to the submitting institution for correction and resubmission.

3.     If submitting recheck diagnoses, flag the diagnosis as a recheck (see VMDB Transmission File Layout scheme and Recheck Guidelines).

4.     Diagnoses may be coded as ‘confirmed’ or ‘suspect’ or ‘probable’.    This designation is given at the concept level, Not, visit level.  VMDB will assume the diagnosis is confirmed unless submitted as suspect (see VMDB Transmission File Layout scheme for flag/field designee information).


·       Malignant neoplasm upper limb (disorder)  - Confirmed

§       Associated morphology (attribute)

§       Chondrosarcoma (morphologic abnormality) - Confirmed

o      Due to (attribute)

o      Drug-induced lesion (disorder) - Suspect

5.     Change in diagnosis:   When a diagnosis is made and is later discovered to be incorrect, that visit should be ‘recoded’ and resubmitted to VMDB to ensure data validity

·       Example:  Initial diagnosis is ‘malignant neoplasia’.  Later it is found to truly be benign and the original record reflects the change with an addendum to the original diagnosis by the clinician.

6.     Post-coordinated terms need to be transmitted in the appropriate sort order


·       Causative agent (attribute)

o      Pseudomonas (organism)

o      Pneumonia (disorder)


·       Pneumonia (disorder)

o      Causative agent (attribute)

o      Pseudomonas (organism).

7.     Do not submit context dependent category codes such as ‘History of’, ‘Family history’, ‘Rule/Out’,  or ‘Differential diagnosis’.  These are used to explain the context of a chosen concept and should be collected in a separate field in the institutions database but not submitted to VMDB at this time.  

8.     In general, do not use ‘qualifer values’ or ‘observable entities’ unless noted in guidelines.

9.     A procedure can be coded and submitted WITHOUT a diagnosis.

10.  Data should be submitted to VMDB for those visits in which the animal was presented to the institution for evaluation or care only (e.g., this excludes lab only, pharmacy only, field service/ambulatory visits, over the counter sales of drugs or supplies, non-client related teaching and research visits).

11.  Also do not submit DOA (dead on arrival) patients, with no clinical intervention, to VMDB. 

12.  Transmit only those groups of codes in which the ‘Rule of Odds” has been applied.










1Written by the VMDB Taskforce on Development of SNOMED Guidelines.  The committee wishes to recognize contributions provided by Dr. Jeff Wilcke and Dr. Penny Livesay