Stream: implementers
Topic: Immunization status
Jeffrey Danford (Jun 23 2017 at 13:32):
We're in the process of migrating from DSTU2 to STU3 and I have a question about Immunization.status. Previously status supported on-hold or in-progress, but now the only options are completed and entered-in-error. In DSTU2 we reported multipart vaccines with a status of in-progress and documented each administration in immunization.protocol, but that's no longer an option. Do we just have it backwards - the resource is for each administration and the protocol indicates it place in the series?
Lloyd McKenzie (Jun 23 2017 at 18:10):
@Joginder Madra ?
Michelle (Moseman) Miller (Jun 23 2017 at 18:24):
It is my understanding that Immunization only represents a single administration (within a possible series). Although Immunization.vaccinationProtocol supports 0..* that is meant for when a single administration satisfies multiple protocols (and is not used to represent all administrations in the series). There is only one Immunization administration date, location, manufacturer, lotNumber, etc. at the root Immunization level, so it is hard for me to see how you could have used Immunization.vaccinationProtocol previously to represent all administrations in the series when each immunization administration could have its own unique date, location, lotNumber, etc.
Jeffrey Danford (Jun 23 2017 at 19:52):
So we have it (mostly) backwards. Thanks for the confirmation and clarification.
Jose Costa Teixeira (Jan 24 2021 at 11:21):
How do we say
- This person is exempt from vaccination because she is pregnant
- This person does not need to be vaccinated now because they were vaccinated last week with J&J vaccine (takes one dose)
- This person does not need to be vaccinated now because they were vaccinated last week with Pfizer vaccine (needs to wait a few weeks)
- This person does not need to be vaccinated because they have antibodies
Jose Costa Teixeira (Jan 24 2021 at 11:21):
I'm interested in the final assertion, not tracking the whole reasoning and evidence
Jose Costa Teixeira (Jan 24 2021 at 11:23):
btw @Craig Newman I presume this is a good reason to introduce the notion of protocols being highly dynamic. In the good old days, a protocol was fixed for a disease for a country for several years. Now this protocol depends on the actual product - if you get one brand, you need one dose, if you get another brand, you need two doses with 21 day interval, or 28 day interval
Christof Gessner (Jan 24 2021 at 11:27):
This implies looking at the past, and making a forecast, possibly including due dates or assumed protection period.
Christof Gessner (Jan 24 2021 at 11:28):
Do we have a register for such protocols across countries and products, past and present?
Craig Newman (Jan 25 2021 at 13:34):
Michelle is correct, Immunization describes a single administration of a vaccine (although depending on which vaccine it is (eg DTaP), a single administration even may help protect against multiple diseases). If a patient has completed a 2 dose series (such as for MMR), we'd expect there to be a minimum of 2 Immunization resources in their history.
Craig Newman (Jan 25 2021 at 13:36):
ImmunizationEvaluation (in R4) can be used to evaluate each Immunization against a specific schedule. As Jose points out, these schedules change (recommended intervals between doses can change, vaccines weren't considered interchangeable are now considered interchangeable, etc). So the evaluation is a snapshot at a point in time against a specific set of recommendations.
Craig Newman (Jan 25 2021 at 13:38):
I don't know of any registry of immunization protocols, but in the US, the CDC has sponsored the CDSi project to convert recommendations from the CDC's ACIP group into machine processable data for ingestion into CDS engines. CDSi documentation can be found at https://www.cdc.gov/vaccines/programs/iis/cdsi.html
Craig Newman (Jan 25 2021 at 13:45):
I'll take shot at answering Jose's questions:
This person is exempt from vaccination because she is pregnant
Craig - ImmunizationRecommendation could be used with .forecastStatus indicating that no vaccine is due and the reason in .forecastReason. Again, because schedules change over time (as do patient circumstances, like being pregnant), an ImmunizationRecommendation is snapshot in time.
