-
Notifications
You must be signed in to change notification settings - Fork 822
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add "cause" as available property for MedicalCondition #2878
Comments
This issue is being nudged due to inactivity. |
cause was removed after a cleanup effort as described here (making https://schema.org/MedicalCause orphaned?) These examples tripped me up as well when I was working on preparing an implementation, especially as causeOf is (still) available. @danbri, as cause was removed, what does this mean for the inverse property causeOf? Should this then be removed too? Or is this an example where MedicalCause would always lead to MedicalCondition and thus causeOf is a valid property, but MedicalCondition can have many causes and would become ambiguous? |
The markup wouldn’t always be pretty but causeOf can express anything that
the old /cause property could express, so long as the entities are
appropriately structured
More important would be to identify anyone who actually does something with
this markup…
…On Sat, 21 Aug 2021 at 10:57, Dutch-Al ***@***.***> wrote:
cause was removed after a cleanup effort as described here (making
https://schema.org/MedicalCause orphaned?)
#2435 <#2435>
So examples should be updated.
These examples tripped me up as well when I was working on preparing an
implementation, especially as causeOf is (still) available.
@danbri <https://github.com/danbri>, as cause was removed, what does this
mean for the inverse property causeOf? Should this then be removed too? Or
is this an example where MedicalCause would always lead to MedicalCondition
and thus causeOf is a valid property, but MedicalCondition can have many
causes and would become ambiguous?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2878 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABJSGK253JYQF6KHRLA6MTT552B3ANCNFSM45CWUX3A>
.
|
Great, thanks for making that clear that it is OK to keep causeOf.
Glad you asked! Allow me to go off topic! I think anyone trying to use health-lifesci schema in their website is only doing this convinced that this has some "SEO benefit", or possible future SEO benefit. The SEO community is always looking for an authoritative answer on till what extent structured data types apart from officially documented are supported/processed by Google (Given Google's market share, if any other search engine would support more, due to cost you'd still not implement it. For Baidu, it does not support Schema.org). So we hear this:
But then 2 days later Gary Illyes says:
So, that leaves us pretty much in the dark till what extent it helps with "entity discovery". But, how how much this would be needed at all now and in the next 5 years. If you insert your medical text into https://aws.amazon.com/comprehend/medical/ you definitely feel less compelled to do all the classification manually. Then we also have the situation where Google is directly working with partners to get medical information into the KG (in supported countries such as US, BR, IN, ID): https://support.google.com/websearch/answer/2364942?hl=en My background in this: I am working on a hospital chain that is migrating CMS and improving implementation of structured data. I started to notice the state of maintenance, especially after looking at MedicalSpecialty. This made me apprehensive to do a full fledged implementation that later may see many revisions apart from not generating much benefit at all ever or at first. I also read your response in #2894 and the discussion in #1598. It seems to me a case of chicken and egg. No adoption as no support from Google and vice versa. Cheers, P.S. defining a medical condition with all of it's symptoms, preventions, possible treatments etc etc is already hard enough in text, let alone then getting this structure into a database. And, if anyone has done this hard work, what would then stop me from scrape this off this website and publish it as my own, maybe augmented with my own data? |
Search engines philosophy toward promoting and sharing health informationThe publishing and sharing of health information that has the potential to save lives should always be openly shared. Google, Bing, and others have long acknowledged this and contribute back to that cause. Google has demonstrated their philosophy and stance on this in my eyes personally by allowing the world to query much of the healthcare sector as a public good. https://console.cloud.google.com/marketplace/browse?filter=solution-type:dataset&filter=category:health I think it's acceptable that a general search against a search engine like Google for a medical condition can be done nowadays and users are presented some quick facts and somewhat authoritative information, while presenting well to users that those quick facts are
SEO of health information with a profit agendaAs health information as a public good, I don't ever look or find appealing that health information is considered IP (intellectual property). For those looking to profit off of health information, there is nothing for us to stop them from that. At the same token, its very unlikely that well structured data on pages from entities that seem to have a "profit agenda" will likely ever show up highly ranked in search engine results. Case in point, https://www.google.com/search?q=heart attack will never show those kinds of non-authoritative sites even in the first 10 pages of results on any search engine. It is well known and established that trying to SEO for any "health information" with a mindset for a "profit agenda" is a waste of time. Those that provide "health information" for the public good are always ranked much more highly and considered a valuable resource of knowledge toward possibly saving human lives. In Summary:I think the benefit for any "hospital chain" to offer deeply structured data of medical conditions is probably not needed. This is already covered by scientific and research communities. MedicalProcedure, MedicalTherapy, and MedicalTest are much more important for any "hospital chain" to let the public know what availableService's they offer along with highlighting any particular MedicalSpecialty that they are well equipped and certified to help the public with should be a focus for those types of entities that can provide care and treatment. |
I'm head over heels for this one.
