Source document located: Content Project Tracker Documents (note version adjusted when uploaded from 1.0 to 0.1)

1 Glossary

1.1 Domain Terms

2 Introduction

2.1 Purpose

The purpose of this project is to describe content issues reported in the subdomain of immunization procedures.

SNOMED CT projects transition from Inception Phase à Elaboration Phase à Construction Phase à 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.

Subject to adequate review by stakeholders and subsequent revision, the inception phase document becomes the primary input to the Elaboration Phase of the project, in which the potential solutions are considered.

2.2 Audience

The audience for this document includes all standards terminology leaders, implementers and users but is especially targeted at those stakeholders from the primary care and immunization programme domains.

2.2.1 Identification of stakeholders

The issue was originally reported by the UK National Release Centre; much of the affected content derives originally from Clinical Terms Version 3 where it had been evolved and used over many years to support national child immunization programmes.

IMMFORM is a division of Public Health England charged with collecting, recording and analyzing data in relation to uptake against national immunisation programmes.

2.2.2 Input from stakeholders

2.2.3 Degree of consensus on the statement of problem

3 Statement of the problem or need

3.1 Summary of problem or need, as reported

In April 2010, Kent Spackman added an Elaboration-type document to a tracker item (Appendix One : Immunizations), including a high-level summary of problems reported in the immunization hierarchy:

1. Immunization procedures are generally primitive and therefore do not subsume all appropriate children
2. Manual assignment of children has not been accurate
3. Definitions do not capture the entity (organism) against which immunization is directed.  

The tracker item itself had been raised one month earlier (March 2010) in relation to several requests from the UK National Release Centre concerning SNOMED CTs vaccination and immunization content. Originally, only the following summary was provided (abridged here from original):

Clinically, the terms immunization and vaccination are often used interchangeably wherever immunization against a particular condition is usually achieved in the majority of the population by administration of a vaccine. However, strictly they’re not the same: vaccination is the act of administering a vaccine; immunization the intended outcome. But immunization as an outcome can be achieved not only through vaccination but also through passive - possibly transient - immunization, typically by an infusion of immunoglobulin or maternal immunoglobulin transfer.

For this reason it is technically incorrect that 398856000|Bordetella vaccine (substance)| is given as the direct substance in the procedure 8605003|Immunization for bordetella (procedure)|: temporary, passive immunization could have been achieved without a vaccine but rather by immunoglobulin infusion.

Should we therefore change the FSN of the existing 8605003|Immunization for bordetella (procedure)| to "Vaccination for bordetella (procedure)” and define it accordingly since, in practical usage, the term means 'active immunization’? Or should we simply recommend post-coordination of a subtype as 33879002|Administration of vaccine to produce active immunity (procedure)| by refining the Direct substance, (in this case, 398856000|Bordetella vaccine (substance)|).

Submitter says their system does not handle role-grouping well, and they have been simply post-coordinating a direct substance of the particular vaccine without role-grouping it with an action of administration.

Also, currently 33879002|Administration of vaccine to produce active immunity (procedure)| is primitive. Can we fully define this, given that Stedman defines “vaccine” as “essentially any preparation intended for active immunologic prophylaxis; e.g., preparations of killed microbes of virulent strains or living microbes of attenuated (variant or mutant) strains; or microbial, fungal, plant, protozoal, or metazoan derivatives or products.”

Should we pre-coordinate a new concept for the submitter request?

3.2 Summary of requested solution

The original requestor suggested that a full solution would involve:

…a much more compositional model where the vaccination hierarchy is entirely defined rather than primitive entities, being at a minimum defined (ultimately) in terms of the organism(s) vaccinated against (which should be separate from the preparation(s) used) and the vaccination schedule dose given, and refinable by e.g. the site of the injection

Spackman’s elaboration makes more detailed recommendations; few had been enacted by 2015:

3.3 Statement of problem as understood

The current SNOMED CT taxonomy (July 2015 International Edition) exhibits the following issues in relation to its modelling and classification of immunization procedures. Remedies suggested in 2010 have been only partially applied.

3.4 Detailed analysis of reported problem, including background

3.4.1 Conflation of immunization with vaccination

The tendency to consider vaccination procedures as universally (rather than only normatively) synonymous with immunization was originally reported in the context of 8605003|Immunization for bordetella (procedure)|. In fact, this concept has never had a modeled relationship of direct substance=398856000|Bordetella vaccine (substance)|; the original complaint therefore may have been based on an expectation that any concept labelled ‘immunization for bordatella’ should have been a subtype of ‘active immunization’ because, currently and normatively, that is the only method by which such immunity can be conferred. However, other concepts carrying the more generic preferred term ‘immunization’ do exist with a direct substance relationship pointing specifically at a vaccine e.g. 243789007|Hepatitis A immunization (procedure)|:direct substance= 396423004:|Hepatitis A virus vaccine (substance)|.