This person does not need to be vaccinated now because they were vaccinated last week with J&J vaccine (takes one dose)
Craig - ImmunizationRecommendation can indicate this with .forecastStatus set to a value like Complete
This person does not need to be vaccinated now because they were vaccinated last week with Pfizer vaccine (needs to wait a few weeks)
Craig - Also handled by ImmunizationRecommendation where .dateCriterion includes a date type of "Date Vaccine Due" and a future date
This person does not need to be vaccinated because they have antibodies
Craig - ImmunizationRecommendation where .forecastStatus is something like Immune
An Observation might also be appropriate although in some of these cases, but it's important that it's clear that a patient's status may change in the future (eg. presumably the woman might need a vaccination once she's no longer pregnant).
Jose Costa Teixeira (Jan 25 2021 at 16:00):
So you'd use the ImmunizationRecommendation to recommend NOT to vaccinate?
Jose Costa Teixeira (Jan 25 2021 at 16:03):
I don't dislike the simplicity at all.
I do wonder if ImmunizationRecommendation is too specific / if there is another fhir resource that says "Dr. D states that Patient P does not need this action". (we'll have to see if there is a difference between "does not need" and "is exempt from..."
Craig Newman (Jan 25 2021 at 21:03):
When you say it THAT way, it does seems a bit odd :grinning: , but I guess the idea is that it's more about guiding actions (even if inactive is the most appropriate course). At one point, we were talking to the CDS group about a more general pattern for offering clinical support, but we never really identified a different way to do it.
Jose Costa Teixeira (Jan 25 2021 at 22:08):
Let me bring the use case (ok maybe not the most relevant use case but is a very good equivalence):
Jose Costa Teixeira (Jan 25 2021 at 22:10):
I want my house to have a electric lock, that only opens if visitors provide a set of FHIR resources that tell me "this person is as vaccinated or has one of these very good excuses)
Jose Costa Teixeira (Jan 25 2021 at 22:10):
(instead of my backyard, it can be an airplane boarding gate, or a office building space)
Jose Costa Teixeira (Jan 25 2021 at 22:12):
I don't want any explanations, I just want "in or out"
Christof Gessner (Jan 26 2021 at 00:17):
...and without disclosure of any personal health data or other data that are not required for the case.
Jose Costa Teixeira (Jan 26 2021 at 11:29):
Christof Gessner said:
...and without disclosure of any personal health data or other data that are not required for the case.
Precisely. That is the whole case - no reasoning artifacts, no tracing, nothing - this is for me to see if I can let guests into the house, not for me to evaluate their clinical background. I am not a healthcare professional.
I have no right to look at their data, and definitely they cannot claim that they don't give consent, so I have no right to lock the door.
Craig Newman (Jan 26 2021 at 13:54):
So it sounds like even Immunization as the "yes" answer is too clinical (too much detail included) and by itself insufficient because there are rules about the immunization sequence. Two Immunization resources for covid vaccine may not be sufficient if the two doses were given only 10 days apart rather than the required 21/28 days apart.
Gino Canessa (Jan 26 2021 at 16:22):
There is a lot of work (on many fronts) being done to standardize some flavor of a verifiable credential (e.g., https://smarthealth.cards/ ). As far as I know, still very much work in progress though.
Josh Mandel (Jan 26 2021 at 16:27):
In my own personal way of thinking:
- we need standardized ways to share verifiable clinical data (low-level details about products, dates, lot numbers, etc) -- i.e., what we call "health cards" in the SMART spec. These are generally immutable and long-lived; and then
- we need standardized ways to generate and share evaluations (e.g., the reuslt of executing a set of rules or polices where the inputs are some combination of health cards, questionnaires, etc) -- i.e., what I call a "pass" or "summary" artifact. These are much more contextual and short-lived.
Peter Jordan (Jan 26 2021 at 22:17):
Craig Newman said:
Two Immunization resources for covid vaccine may not be sufficient if the two doses were given only 10 days apart rather than the required 21/28 days apart.