:-)
…On Sat, Aug 21, 2021 at 12:08 PM Dutch-Al ***@***.***> wrote:
Great, thanks for making that clear that it is OK to keep causeOf.
More important would be to identify anyone who actually does something with
this markup…
Glad you asked! Allow me to go off topic!
I think anyone trying to use health-lifesci schema in their website is
only doing this convinced that this has some "SEO benefit", or possible
future SEO benefit. The SEO community is always looking for an
authoritative answer on till what extent structured data types apart from
officially documented are supported/processed by *Google* (Given Google's
market share, if any other search engine would support more, due to cost
you'd still not implement it. For Baidu, it does not support Schema.org).
So we hear this:
Mueller notes that Google does not crawl structured data to learn more
about the page.
https://www.searchenginejournal.com/google-structured-data-rarely-gives-us-unique-information/389729/
But then 2 days later Gary Illyes says:
but we do extract all kinds of structured data it's not just what the Rich
Result testing tool
validates, but pretty much any structured data that you put on the page
https://twitter.com/thetafferboy/status/1335361054654783492
So, that leaves us pretty much in the dark till what extent it helps with
"entity discovery". But, how how much this would be needed at all now and
in the next 5 years. If you insert your medical text into
https://aws.amazon.com/comprehend/medical/ you definitely feel less
compelled to do all the classification manually.
Then we also have the situation where Google is directly working with
partners to get medical information into the KG (in supported countries
such as US, BR, IN, ID):
https://support.google.com/websearch/answer/2364942?hl=en
This data is not connected to the structured data at all. The KG from Mayo
Clinic on "heart attack" in the US is much more complete than the mark-up
on MedicalCondition:
https://www.mayoclinic.org/diseases-conditions/heart-attack/symptoms-causes/syc-20373106
My background in this: I am working on a hospital chain that is migrating
CMS and improving implementation of structured data. I started to notice
the state of maintenance, especially after looking at MedicalSpecialty.
This made me apprehensive to do a full fledged implementation that later
may see many revisions apart from not generating much benefit at all ever
or at first.
I also read your response in #2894
<#2894> and the discussion
in #1598 <#1598>. It seems
to me a case of chicken and egg. No adoption as no support from Google and
vice versa.
Cheers,
Aldert
P.S. defining a medical condition with all of it's symptoms, preventions,
possible treatments etc etc is already hard enough in text, let alone then
getting this structure into a database. And, if anyone has done this hard
work, what would then stop me from scrape this off this website and publish
it as my own, maybe augmented with my own data?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#2878 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADFVNUTYPQXCAEX7CKOBNBDT57MQ5ANCNFSM45CWUX3A>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email>
.
|
@thadguidry thanks for weighing in on this. I agree with your philosophical POV. Justifying implementation costsThe SEO aspect remains unanswered. There is no evidence that MedicalCondition or MedicalProcedure, -Therapy, -Test are consumed by Google. When I read Dan's comments like below, I don't get convinced otherwise.