This combination of terming and modelling implies not only that ‘immunization against Hepatitis A’ has always been and will only ever be possible by active immunization and correspondingly that no mechanism for passive immunization against Hepatitis A could ever exist, but also that even to merely contemplate the idea of ‘passive immunization against Hepatitis A’ is as semantically invalid as would be the notions of ‘mammalian fish’ or ‘myocardial infarction of the brain’.

Although the continued presence of both 359950007|Booster immunization (procedure)| and 359953009|Booster vaccination (procedure)| as separate, mutual siblings might suggest that ‘immunization’ is not always considered synonymous with ‘vaccination’, the conflation is still clearly pervasive. For example, the descendants of 33879002|Active immunization| include 70322007|Immunization in infancy (procedure)|, 91223004|Emergency immunization during epidemic (procedure)|, 359950007|Booster immunization (procedure)|, 312871001|Immunization against bacterial disease (procedure)| and 49083007|Viral immunization (procedure)|, thus apparently discounting the possibility that immunization in each case may be achieved passively.

Although pervasive, the conflation is not consistently applied:  49083007|Viral immunization (procedure)| - another primitive child of 33879002|Active immunization (procedure)| - has direct children including both 117089007|Administration of Respiratory Syncytial virus immune globulin, human (procedure)| and 117095008|Administration of Vaccinia immune globulin, human (procedure)|, which are both  of immunization. Further, because of this misclassification, both of these passive immunization procedures also currently erroneously inherit the modelled relationship of direct substance= 418310008|Viral vaccine (substance)|.

1.1.1 Immunization hierarchy mostly primitives

In the July 2015 International Edition, 127785005|Administration of substance to produce immunity, either active or passive (procedure)| currently has 234 descendant concepts, 80% of which are primitive. 82% of 119 concepts below 312871001|Immunization against bacterial disease (procedure)| are primitive, and 79% of 111 concepts below 49083007|Viral immunization (procedure)|. Altogether, fewer than 20% of all active immunization procedure concepts are defined entities.

3.4.3 Hierarchy contains many errors

127785005|Administration of substance to produce immunity, either active or passive (procedure)| is itself a primitive concept but with a modeled relationship Direct substance=Vaccine, immunoglobulin, AND/OR antiserum. 51116004|Passive immunization (procedure)| is correspondingly a primitive child of 127785005 but therefore still inherits from this parent the Direct substance=Vaccine, immunoglobulin, AND/OR antiserum modeled relationship. This looks odd since it suggests that some forms of passive immunization might be performed using a vaccine. This issue could be corrected if the taxonomy below 418039004|Vaccine, immunoglobulin, AND/OR antiserum (substance)| was subdivided by the introduction of an ‘immoglobulin and/or antiserum’ grouper, as the disjunctive complement sibling of 398827000|Vaccine (substance)|.

51116004|Passive immunization (procedure)| does not currently subsume all codes that properly correspond to forms of passive immunization, including 275844006|Gamma globulin administration (procedure)|.

Many “missed classification” errors are evident within the taxonomy of 212 concepts below 33879002|Active immunization|, including but not limited to:virus vaccination (procedure)| should be a 49083007|Viral immunization (procedure)|

170392001|Leishmaniasis vaccination (procedure)| should be a 312871001|Immunization against bacterial disease (procedure)|

30338008|Typhus vaccination (procedure)| should be a 312871001|Immunization against bacterial disease (procedure)|

312872008|Combined immunization against bacterial and viral disease (procedure)| should be a 49083007|Viral immunization (procedure)|

312872008|Combined immunization against bacterial and viral disease (procedure)| should be a 312871001|Immunization against bacterial disease (procedure)|

312867004|Typhoid and paratyphoid vaccination (procedure)| should be a 83798008|Paratyphoid fever vaccination (procedure)|

414001002|Diphtheria tetanus and five component acellular pertussis, haemophilus influenzae type b, inactivated polio vaccination (procedure)|

should be a 268499008|Tetanus and polio vaccination (procedure)|

should be a 73152006|Tetanus diphtheria vaccination (procedure)|

414259000|First diphtheria tetanus and five component acellular pertussis, haemophilus influenzae type b, inactivated polio vaccination (procedure)|

should be a 414001002|Diphtheria tetanus and five component acellular pertussis, haemophilus influenzae type b, inactivated polio vaccination (procedure)|

3.4.3.1 Classification of multivalent vaccination

The current vaccination content includes concepts for procedures involving multivalent vaccination preparations (ie administration of a single vaccine preparation capable of inducing an immune response in the recipient against more than one microorganism), for example:

268499008|Tetanus and polio vaccination (procedure)|

270898008|Diphtheria and tetanus double and polio vaccination (procedure)|

As with the rest of the vaccination content, these are generally primitive.