Surely, the key purpose of an Immunization Resource is to record an individual administration of a vaccine regardless of the recommended interval between administrations of the same vaccine - or am I missing something here?
Craig Newman (Jan 27 2021 at 13:52):
@Peter Jordan that's right, the Immunization resource is about the immunization event itself (what actually happened). But just because an immunization happened doesn't mean that it meets the clinical requirements to start the patient on the road to (presumed) immunity. For example, if a dose of MMR was administered to a 10 month old, that event would be recorded in an Immunization resource. But in the US, that dose wouldn't be considered valid for the child's recommended series of MMR vaccines because it was administered too early. So just the presence of the Immunization resource doesn't mean that the child has fulfilled the recommended vaccination.
Christof Gessner (Jan 28 2021 at 19:24):
copied from "Immunization manufacturer" topic:
Giorgio Cangioli said:
Looking at the immunization manufacturer element I was wondering if this element represents only the actual manufacturer ("the organisation that holds the authorisation for the manufacturing process") or if it is a more comprehensive element covering any organization having any kind of 'ownership' with the product (as manufacturer , as marketing authorization holder (MAH), as dealer...).
I'm asking that because looking at some immunization data sets currently under development there are some referring the MAH, others the manufacturer and assuming they are using these terms appropriately (i.e. by using the IDMP definitions) I was wondering if both map in the manufacturer element or not....
Floyd Eisenberg (Jan 29 2021 at 19:46):
Sorry because I'm new to this thread and you may have considered this option. The MedicationKnowledge http://build.fhir.org/medicationknowledge.html resource may be a way to handle the immunization product details/metadata.
Richard Townley-O'Neill (Feb 03 2021 at 05:16):
Jose Costa Teixeira said:
Precisely. That is the whole case - no reasoning artifacts, no tracing, nothing - this is for me to see if I can let guests into the house, not for me to evaluate their clinical background. I am not a healthcare professional.
I have no right to look at their data, and definitely they cannot claim that they don't give consent, so I have no right to lock the door.
That seems like an Observation.
Jose Costa Teixeira (Feb 06 2021 at 22:41):
Floyd Eisenberg said:
Sorry because I'm new to this thread and you may have considered this option. The MedicationKnowledge http://build.fhir.org/medicationknowledge.html resource may be a way to handle the immunization product details/metadata.
Medication contains things like Lot nunmber, serial number, etc. and is the one to use in transactions like this one. Medication has too little attributes to fully define medication, MedicationKnowledge fits that purpose. The only issue is that there's no link between them - We cannot have a Medication in an immunization and univocally refer to a medicationKnowledge
Jose Costa Teixeira (Feb 06 2021 at 22:43):
Richard Townley-O'Neill said:
Jose Costa Teixeira said:
Precisely. That is the whole case - no reasoning artifacts, no tracing, nothing - this is for me to see if I can let guests into the house, not for me to evaluate their clinical background. I am not a healthcare professional.
I have no right to look at their data, and definitely they cannot claim that they don't give consent, so I have no right to lock the door.That seems like an Observation.
I can perfectly accept that indeed. Someone is stating something based on whatever criteria is valid on their end, and it is for a very specific purpose. Thanks @Richard Townley-O'Neill
Josh Lamb (Feb 06 2021 at 22:53):
I have followed a few threads and apps about tracking immunization status. I am a bit confused about why this is complex. Do we not have a consistent way to record that a vaccine is administered? I have also heard cases where people are unable to certify the vaccine because they cannot certify the codes (there is like 3, which should be easy, right?)
Lloyd McKenzie (Feb 07 2021 at 01:27):
There's rarely any such thing as "easy" in healthcare :(
Note: there being 3 vaccines is not the same as there being only 3 codes.
