Stream: implementers
Topic: ConditionDefinition
Jose Costa Teixeira (May 13 2020 at 22:11):
Hi. ConditionDefinition says it doesn't follow any patterns, but it would follow the Definition pattern, right?
Grahame Grieve (May 13 2020 at 22:44):
yes it should; I didn't get around to doing the mappings for that
Jose Costa Teixeira (Jun 27 2020 at 19:13):
what is the difference /relation between ConditionDefinition and ClinicalUseIssue?
If i wanted to say "this medication has an indication X", I' assume we'd have a ConditionDefinition for that, or at least a path to a ConditionDefinition.
Lloyd McKenzie (Jun 27 2020 at 19:16):
ClinicalUseIssue is essentially the definition for a DetectedIssue (i.e. a contraindication or similar issue). ConditionDefinition seems a little odd to me, but appears to be a grab-bag of "stuff related to Condition X"
Rik Smithies (Jun 27 2020 at 20:59):
ClinicalUseIssue just points to coded things but could have codeableReference if there are FHIR definitions of conditions. But ConditionDefinition doesn't seem to define a condition (or a disease/symptom), but how to diagnose and manage one. I wouldn't expect that from the name. Maybe DiagnosisDefinition? Diagnosis has a connotation both of the process of diagnosing, and then it becomes a diagnosis, aka condition, which then needs management. Or Maybe DiagnosisKnowledge, because it seems more informational that definitional.
Jose Costa Teixeira (Jun 27 2020 at 21:13):
There seems to be some overlap (functionally).
In some cases, this included, it seems that master data is spread around different resources - which is not a problem if that is how implementers do it, but I don't know if there is one way that implementers do it.
Rik Smithies (Jun 27 2020 at 22:00):
where do you see overlap Jose? Between ClinicalUseIssue and ConditionDefinition?
Jose Costa Teixeira (Jun 28 2020 at 07:24):
Well, one seems to be a definition of a condition, the other one seems to be a definition of a condition...
Jose Costa Teixeira (Jun 28 2020 at 07:29):
the scope boundaries are not clear to me. ClinicalUseIssue seems more specific than the other, and is about medication, the other more general. I don't see this as something that is medication-specific.
Jose Costa Teixeira (Jun 28 2020 at 07:41):
I cannot comment beyond my baseline opinion about "knowledge" vs "definition": I do not think systems behave differently. To me, a non-native english speaker, this difference does not compute.
If a system asserts some information about some type that means "here's what you should know about this type", that's a definition, regardless of whether this is "raw" knowledge (does that exist)?) or local, administrative or regulatory definition (which is a continuum and i do not differentiate).
I do not understand / agree why we have xxxDefinition vs xxxKnowledge. And the first time I heard "knowledge" was from drug DBs wanting to add a buzz word for marketing points, so that marked me ;)
Again I am not a native english speaker, so I may be missing some subtleties that most people will grasp. I just hope we keep things a) consistent if possible, and b) on a continuum because silo boundaries are never fixed.
Lloyd McKenzie (Jun 28 2020 at 14:01):
Definition of a contraindication and definition of a condition aren't really the same thing... ClinicalUseIssue says "don't use X when Y".
Rik Smithies (Jun 28 2020 at 14:03):
@Jose Costa Teixeira ClinicalUseIssue is not the definition of a condition. It does not tell you what hypertension is (which would be its definition), but it tells you that a drug may cause it.
ClinicalUseIssue is, roughly, the assertion of a known link between a medication and a condition caused by its use. It doesn't define the condition, it just refers to it.
(It need not be medication, it could be a procedure etc, and it need not be something that is described with an actual FHIR Condition, but just some other negative thing e.g that it stops working properly)
Neither the word definition nor knowledge is perfectly, err, defined, to only mean one thing in all contexts, but there are some differences. "Continuums" and overlapping concepts are a fact of life but we do what we can.
Definitions tend to be more fundamental and about unchanging intrinsic things.
Then you can add other facts that are more about context and not so intrinsic.
It's fairly definitional that a tablet is made of a certain substance. If it was a different substance it would have to be a different tablet.
But its price and how it is used are subject to variation, over time and from place to place. So those are less definitional.
ConditionDefinition does not seem to me to be the fundamental definition of a condition, so that doesn't seem a good name. It seems more about diagnosis and management of a condition. If we did need something that was truly definitional, e.g. that hypertension is high blood pressure, then I wonder what that resource would be called.
Raed MO (Jun 28 2020 at 14:43):
Hi,
I'm new to HAPI, docker and linux .. For a school project , I need to implement today a HAPI FHIR server
Raed MO (Jun 28 2020 at 14:44):
I cloned the hapi-fhir-jpaserver-starter and modified the code to meet my requirements then did as they said in the README file:
mvn clean install
docker-compose up -d --build
It did deploy the server but with a new fresh HAPI server, not the one I modified and built.
How can I use docker compose to deploy my build not the version he gets from the docker repo ?
Raed MO (Jun 28 2020 at 14:44):
Here is my docker-compose.yml (same as the github repo) :
version: "3"
services:
hapi-fhir-jpaserver-start:
build: .
container_name: hapi-fhir-jpaserver-start
restart: on-failure
ports:
- "8888:8080"
hapi-fhir-mysql:
image: mysql:latest
container_name: hapi-fhir-mysql
restart: always
environment:
MYSQL_DATABASE: 'hapi'
MYSQL_USER: 'admin'
MYSQL_PASSWORD: 'admin'
MYSQL_ROOT_PASSWORD: 'admin'
volumes:
- hapi-fhir-mysql:/var/lib/mysql
volumes:
hapi-fhir-mysql:
Raed MO (Jun 28 2020 at 14:45):
@Jens Villadsen Can you please help with the file ?
Raed MO (Jun 28 2020 at 14:47):
I tried to deploy the normal way with WAR file in tomcat but it didn't work, the file is too big 144 MB and if I copy it into /webapps and execute it from host manager it says unable to start
I'm stuck
Jose Costa Teixeira (Jun 28 2020 at 14:48):
@Raed MO perhaps you should ask this in the #hapi stream
Jose Costa Teixeira (Jun 28 2020 at 15:51):
I think I understand roughly the difference, at least partially. ClinicalUseIssue is the set of relations that are defined/known between a medication and a condition (I like the fact it exposes relations). And ConditionDefinition contains other set of information (still seems definitional).
Jose Costa Teixeira (Jun 28 2020 at 15:52):
I don't want to say they are fully overlapping, just a bit of overlap and perhaps scoping would be useful. If I were to implement something like a catalog, before I raised this, I would probably start with ConditionDefinition, based on the name and scope.
Jose Costa Teixeira (Jun 28 2020 at 15:56):
It's fairly definitional that a tablet is made of a certain substance. If it was a different substance it would have to be a different tablet.
But its price and how it is used are subject to variation, over time and from place to place. So those are less definitional.
Correct. Price is actually something that is asserted: some authority determines / defines the price for a drug- so I would not call it "knowledge".
If Definition means "intrinsic", i don't know how that will work - and I don't have a better word for that.
Jose Costa Teixeira (Jun 28 2020 at 16:06):
Neither the word definition nor knowledge is perfectly, err, defined, to only mean one thing in all contexts, but there are some differences.
perhaps it' s just me, or my romance language background, but those differences have not been clear, and that is the feeedback I give to you native speakers.
"Continuums" and overlapping concepts are a fact of life but we do what we can.
Precisely, hence my point - if in reality the boundaries are not clear, our model should not try to assert them by assigning different words that different people may understand differently.
Jose Costa Teixeira (Jun 28 2020 at 16:08):
anyway, this discussion has already been clarifying . Hopefully the fact that I'm asking might may help consider if we need further clarification
Lloyd McKenzie (Jun 28 2020 at 16:13):
I do lean towards Jose's opinion that ConditionKnowledge would be a better label. As used elsewhere in FHIR, Definition describes an activity that could potentially be instantiated. Knowledge describes information known about a kind of thing.
Rik Smithies (Jun 28 2020 at 17:25):
@Jose Costa Teixeira ConditionDefinition is definitional, yes, agree, but imho it just doesn't define a condition.
@Lloyd McKenzie agree that "Knowledge" is better. I am not sure "Condition" is ideal either, since that might seem patient/instance centric. Hence suggestion "DiagnosisKnowledge" or maybe "DiagnosingKnowledge"
Jose Costa Teixeira (Jun 28 2020 at 18:41):
Why diagnosis? That presumes someone has actually inferred some knowledge from existing information
Jose Costa Teixeira (Jun 28 2020 at 18:44):
For example, I would think that we should be able to express either
- Contraindications:
- "if hepatitis"
- "if liver function impaired"
- "if ggt > 35 IU/l"
I'd think only first one is diagnosis...?
Jose Costa Teixeira (Jun 28 2020 at 18:46):
btw, diaGNOSIS_KNOWledge is redundant in terms, right?
Jose Costa Teixeira (Jun 28 2020 at 20:19):
I do lean towards Jose's opinion that ConditionKnowledge would be a better label. As used elsewhere in FHIR, Definition describes an activity that could potentially be instantiated. Knowledge describes information known about a kind of thing.
I don't remember having that opinion :) ... But I guess the boundary clarification is useful.
Grahame Grieve (Jun 28 2020 at 20:49):
ConditionDefinition (or ConditionKnowledge, which I don't mind), links it to the Condition resource
Rik Smithies (Jun 28 2020 at 21:18):
@Jose Costa Teixeira we are at cross purposes, I am not talking about a contraindication being a diagnosis, but a ConditionDefinition being about the process of diagnosing - which is what it claims. ("...how the condition is identified" == diagnosis)
And yes both words ultimately mean knowledge but you can have knowledge about knowledge, right? ;-)
@Grahame Grieve Yes, I understand the intent, but I am questioning the intent.
If I hear the word "condition" I tend to think of it in the context of something needing managing for an actual patient (though I realise that is not the only interpretation).
Whereas, say "disease", as an example, makes me think of the medical issue in the abstract.
I assume this is not about the management of any particular instance of a disease (or other issue), but the disease in general.
Grahame Grieve (Jun 28 2020 at 21:26):
but it's not contained to 'disease'. Is pregnancy a disease?
Rik Smithies (Jun 28 2020 at 21:28):
Disease was just a example of a word that is not as patient centric as "condition". I am not suggesting it be called disease.
Grahame Grieve (Jun 28 2020 at 21:29):
this same discussion led us to the name Condition. I think that "something needing managing for an actual patient" is what the resource is about
Rik Smithies (Jun 28 2020 at 21:30):
Condition is about that, yes. Patient context. But does this new resource have a patient context? I thought not.
Rik Smithies (Jun 28 2020 at 21:30):
I though medical information in the abstract
Grahame Grieve (Jun 28 2020 at 21:34):
right. There's no patient context. It's information about the underlying condition itself
Rik Smithies (Jun 28 2020 at 21:36):
well yes but I think, and maybe I am the only one :-) that the word "condition" somewhat implies it is related to a patient.
Lloyd McKenzie (Jun 28 2020 at 22:28):
Condition has a patient (or Group) context but there's no reason to think that ConditionDefinition would - any more than we'd expect ObservationDefinition to.
Rik Smithies (Jun 28 2020 at 22:46):
True, but we are considering "ConditionKnowledge", because it really isn't a definition of a condition :-) Without the (inaccurate imho) word "definition", it sounds more like supporting information about an actual condition.
Jose Costa Teixeira (Jun 29 2020 at 07:46):
I think definition is not limited to the core definition, but can be other attributes. I don't understand thte boundary "intrinsic vs not intrinsic".
Jose Costa Teixeira (Jun 29 2020 at 07:50):
For what it's worth, here a couple of terms we needed to sharpen in a past project, not sure whether/how they apply:
- Definition: when we provide attributes about something, to define it. For a medication, this is the price, presentation, substances, indications...
- Specification: when someone indicates a set of attributes to indicate what they mean. In a prescription, this would be the GP saying "Paracetamol 1g, box of 20"
- Identification: when someone takes the set of attributes, and identifies what was meant. In a prescription, this is the pharmacist "Paracetamol 1g box of 20. I have a few of those, in capsules or tablets or ..."
Jose Costa Teixeira (Jun 29 2020 at 07:55):
I'd like to align this glossary with whatever FHIR is coming up with, this is why I ask these questions. We have substanceDefinition and medicinalProductDefinition, and medicationKnowledge. We don't have a medicineKnowledge even if in English there is also some difference between Medicine and Medication.
Jose Costa Teixeira (Jun 29 2020 at 07:55):
(should I push this to a new thread)?
liu2972505 (Jun 29 2020 at 11:26):
Jose Costa Teixeira said:
Neither the word definition nor knowledge is perfectly, err, defined, to only mean one thing in all contexts, but there are some differences.
perhaps it' s just me, or my romance language background, but those differences have not been clear, and that is the feeedback I give to you native speakers.
"Continuums" and overlapping concepts are a fact of life but we do what we can.
Precisely, hence my point - if in reality the boundaries are not clear, our model should not try to assert them by assigning different words that different people may understand differently.
Rik Smithies (Jun 29 2020 at 13:36):
@Jose Costa Teixeira
Intrinsic means a property of the thing itself, about its own nature or essence. "not intrinsic", or extrinsic, means the opposite, so perhaps information relating to a thing.
Definition is usually "intrinsic", by definition ;-) But maybe for your project you have intrinsic and extrinsic definitions. But since we have other words to use and "extrinsic definition" is a mouthful, using "knowledge" for extrinsic definitions seems helpful.
Your 2 and 3, specification and identification, in the context of medication, sound a lot like prescription and dispense, so maybe it is better to stick to those words, unless you really want to take things to a more abstract level.
Mothers say medicine, pharmacists say medication. And pharma tends to say medicinal product, because they are concerned with manufactured things, less so than an administered things. So those terms help keep the day to day meds resources separate from the others.
"Medicine" also refers to "the science of health", so is further overloaded.
Jose Costa Teixeira (Jun 29 2020 at 14:33):
specification and definition are not prescription and dispense. tey are the act of specifying what you want vs the act of understanding what was specified.
Jose Costa Teixeira (Jun 29 2020 at 14:35):
in some languages we do have a difference between medication,medicine (medicinal product) and medicine (science). Since the EMA is called Medicines agency, I thought "finally it's there", but apparently i was wrong
Rik Smithies (Jun 29 2020 at 16:19):
they are the act of specifying what you want vs the act of understanding what was specified.
Sounds like you need the RIM...
When medicines is plural the meaning is a little more clear. It cannot be the science. Anyway, we can discuss that forever. FHIR has made a choice, some time ago, and it's a reasonable choice, so we need to be consistent with it.
Last updated: Apr 12 2022 at 19:14 UTC