In determining the correct manual classification of these multivalent vaccination procedures, therefore, the authors must decide whether ‘immunization against A and B’ should be interpreted as ‘immunization against A and B and nothing else’ or as ‘immunization against at least A and B’. In general, the manually asserted hierarchy appears to have been constructed using the latter interpretation, so that ‘immunization against A and B and C’ is mostly classified as a subtype of ‘immunization against A and B’. There are, however, exceptions, for example:

414001002|Diphtheria tetanus and five component acellular pertussis, haemophilus influenzae type b, inactivated polio vaccination (procedure)|

should be a 268499008|Tetanus and polio vaccination (procedure)|

should be a 73152006|Tetanus diphtheria vaccination (procedure)|

3.4.4 Relationship between immunization procedures and disease/organism

Although all active immunization procedures are modelled with a direct substance relationship pointing at an appropriate vaccine substance, the latter vaccine substance itself is always a primitive. As a result, immunization procedures intended to confer immunity specifically against infectious conditions have no modelled relationship - or chain of such relationships –linking them to the disease(s) to be prevented or to their pathogenic organism.

Such a link would be desirable (possibly even required) if the concepts below 312871001|Immunization against bacterial disease (procedure)| and 49083007|Viral immunization (procedure)| were to be autoclassified.

3.4.5 Parts of multi-part vaccination schedules

The current immunization procedure taxonomy includes 29 concepts for booster doses (also occasionally called ‘reinforcing doses’), and a similar number for ‘first vaccination’; the latter is assumed to mean the first part of a multi-part schedule for primary immunization. Smaller numbers of concepts similarly exist for second, third, fourth and even fifth parts of the primary immunization schedule.

Two particular problems arise from the current approach:

1)    If the locally endorsed immunization schedule changes, possibly to include an additional part, then it would be reasonable to expect the SNOMED CT concept model to offer a postcoordinated solution by which this new part could be defined, and so avoid the often lengthy delay between requesting new codes and getting them added to the international edition (or a national extension) before the modified schedule can be rolled out and documented.

2)    For a variety of reasons, though most commonly occurring during national transitions to new multivalent vaccine preparations and schedules, most practical vaccination uptake monitoring programmes will experience patients for whom the administration of a particular multivalent vaccine dose will not be functioning as the same part of the primary immunization schedule for each organism involved. Thus, for example, a patient who first receives the following:

170395004|First diphtheria, pertussis and tetanus triple vaccination|

170396003|Second diphtheria, pertussis and tetanus triple vaccination|

…but, on their third attendance, is offered

312870000|Diphtheria, tetanus, pertussis, polio and haemophilus influenza B vaccination|

…is NOT receiving a

310308006|Third DTP polio and Hib vaccination|

..but in fact all of:

170397007|Third diphtheria, pertussis and tetanus triple vaccination|

170343007|First haemophilus B vaccination|

170353008|First polio vaccination|

Such single ‘same scheduled part’ codes for multivalent vaccine administrations certainly provides significant convenience at the point of clinical data capture, but this may no longer be appropriate now that national vaccination programmes have become considerably more dynamic than they were in the 1980s and 1990s when the current terms were devised; in certain circumstances, they make it harder to work out who really has completed their primary immunization schedules.

3.4.6 Boosters

Spackman’s original analysis suggested that 359950007|Booster immunization (procedure)| could be retired as a duplicate of 359953009|Booster vaccination (procedure)|. An alternative approach would be to move 359950007|Booster immunization (procedure)| so that it was no longer a child of 33879002|Active immunization (procedure)| but became a child of 127785005|Immunisation| with 359953009|Booster vaccination| as its own child. This change would recognize the possibility of passive treatments to boost an individual’s immunity.

3.4.7 Different forms of passive immunization

3.4.7.1 Antitoxins

Procedures to administer bacterial antitoxins (tetanus, botulinum, diphtheria etc) are logically types of passive immunization against a bacterial disease; 71862009|Antitoxin (substance)| itself is a descendant of 68498002|Antibody (substance)|. (Though, note that 68498002|Antibody (substance)| is not itself a subtype of any ‘immunoglobulin’ ancestor).

However, 27513008|Administration of antitoxin (procedure)| is not currently autoclassified as a subtype of either 51116004|Passive immunization (procedure)| or of 48556009|Administration of immune serum (procedure)| because both are primitive entities, and the latter has no modeled Direct substance=Antiserum relationship.

3.4.8 Autogenous vaccination/immunization

Autogenous vaccines are preparations made from the patient's own microorganisms, sometimes from their part-treated urine. It is a seldom-used therapy; the medical literature is limited but describes uses in the treatment of chronic refractory bacterial infection (furunculosis, osetomyelitis etc) and also possibly for allergy.