Jose Costa Teixeira (Feb 07 2021 at 09:29):
Josh Lamb said:
I have followed a few threads and apps about tracking immunization status. I am a bit confused about why this is complex.
While I am after something simple (hence my silly example of a door lock that opens with a super-simple vaccination status)
Some topics I recall have added to this complexity:
Jose Costa Teixeira (Feb 07 2021 at 09:29):
- Immunization does not mean immunity
Jose Costa Teixeira (Feb 07 2021 at 09:43):
- a bit more than 3 codes... Let's see the Pfizer vaccine:
- One code for "Covid Vaccine" (SNOMED 1119305005, ATC J07BX03)
- ICD-11 has a code 91300
- There's a specific code for "messenger RNA COVID-19 Vaccine": in SNOMED, 1119349007 fits, but 1142178009 also seems to fit.
- For the active substance, there is a UNII (where the U means "it's unique, doesn't mean it's the only one"). for Pfizer it is TOZINAMERAN 5085ZFP6SJ https://fda.report/UNII/5085ZFP6SJ
- one code per country per authorization - so if the vaccine has 2 authorization procedures you may get 2 codes (not sure if anyone is doing it like that, but they can)
- I don't know if there are GTIN codes for that.
- One code for "Covid Vaccine" (SNOMED 1119305005, ATC J07BX03)
Josh Lamb (Feb 07 2021 at 18:21):
Thanks for sharing @Jose Costa Teixeira . I have spoken to a few devs who have apps that display vaccination status. Their feedback is similar to yours.
I am also concerned about the intent of these vaccine cards. Why do other people need to know if I am vaccinated? Given the pace of the tech and vaccine distribution it seems that vaccine cards and distribution will be even.
If someone does not want a vaccine that is their choice. If the vaccine does not grant immunity I do not see what we are declaring with the status (partial immunity?)
If someone is vaccinated I am not sure why they need to know if other people are vaccinated.
I am concerned because vaccines are a polarizing topic and there is a lot of misinformation. I hope we are not considering using vaccine cards as a way to reopen or to determine if someone can receive a service (like Uber or starbucks).
Josh Lamb (Feb 07 2021 at 18:22):
I 100% support vaccine cards and people being allowed to easily communicate their vaccine status but I hope we do not get extreme with the cards.
Jose Costa Teixeira (Feb 07 2021 at 19:28):
a few replies on the topics I think I should comment:
Jose Costa Teixeira (Feb 07 2021 at 19:29):
(deleted)
Jose Costa Teixeira (Feb 07 2021 at 19:31):
Josh Lamb said:
I am also concerned about the intent of these vaccine cards. Why do other people need to know if I am vaccinated? Given the pace of the tech and vaccine distribution it seems that vaccine cards and distribution will be even.
I don't think that everyone is going to be immune/vaccinated against covid that soon. Maybe most people in developed countries will get their 1st covid vaccination (1 or 2 doses) but that won't be the end of the story, I guess
Jose Costa Teixeira (Feb 07 2021 at 19:49):
Finally, some vaccines are mandatory in some places. Some schools require vaccination certificates.
In this forum we only discuss the technical considerations.
Lloyd McKenzie (Feb 07 2021 at 20:05):
The cards are also an indication of probable risk of being able to get sick and possibly of risk of being able to spread virus. There are lots of organizations that will want to reduce risk. There will absolutely be employers and potentially others who, once vaccine access is widespread, will make vaccination (or medical documentation indicating inelligibility) a requirement for employment. (Not getting a vaccine may be the employee's choice, but continuing to employ them is the employer's choice.) Also, there may come appoint where testing people who have been properly vaccinated is deemed a waste of resources. Finally, some vaccines may need boosters or may be more or less effective against certain variants, so the cards will be key to determining if/when subsequent vaccinations are necessary. It seems very unlikely that this is going to be a situation where you get 2 shots and are done forever.
Last updated: Apr 12 2022 at 19:14 UTC