If it costs $30k to implement these schemas and, more importantly, another X amount to maintain it, how would I justify such costs? Meanwhile it could still make sense to modularize the actual content so that the schemas eventually could be used (and for improved display of content to visitors). Examples in the wildMedicalProcedure has been implemented on less websites than MedicalCondition (if I do a source code search with https://publicwww.com/). BuiltWith tracks usage of MedicalCondition and MedicalSpecialty. Less than 100 of all top 1 million websites use these schemas: https://trends.builtwith.com/framework/MedicalCondition-Schema. BuiltWith does not support the more obscure Medical schemas. Even though well known sites are using the schema (pennmedicine.org, franciscanhealth.org, CDC.gov, MayoClinic.org, NHS.uk) ), the question remains, did it help anyone that they added the schema mark-up on the front-end of the website? AFAIK, the medical condition KG is not populated by Google crawling the JSON data, but relies on a specific partnership. Evidence also shows that Mayo Clinic KG does not match the data inside MedicalCondition on their webpage. No standardized implementationNHS has also created modularized content and is now available in the KG (https://www.nhs.uk/conditions/heart-attack/) Their approach to the Medical schema vastly differs from Mayo Clinic with a heavy reliance on HealthTopicContent instead. MedicalCondition is found under about of MedicalWebPage: KG ahead of Schema.org for MedicalProcedureThe KG below provides important topics of the procedure that are not part of a property under https://schema.org/MedicalProcedure. The accordion could offer different subtopics depending on the procedure. Still uncertain on benefit for mankind
All such information is important for the public and it can be modularized accordingly (for which the schemas are a big inspiration). But the public interfacing with website does not require Schema at all... So I fail to understand why you say it is important (to implement). You'd implement Schema so that the public interfacing with a search engine quicker/better finds the answer through improved matching of content with query. I can appreciate some vagueness on actual support or, more importantly, the benefit to prevent overpromising. I am not looking for special ranking boosts or fancy search features. But I do need to understand if search engines are helped in any way (no matter how little) with the medical structured data, or whether Googlebot blatantly ignores it and will always do so. As long as this is unclear, how can I become an advocate for implementing Medical Schemas and waste my time, others people time, and money for no good? |
I wouldn’t recommend putting this markup in solely in hope Google uses it.
I am not saying it would be entirely useless or what the future might hold,
but I suspect you’d often get more value from focussing on usability, web
accessibility, performance on mobile, and all the other things that help
people make the most of your content.
…On Sun, 22 Aug 2021 at 13:31, Dutch-Al ***@***.***> wrote:
@thadguidry <https://github.com/thadguidry> thanks for weighing in on
this. I agree with your philosophical POV.
Justifying implementation costs
The SEO aspect remains unanswered. There is no evidence that
MedicalCondition or MedicalProcedure, -Therapy, -Test are consumed by
Google. When I read Dan's comments like below, I don't get convinced
otherwise.
More important would be to identify anyone who actually does something
with this markup…
@danbri <https://github.com/danbri> <#2878 (comment)
<#2878 (comment)>
>
*Context* - This is a proposal from Google based on our experience
consuming schema.org HealthAspectEnumeration markup and working with
health-related content. *If it were accepted, it could (this is a tricky
area...)* make it easier for us and others to understand the different
kinds of health aspects addressed by a piece of Web content.
@danbri <https://github.com/danbri> <#2799 (comment)
<#2799 (comment)>>
(Emphasis mine)
If it costs $30k to implement these schemas and, more importantly, another
X amount to maintain it, how would I justify such costs? Meanwhile it could
still make sense to modularize the actual content so that the schemas
eventually could be used (and for improved display of content to visitors).
Examples in the wild
MedicalProcedure has been implemented on less websites than
MedicalCondition (if I do a source code search with https://publicwww.com/).
BuiltWith tracks usage of MedicalCondition and MedicalSpecialty. Less than
100 of all top 1 million websites use these schemas:
https://trends.builtwith.com/framework/MedicalCondition-Schema. BuiltWith
does not support the more obscure Medical schemas.
Even though well known sites are using the schema (pennmedicine.org,
franciscanhealth.org, CDC.gov, MayoClinic.org, NHS.uk) ), the question
remains, did it help anyone that they added the schema mark-up on the
front-end of the website? AFAIK, the medical condition KG is *not*
populated by Google crawling the JSON data, but relies on a specific
partnership. Evidence also shows that Mayo Clinic KG does not match the
data inside MedicalCondition on their webpage.