359959008|Autogenous vaccination (procedure)| and 359956001|Autogenous immunization (procedure)| are siblings, but both are subtypes of 33879002|Administration of vaccine to produce active immunity (procedure)|. This hierarchical positioning is consistent with the literature, which implies that a passive form of autogenous immunization is impossible. However, if we follow through with this logic, then  359956001|Autogenous immunization (procedure)| is a redundant concept and could be inactivated SAME_AS 359959008|Autogenous vaccination (procedure)|.

3.5 Subsidiary and interrelated problems

3.5.1 Vaccine programme management

384810002|Immunization/vaccination management (procedure)| is a primitive, but carries a Has focus=127785005| Immunisation (procedure)|modeled relationship. If it were made a defined entity, and the same modeled relationship added to all of :

308433002|Immunization call (procedure)|

171044003|Immunization education (procedure)|

308432007|Immunization recall (procedure)|

268558004|Immunization status screening (procedure)|

702972007|Counseling for influenza immunization (procedure)|

702971000|Counseling for childhood immunization (procedure)|

…then these would all autoclassify correctly below it. It would also become possible to create postcoordinated expressions to record call/recall activities in relation to specific immunization management procedures due.

4 Risks / Benefits

4.1.1 Risks of not addressing the problem

SNOMED CT’s current content and taxonomy is incorrect in a number of significant areas relating to immunization, with consequences to both the management of individual patients within national immunization programmes, and aggregated reporting for the purposes of monitoring the same programmes. These programmes cover public health and preventative medicine activities that are experienced by almost all citizens and so these content failings have very high functional and reputational impact.

4.1.2 Risks of addressing the problem

Most of the likely content changes would be intended to improve the existing taxonomy, and the extensibility of the content in support of more rapidly changing national immunization programmes.

Possible changes to the representation of schedule-parts, however, could include e.g. retirement of the existing ‘same part’ code content and the recommendation that all future clinical recording should separately record the schedule part for each organism, including in the majority of instances where a single multivalent vaccination procedure does, indeed, function as the same schedule part for all organisms involved. Such a change, however, would have the potential to seriously disrupt any live vaccine uptake monitoring programmes and so should be handled very carefully.

5 Requirements: criteria for success and completion

5.1 Criteria for success/completion

Technical criteria

1. All descendants of 127785005|Administration of substance to produce immunity, either active or passive (procedure)| shall be defined entities with modelled relationships that link them to the active organisms or other antigens for which a trained or translanted immune response is to be achieved, and separately to the disease entities to be protected against
2. Classification of all immunization related procedures to be achieved fully automatically
3. A coherent and extensible model for modelling immunization schedule parts and booster doses

Operational criteria

1. No adverse impact on established mass immunization programmes

5.2 Strategic and/or specific operational use cases

5.2.1 Use case 1

Medicolegal documentation of vaccine administration procedures

5.2.1.1 Fit with IHTSDO strategy

1F. Public Health, including notifiable disease and organism reporting (public health surveillance)

5.2.2 Use case 2

Management and monitoring of mass vaccination programmes (call, recall, education)

5.2.2.1 Fit with IHTSDO strategy

1F. Public Health, including notifiable disease and organism reporting (public health surveillance)

5.2.3 Use case 3

Monitoring of vaccine efficacy (by populations that contract diseases despite vaccination)

5.2.3.1 Fit with IHTSDO strategy

1F. Public Health, including notifiable disease and organism reporting (public health surveillance)

6 Outline Possible Technical Approaches and Concept Model

6.1 Indicative Solutions

6.1.1 Approach One

Most of Spackman’s original five recommendations remain valid but it would be prudent to verify that the approach as set out will yield all the results now required, for example by means of some sandpit modelling in either Protégé or the new IHTSDO collaborative editing platform.

This testing would also reveal whether :

• “schedule parts” can be represented using a compositional approach but without changing the concept model
• the desired autoclassification can be achieved without using further right identity/role inclusion axioms

Future use of the terms including the words ‘immunization’ and ‘vaccination’ should be subject to strong editorial guidance and rules - ideally automatically enforced - to avoid future conflation of vaccination and immunization.

6.1.1.1 End user Impact of approach one

A reduction in the number of false negative and false positive subsumptions within the taxonomy of immunizations will reduce the complexity of data extract and other report specifications in this subdomain and increase their reliability.

A fully modelled approach would reduce the risk that any fix to the taxonomy would be only temporary, with the next primitive additions serving to break it again.

An extensible compositional model for defining new vaccination programmes and new schedule parts for existing programmes would provide a mechanism by which changes in mass vaccination programmes could eventually decouple themselves from the current absolute dependency on new terminology codes being provided from the centre.

7 Indicative Project Plan

The project requires reworking the modelling of around 300 concepts in total, most of which are currently found in a single and self-contained taxonomy. A minority of this number will also need changes to their preferred terms and fully specified names (ie retire-and-replace)

7.1 Scope of elaboration phase

The project divides into four workpackages:

1. Elaboration : Initial sandbox prototyping of the content revision required
   Skills:   Consultant terminologist
   Resource:       1 person month
   Elapsed time:   1 month
2. Dissemination to stakeholders and request for impact analysis
   Skills:               Chief Terminologist
   Resource:       2 person months
   Elapsed time:   6 months
3. Implementation of agreed changes in the release data
   Skills:               Junior terminologists
   Resource:       2 person months
   Elapsed time:   1 month          
4. Implementation of editing environment rules to prevent recurrence

Skills:               Technical
Resource:       1 person month
Elapsed time:   1 month

7.2 Projection of overall project size and resource requirements

7.2.1 Expected project resource requirement

The project resource requirement is classed as SMALL – less than 1 person year

7.2.2 Expected project impact and benefit

The project impact is HUGE – significant improvement to almost all citizens and in a high profile and internationally common use case

7.2.3 Indicative resource estimates for elaboration, construction, transition and maintenance:

Elaboration phase:                  1 person month effort, 1-2 months elapsed time

Construction and transition phase:     100 new concepts to be authored, 200 to be remodelled

Maintenance phase:                50 new ‘frequent usage’ concept requests in 1^st 3 years

8 Appendices

8.1 Appendix One : Immunizations  Kent Spackman, Apr 2010

Several prime complaints have been lodged about the current hierarchy of immunization procedures:

1. Procedures are primitive and therefore do not subsume appropriate children
2. Manual assignment of children has not been accurate
3. Definitions do not capture entity (organism) against which immunization is directed.  

Per Ian Green (at the time, a terminologist at the UK NRC): “What's really needed here, and would be particularly valuable for any real-world applications, is a much more compositional model where the vaccination hierarchy is entirely defined rather than primitive entities, being at a minimum defined (ultimately) in terms of the organism(s) vaccinated against (which should be separate from the preparation(s) used) and the vaccination schedule dose given, and refinable by e.g. the site of the injection.”

8.1.1 Definitions (Dorland):

Immunity: the condition of being immune; the protection against infectious disease conferred either by the immune response generated by immunization or previous infection or by other nonimmunologic factors

Active immunization: stimulation of the immune system to confer protection against disease, e.g., by administration of a vaccine or toxoid

Passive immunization: the conferring of specific immune reactivity on previously nonimmune individuals by the administration of sensitized lymphoid cells or serum from immune individuals.  (Note: SNOMED’s current concept Passive immunization (procedure) has an incorrect synonym of Immunoglobulin injection.)

Adoptive immunization: passive immunization by transfer of sensitized lymphocytes from an immune donor to a previously nonimmune recipient.

Vaccination: the introduction of vaccine into the body for the purpose of inducing immunity.

Vaccine: a suspension of attenuated or killed microorganisms (bacteria, viruses, or rickettsiae), or of antigenic proteins derived from them, administered for the prevention, amelioration, or treatment of infectious diseases

8.1.2 Modelling Options

Can these root procedures be fully defined so that they will subsume (only) their legitimate offspring?  There are several options:

1. Define them by substance administered
2. Define them by procedure action
3. Define them by procedure intent
4. Define them by object of the immunity

The above definitions from Dorlands somewhat understate the complexities to be considered.  To review each of the above options:

8.1.2.1 Definition by Substance administered

127785005|Administration of substance to produce immunity, either active or passive (procedure)|

syn:  Immunization

IS A Administration of therapeutic substance (procedure)

Role group 1

{Has Method=Introduction of substance –action,

Direct substance=Vaccine, immunoglobulin, AND/OR antiserum (substance)}

Problems:

In SNOMED’s ‘intents’ hierarchy, “preventative” is a synonym, not a subtype of “therapeutic.”  This is inconsistent with the assignment of immunizations, which usually (though not always!) have a prophylactic intent, as subtypes of Administration of therapeutic substance—and with giving Administration of therapeutic substance a synonym of Administration of substance.  (Is a contrast agent for radiography a “therapeutic substance”?)

The direct substance (Vaccine, immunoglobulin, AND/OR antiserum) does not include the sensitized lymphocytes used in adoptive immunization (see definitions above).  Nor can we simply add sensitized lymphocytes as subtypes of this grouper concept, because cells and substances reside in separate hierarchies.  (We also don’t currently have a mechanism by which cells used with a medical intent can be used as values for Direct substance. This issue is larger than vaccination: spermatozoa/ova in artificial insemination or IVF; red blood cells, platelets, etc in transfusions, stem cells, are similarly baffling to model.) 

This problem also prevents full definition of passive immunization by means of its direct substance, because merely giving that concept a direct substance of Immunoglobulin AND/OR antiserum would not accommodate its subtype, Adoptive immunity

Conclusion:

This approach would probably be adequate to fully define

New concept Administration of vaccine to produce active immunity (procedure) as:

IS A Administration of substance to produce immunity, either active or passive

Group 1

Has Method Introduction of substance –action

Direct substance Vaccine

However, this model does not allow either Immunization or Passive Immunization to be fully defined under current rules. 

8.1.2.2 Definition by Procedure method

We used to have a concept 129446007|Immunization - action (qualifier value)|. It was retired in 2003 as erroneous with REPLACED_BY 303430001|Introduction of substance (qualifier value)|, but then in 2008 moved to ‘retired no reason’ and lost its active concept substitute.  We could reactivate this concept and possibly add subtypes of Active immunization and Passive immunization, using these to fully define the procedures in question.   This would probably be the simplest of the available means to fully define immunization procedures.

Problems:

I agree with what I suspect was the rationale behind retiring 129446007|Immunization - action (qualifier value)| —that one is not actually performing the immunization, but rather an action one hopes will result in the production of immunity—that immunization is the intent, not the action.  However, we don’t appear to adhere very strictly to this rule: consider Fixation – action, or Repair – action.   That may not be an argument in favor of further fudging—Repair-action causes a lot of problems in procedures because many other actions that may be involved in a repair (like fixation) neither are nor can be defined as subtypes of repair, so procedures assigned repair as a parent often acquire an additional and often ungrouped action.  Immunization procedures would similarly have two ‘actions’—Administration of substance and Immunization.

Conclusion:

A quick and simple solution, if a somewhat ‘dirty’ one.

8.1.2.3 Definition by Procedure intent

In some respects, this is the most logical approach.  Immunization is, truly, the intent of these procedures, as suggested by the fully specified name of the root concept, Administration of substance to produce immunity, either active or passive.  This concept would then be defined as:

127785005|Administration of substance to produce immunity, either active or passive (procedure)|

IS A Administration of substance (procedure)

Role group 1

{Has Method=Introduction of substance –action,

Direct substance=Vaccine, immunoglobulin, AND/OR antiserum (substance),

Has intent=Immunization intent}

Problems:

Immunization is considerably more specific than any of the other modelled values of Has intent (diagnostic, therapeutic, preventative, palliative, screening) and so represents a precedent that might raise questions about whether we ought not model other procedures with an implied intent (like Surgical repair) in a similar way.  This might be a good thing, ultimately, but definitely could lead to an expanded scope of work. 

Due to the problems with the substance/cell issue, we still couldn’t fully define Passive immunization with this method except by adding a subtype “Passive immunization – intent”—which really would be pushing the boundaries of the meaning of “intent.”

If we do add Immunization as an intent, what should be its immediate parent?  The obvious answer would be Preventive - procedure intent (qualifier value), but I’m not sure that’s always true.  What about the use of cancer vaccines as adjunctive therapy, where tumor-derived vaccines are used therapeutically rather than prophylactically?  Is giving a “cancer vaccine” an “Immunization (procedure)”?  Also, autogenous vaccination “A second vaccination with virus from a vaccine sore or liberation of antigenic products from invading microorganisms on the same individual. (Stedman)’ is clearly not ‘preventative.’

Conclusion:

Tempting, and superficially logical, but using this approach to allow full definition of passive immunity would strain the meaning of intent.

8.1.2.4 Definition by Object of the immunity

Adding attribute-value pairs to define the object of immunity would not, by itself, allow full definition of any of the immunization concepts.  However, it would meet the expressed desire to define immunizations “in terms of the organism(s) [separately from] the preparation(s) used.” 

I think we could logically use the existing Has focus attribute to meet this need.  We could do so within the existing model as follows:

Measles vaccination (procedure)

IS A Administration of substance to produce immunity, either active or passive

Has Method Introduction of substance –action

Direct substance Measles vaccine (substance)

Has focus Measles (disorder)

Problem:

Would require reliance on right identity to find the relationship between the immunization procedure and the actual organism.  Will users accept this?

Conceivably, we could get around this problem by expanding the allowed value set for the Has focus attribute to include organisms:

Measles vaccination (procedure)

IS A Administration of vaccine to produce active immunity (procedure)

Has Method Introduction of substance –action

Direct substance Measles vaccine (substance)

Has focus Measles virus (organism)           

I suspect we should stick to the former approach, as immunizations, strictly speaking, don’t prevent organisms but diseases due to organisms.  This would also allow a consistent approach when the immunization is not against an organism but against a toxin (botulism, clostridium, tetanus) or against cancerous cells.

8.1.3 Additional questions raised:

8.1.3.1 Booster vaccination/immunization

What should be the relationship between Booster vaccination (procedure) and Booster immunization?

Recommendation:

Booster immunization should be retired as ambiguous with a single MAY BE A link to Booster vaccination.

Rationale:

Booster immunization is presently defined as a subtype of Administration of vaccine to produce active immunity (procedure).  This definition makes it a synonym of Booster vaccination.

One could argue for two distinct concepts, Booster immunization and Booster vaccination, ONLY if some booster immunizations are not vaccinations, but passive immunization.  Are there any cases in which repeated injections of immunoglobulin, antiserum, or sensitized lymphocytes are referred to as “boosters”?  Stedman defines “booster dose” as “a dose given at some time after an initial dose to enhance the effect, said usually of antigens for the production of antibodies.”  Dorland states even more strongly that a booster dose is “a dose of an active immunizing agent, usually smaller than the initial dose, given to maintain immunity.” 

Given these definitions, and the fact that Booster immunization has NO subtypes in SNOMED, I think we can retire it as suggested, or even as duplicate to Booster vaccination.

We also have a request to add, as primitives, “concepts for first, second, third, fourth etc in a primary vaccination schedule.”  We could add these—but how would they function to meet user needs?  I don’t think it’s feasible to precoordinate every first and subsequent vaccination for all the immunization procedures in SNOMED.  We may need to consider developing a template for post-coordinating this sort of information. 

8.1.3.2 Autogenous vaccination/immunization

What should be the relationship between Autogenous vaccination (procedure) and Autogenous immunization (procedure)?

Recommendation:

Autogenous immunization should be retired SAME AS Autogenous vaccination

Rationale: 

There are almost no references to Autogenous immunization; a Google search on this term turns up only 8 references (most of them to SNOMED itself) with one or two sources referring to “Autogenous immunization therapy: a medical technique that involves the injection of a person's own sterilized urine for the purpose of treating allergies.”  (There are more references to urine autoinjection, most disparaging.  Apparently it’s sometimes used as a test for allergies, as well as a therapy.  Ugh).  Autogenous vaccine is defined by Stedman’s as “a vaccine made from the patient's own microorganisms”

8.1.3.3 Administration of antitoxins

How should we define Administration of tetanus antitoxin and Administration of botulism antitoxin (procedure) and their subtypes?

Recommendation:

As subtypes of a new concept, Administration of bacterial antiserum (procedure)

Rationale:

Administrations of bacterial antitoxins are logically types of passive immunization against a bacterial disease.  However, Immunization against bacterial disease (procedure) is presently defined with a direct substance of “Bacterial vaccine.”   This is incorrect.  Administration of bacterial vaccine should be created as a subtype of Immunization against bacterial disease (procedure), and it (the subtype) can then be defined with a direct substance of “Bacterial vaccine.” 

Immunization against bacterial disease (procedure) is itself therefore an ambiguous FSN, because it isn’t clear whether it is meant to refer only to bacterial infections or to any disease caused by bacteria or their toxins, which may be absorbed via ingestion rather than by infection, such as botulism.  I think Immunization against bacterial disease (procedure) should be retired as ambiguous, with a MAY BE A link to a two new concepts, Administration of vaccine against bacterial infection and Administration of bacterial antitoxin.

8.1.3.4 Antiserum vs Immunoglobulin

What should be the proper relationship between Antiserum (substance) and Immunoglobulin (substance), and how should they be used in defining passive immunization procedures? 

Recommendation:

Immunoglobulin (substance) and its subtypes should be used, rather than antiserum.

Rationale:

At the moment they are siblings under Vaccine, immunoglobulin, AND/OR antiserum (substance), with Immunoglobulin having an additional parent of Plasma protein (substance).  In a purely logical sense, it seems to me that Immunoglobulin is the "active ingredient" of Antiserum, but we have no legitimate way to represent this kind of relationship within the substance hierarchy (so far as I know), and neither substance seems to me a subtype of the other.  Antiserum (substance) has only two subtypes: antivenin and antitoxin (actually, antivenin ought to be a subtype of antitoxin).  Given that Stedman defines antitoxin as “Antibody formed in response to antigenic poisonous substances of biologic origin” I think we should move antitoxin and antivenin over under Immunoglobulin (substance).  

In the procedure hierarchy, we have Administration of immune serum (procedure), but it is primitive—the substance Antiserum is not used in its definition, and its only subtype is Intravenous infusion of Respiratory syncytial virus immune globulin, human (procedure).

Given that all administrations of immune serum are, by definition, introductions of immunoglobulin, I’m not sure that having either antiserum (substance) or Administration of immune serum (procedure) has any benefits that outweigh the complications they introduce.  Since neither is presently subsuming appropriate subtypes, I’d prefer to retire both with MAY BE A links to the equivalent “immunoglobulin” substance and procedure.

8.1.3.5 Postcoordinating models for vaccination

How might we best meet the desire of users to make vaccination procedures "refinable” by e.g. the site of the injection”?

Recommendation:

Indirect procedure – site could adequately capture this information in post-coordination, though adding  precoordinated concepts for all the possible sites of all the vaccines may not be feasible. 

Rationale: 

I considered whether we shouldn’t use Route of administration, which seems a logical option, but it would not allow specific anatomical sites of administration to be specified (which body region for a subcutaneous vaccine, for instance, or which muscle for an intramuscular vaccine).  This information might be important to track—for instance, in determining which vaccine might have provoked a vaccine-site fibrosarcoma.  To capture this information, we would need to deploy Procedure site – indirect as a second attribute in defining Route of administration.  By contrast, Procedure site – indirect alone would, I think, allow adequate characterization of the vaccination site.  Also, the use of Route of administration has so far been restricted to a single skeleton hierarchy, Administration of treatment via specific route (procedure) and its subtypes. I am not sure what purpose is served by retaining either Administration of treatment via specific route (procedure) or Route of administration (attribute) since neither is used to define such seemingly logical subtypes as children as Intramuscular drug injection or Oral fluid replacement.  In any case, existing SNOMED precedent does not seem to encourage use of Route of administration.

Since vaccinations are administrations of a substance either by intramuscular or subcutaneous injection, intranasally, or orally, I looked at some procedures that seemed like logical parallels and found that we have used Indirect procedure site at least some of the time to capture the site of administration:

Intramuscular

Intramuscular drug injection (procedure)

Role group 1

Method= Injection – action

Direct substance =Drug or medicament

Procedure site – indirect = Skeletal muscle structure

Theoretically, for intramuscular vaccines, we could simply follow this approach, refining the indirect procedure site to a more specific site.

Subcutaneous

Subcutaneous chemotherapy (procedure)

Role group 1

Method = Injection – action

Direct substance = Drug or medicament

Procedure site – indirect = Subcutaneous tissue structure

Intranasal

Intranasal oxygen therapy (procedure)

Procedure site direct = Nasal structure

Direct substance = Oxygen 

This concept is clearly incorrectly modeled --the procedure site should be indirect and probably Nasal cavity structure would be more accurate, plus the site is not role-grouped with the inherited Direct substance Oxygen.  However, I think we could reasonably model:

Administration of intranasal vaccine

Method = Introduction – action

Direct substance = Vaccine

Procedure site – indirect = Nasal cavity structure

Oral

None of the existing procedure concepts involving oral administrations (such as Oral fluid replacement) are modeled in such a way as to define the ‘oral’ aspect of the procedure.  We could use an approach similar to that suggested for intranasal vaccination, e.g.

Administration of oral vaccine

Method = Introduction – action

Direct substance = Vaccine

Procedure site – indirect = Oral cavity structure

8.1.4 SUMMARY

While it will be extremely difficult to fully define some top-level concepts, such as

Administration of substance to produce immunity, either active or passive (procedure)

Passive immunization (procedure), we can, I believe, greatly improve the usability of this hierarchy by making certain changes consistent with the existing SNOMED model for procedures.

8.1.5 RECOMMENDATIONS

1. Allow 127785005|Administration of substance to produce immunity, either active or passive (procedure)| to remain primitive.  Ultimately, this concept should have relatively few direct subtypes, which can be manually assigned.
2. Retire existing concept 51116004|Passive immunization (procedure)| as ambiguous:

MAY BE A Administration of substance to produce passive immunity (procedure)

MAY BE A Administration of immunoglobulin (procedure) (new concept)

Administration of substance to produce passive immunity (procedure) cannot be fully defined; however it needs only a few children:

Administration of immunoglobulin

      Administration of antitoxin

      Administration of antivenin

Administration of sensitized lymphocytes (syn: Adoptive immunization)

1. Fully define Administration of vaccine to produce active immunity (procedure) as

IS A Administration of substance to produce immunity, either active or passive

Role Group 1

Has Method Introduction of substance –action

Direct substance Vaccine

1. Create and fully define

Administration of vaccine against bacterial infection (procedure)

IS A Administration of vaccine to produce active immunity (procedure)

Has focus Bacterial infectious disease (disorder)

Role Group 1

Direct substance Bacterial vaccine (substance)

Method Introduction of substance (qualifier value)

Fully define other vaccinations against infectious disorders using the same model.   It may seem redundant to use both “Direct substance Bacterial vaccine” and “Has focus Bacterial infection” in defining Administration of vaccine against bacterial infection.  However, while “Bacterial vaccine” is a useful grouper in the substance hierarchy, neither it nor its subtypes have any link to the organisms or the disorders, which our