No standardized implementation
NHS has also created modularized content and is now available in the KG (
https://www.nhs.uk/conditions/heart-attack/)
[image: image]
<https://user-images.githubusercontent.com/46360035/130350296-faf660eb-5bcb-41d8-8275-fa2167e57c33.png>
Their approach to the Medical schema vastly differs from Mayo Clinic with
a heavy reliance on HealthTopicContent instead. MedicalCondition is found
under about of MedicalWebPage:
[image: image]
<https://user-images.githubusercontent.com/46360035/130349965-afe6720f-818a-4d64-ae8c-315dd1702fa3.png>
KG ahead of Schema.org for MedicalProcedure
The KG below provides important topics of the procedure that are not part
of a property under https://schema.org/MedicalProcedure. The accordion
could offer different subtopics depending on the procedure.
[image: image]
<https://user-images.githubusercontent.com/46360035/130350386-e81df546-992f-4bed-86cc-ac5cde3b11bb.png>
Still uncertain on benefit for mankind
MedicalProcedure, MedicalTherapy, and MedicalTest are much more important
for any "hospital chain" to let the public know what availableService's
they offer along with highlighting any particular MedicalSpecialty that
they are well equipped and certified to help the public with should be a
focus for those types of entities that can provide care and treatment.
All such information is important for the public and it can be modularized
accordingly (for which the schemas are a big inspiration). But the public
interfacing with website does not require Schema at all... So I fail to
understand why you say it is important (to implement). You'd implement
Schema so that the public interfacing with a search engine quicker/better
finds the answer through improved matching of content with query.
I can appreciate *some* vagueness on actual support or, more importantly,
the benefit to prevent overpromising. I am *not* looking for special
ranking boosts or fancy search features. But I do need to understand if
search engines are helped in any way (no matter how little) with the
medical structured data, or whether Googlebot blatantly ignores it and will
always do so. As long as this is unclear, how can I become an advocate for
implementing Medical Schemas and waste my time, others people time, and
money for no good?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2878 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABJSGI2VA3AYMKIFIS5TTLT6DN2FANCNFSM45CWUX3A>
.
|
I should add, I have not read this full thread and am ooo for the week -
will reply in more detail in Sept
…On Sun, 22 Aug 2021 at 14:47, Dan Brickley ***@***.***> wrote:
I wouldn’t recommend putting this markup in solely in hope Google uses it.
I am not saying it would be entirely useless or what the future might hold,
but I suspect you’d often get more value from focussing on usability, web
accessibility, performance on mobile, and all the other things that help
people make the most of your content.
On Sun, 22 Aug 2021 at 13:31, Dutch-Al ***@***.***> wrote:
> @thadguidry <https://github.com/thadguidry> thanks for weighing in on
> this. I agree with your philosophical POV.
> Justifying implementation costs
>
> The SEO aspect remains unanswered. There is no evidence that
> MedicalCondition or MedicalProcedure, -Therapy, -Test are consumed by
> Google. When I read Dan's comments like below, I don't get convinced
> otherwise.
>
> More important would be to identify anyone who actually does something
> with this markup…
> @danbri <https://github.com/danbri> <#2878 (comment)
> <
#2878 (comment)>
> >
>
> *Context* - This is a proposal from Google based on our experience
> consuming schema.org HealthAspectEnumeration markup and working with
> health-related content. *If it were accepted, it could (this is a tricky
> area...)* make it easier for us and others to understand the different
> kinds of health aspects addressed by a piece of Web content.
> @danbri <https://github.com/danbri> <#2799 (comment)
> <#2799 (comment)>>
> (Emphasis mine)
>
> If it costs $30k to implement these schemas and, more importantly,
another
> X amount to maintain it, how would I justify such costs? Meanwhile it
could
> still make sense to modularize the actual content so that the schemas
> eventually could be used (and for improved display of content to
visitors).
> Examples in the wild
>
> MedicalProcedure has been implemented on less websites than
> MedicalCondition (if I do a source code search with
https://publicwww.com/).
> BuiltWith tracks usage of MedicalCondition and MedicalSpecialty. Less
than
> 100 of all top 1 million websites use these schemas:
> https://trends.builtwith.com/framework/MedicalCondition-Schema.
BuiltWith
> does not support the more obscure Medical schemas.
>
> Even though well known sites are using the schema (pennmedicine.org,
> franciscanhealth.org, CDC.gov, MayoClinic.org, NHS.uk) ), the question
> remains, did it help anyone that they added the schema mark-up on the
> front-end of the website? AFAIK, the medical condition KG is *not*
> populated by Google crawling the JSON data, but relies on a specific
> partnership. Evidence also shows that Mayo Clinic KG does not match the
> data inside MedicalCondition on their webpage.
> No standardized implementation
>
> NHS has also created modularized content and is now available in the KG (
> https://www.nhs.uk/conditions/heart-attack/)
> [image: image]
> <
https://user-images.githubusercontent.com/46360035/130350296-faf660eb-5bcb-41d8-8275-fa2167e57c33.png
>
>
> Their approach to the Medical schema vastly differs from Mayo Clinic with
> a heavy reliance on HealthTopicContent instead. MedicalCondition is found
> under about of MedicalWebPage:
> [image: image]
> <
https://user-images.githubusercontent.com/46360035/130349965-afe6720f-818a-4d64-ae8c-315dd1702fa3.png
>
> KG ahead of Schema.org for MedicalProcedure
>
> The KG below provides important topics of the procedure that are not part
> of a property under https://schema.org/MedicalProcedure. The accordion
> could offer different subtopics depending on the procedure.
>
> [image: image]
> <
https://user-images.githubusercontent.com/46360035/130350386-e81df546-992f-4bed-86cc-ac5cde3b11bb.png
>
> Still uncertain on benefit for mankind
>
> MedicalProcedure, MedicalTherapy, and MedicalTest are much more important
> for any "hospital chain" to let the public know what availableService's
> they offer along with highlighting any particular MedicalSpecialty that
> they are well equipped and certified to help the public with should be a
> focus for those types of entities that can provide care and treatment.
>
> All such information is important for the public and it can be
modularized
> accordingly (for which the schemas are a big inspiration). But the public
> interfacing with website does not require Schema at all... So I fail to
> understand why you say it is important (to implement). You'd implement
> Schema so that the public interfacing with a search engine quicker/better
> finds the answer through improved matching of content with query.
>
> I can appreciate *some* vagueness on actual support or, more importantly,
> the benefit to prevent overpromising. I am *not* looking for special
> ranking boosts or fancy search features. But I do need to understand if
> search engines are helped in any way (no matter how little) with the
> medical structured data, or whether Googlebot blatantly ignores it and
will
> always do so. As long as this is unclear, how can I become an advocate
for
> implementing Medical Schemas and waste my time, others people time, and
> money for no good?
>
> —
> You are receiving this because you were mentioned.
>
>
> Reply to this email directly, view it on GitHub
> <
#2878 (comment)
>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AABJSGI2VA3AYMKIFIS5TTLT6DN2FANCNFSM45CWUX3A
>
> .
>
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#2878 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABJSGJZ6ROBHX3Q4GJKNYDT6DWWBANCNFSM45CWUX3A>
.
|
Thanks for the initial reply on this so quickly. It will help me for now this week and looking forward to further details in September, have a great break :) |
I would tread lightly here, the medical profession is notorious for revamping their coding every couple of years. The privacy issues are probably already mute though, Google seems to know which maladies I'm interested before I even search for the term. |
Outside the scope of Google's use of medical structured data, sites that apply this markup are also interested in the internal reuse of their structured data knowledge graph. The specificity permitted by bilateral relationships is useful. I would like to know more about the rationale that dictates when inverse properties are allowed, and when they are not (as with causeOf vs. cause). |
This is still needed, it was removed during the health-lifesci cleanup.
Example code block from https://schema.org/MedicalCondition (truncated)
|
Today I noticed that the JSON-LD example for MedicalCause shows it can be embedded within MedicalCondition using the "cause" property:
HOWEVER! "cause" isn't listed as an available property within https://schema.org/MedicalCondition and there isn't any entry for a "cause" property that I can find within Schema.org.
The text was updated successfully, but these errors were encountered: