As Aravind pointed out, her 306 reports they labeled "batshit crazy" are not all 100% crazy when one reads the full writeups but just assuming they ALL are, with ~1.46 million reports of harm from a covid vaccine, there are a whopping 4,777 she could not criticize for every 1 of her 306.
On another note, the scientist in the video has successfully shown that CDC's 50 employees (I recall reading that somewhere) who are vetting reports are sleepy. Or perhaps were told to leave those reports in there to make the system criticizable? Personally as director of the CDC, I would set them aside in a separate file publicly available labeled "Questionable and/or Batshit Crazy".
Since they often DO delete reports, so long as the 306 remain in place, allowing the system to be mocked, those sit as ongoing indication that CDC is in wilful violation of its duty in the 1986 VAERS law "to achieve optimal prevention against adverse reactions to vaccines".
You can see the delight in her emotions as she picked 4 favorites of the 12 pranks involving a penis out of the 306 as her way of devalidating the 1.4 million destroyed lives, unless that was all just a parody skit to make the critics of VAERS look silly, which would not be very nice.
>>On another note, the scientist in the video has successfully shown that CDC's 50 employees (I recall reading that somewhere) who are vetting reports are sleepy
Indeed, it is remarkable, because some of those "crazy" reports are from many YEARS back.
Great great article! Now I have a quick link to slap the kool-aid drinkers with, when they point to the "Hulk" and snake bite and "bent penis" reports! There is another "type" of report that seems to discredit VAERS and used by the kool-aid drinkers. These reports will say "I saw on social media" hundreds of people died or have myocarditis. Basically there are some people but very few in the grand picture that have filed a report based on something they saw in the "news". I have seen two other reports filed against my Alaskan Baby death report that was never published story as a direct example. See here: https://www.vaersaware.com/post/this-is-the-most-infamous-deleted-vaers-report-of-all-24-000-deleted-reports . There was a spot in this piece where you mention "false positives", which was great! So much philosophy and logic in those two words that industries have sprung up around it and in the context we are using it here. Unsupervised Extraction of Diagnosis Codes from EMRs Using Knowledge-Based and Extractive Text Summarization Techniques is such an endeavor and beaten path: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5524149/. Speaking of "false positives" have you noticed how much the "myocarditis red box" has come down in OpenVAERS? I'm talking above and beyond from the supposed lose of "signal" from the foreign data scrubbing. An additional ~9K have been corrected because of this "false positive" you speak of. I have lamented for over three months these openvaers myocarditis stats and any stats Liz produces that rely on data "extraction" from the summary narrative and I finally called her and Rose to task and behind the scenes with all the VIP's in the email thread audience. This was my original exchange with Liz when I questioned how she was arriving at her then 52K myocarditis reports? https://i.imgur.com/22P2JcR.jpg
You should consider putting your blog posts front and center on the VaersAware website.
There are a lot of dashboards - which the technical folks might appreciate - but I think a much larger group want to see static posts interspersed with a few summary images. You can always link to the dashboard from the summary image's caption if you like.
Edit:
Also another great help would be to add a giant VAERS FAQ page and answer all these questions, with the answers linking to more detailed articles where it makes sense.
I think your website will automatically become a link magnet from the "anti-vaxxers" :-)
Although I have a feeling that won't help with any SEO, given search engines are now more or less actively hostile to such content.
Thank you for your detailed explanation and assessment on this particular situation. I know as much as anyone the immense amount of sophistication that goes into creating sound, valid and cogent coding rules for “Unsupervised extraction of diagnosis codes from EMRs Using Knowledge-Based and Extractive Text Summarization Techniques”. Industries have sprung up going down this beaten path (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5524149/ ). It is an area I was very familiar with managing 12 EMR systems at once for an enterprise level MSO/HMO here is San Jose aka Verity Medical Foundation. I will tell you this however, it’s going to take more than an engineer to fine tune the logic but it’s a great start. Just focusing on the natural language within summaries to exclude the obvious “false positives” is a huge first step. Of course, there will be great challenges determining previous patient history and if it was myocarditis history from years ago, or history from the first round of covid jabs last year? Yikes. In the perfect world the algorithms could deduce lab results where applicable. That’s a whole other level, but something I’m not unfamiliar with. It’s something we achieved in my journey with Chris Riedel of Hunter Laboratories in the eventual creation of the HunterHeart Panel. Maybe Dr. McCullough is familiar with this advanced heart panel? After all it was issued a McKesson Moldx “Z” code by CMS and can be order by any physician in the country. https://casetext.com/case/hunterheart-inc-v-bio-reference-labs-inc . I digress.
Thank you Liz for taking the time to perfect your system. Coding and algorithms aside, I will be the first to say there is probably much more “carditis” than is being capture by VAERS summary narratives and symptom fields. The mere fact ~40K reports of “Chest Pain” that hadn’t been clinically diagnosed before reports were submitted. Hence no “carditis” symptoms could be extracted by CDC’s inhouse certified professional coders. How many of these people walking around today have since developed clinical “carditis”? How many more are now since dead? How many kids with reports only for “Inappropriate age” are now walking around with myocarditis? We will never know since ONLY INITIAL REPORTS ARE MADE PUBLIC! I don’t deserve a thank you, but I can tell you, your welcome. I did state back in September in no uncertain terms I was concerned with the statistics presented. I do now see your statistics dropped from 35,917 to 26,045 this week. I can state given my personal in-depth analysis there is an additional ~8-9K reports that will need additional scrutiny, and I’m sure you’ll address them on future iterations. Keep up the great work. Let me know if you need help. Having full transparency to your code or at least those reports in the gray area will be helpful. At the very least, it will create a discussion about the CDC’s coders what kind of malfeasance is going on. Looking forward future audits. God Bless.
James,
“Fate whispers to the warrior, ‘You cannot withstand the storm.’ The warrior whispers back, ‘I am the storm.” -Some Great Warrior
So young Aravind, what do you think of them apples? What are these implications going to do to any future musings of Dr. Rose substack articles on this subject? In hindsight this means Jessica's original assertion of "32K missing myocarditis signals/reports" was over inflated in the first place because the "false positives" were not addressed. This means Dr. Wolf's piece on the missing signal was also incorrect in hindsight. No worries at least our side is sharpening iron so we can be rock solid at Nuremburg 2.0. I have since met Dr. Henry Ealy and will be collaborating with him, he seems to have a little bit of resources and could help bring a better VAERS Interactive Dashboard forward and one that incorporates all the machine learning Python Probability spaCy Logic you speak of to better "cleanse" data all the while keeping track of every edit in effect creating a "before and after" version of VAERS which is what my current dashboards try to do! God Bless https://www.vaersaware.com/
This was openvaers response on Sunday New Years Day, keep in mind how disingenuous this admission is knowing that I was poking around the issue 3 months prior: Last fall OpenVAERS underwent its 3rd major revision. We brought in an engineer to redesign the database, make searches more efficient, and give me back my Fridays.
In the process of that revision, we rewrote most of our queries, including that of myocarditis. Since the change to the European data on November 18th, 2022, I have been reevaluating how that change affected the site counts, and subsequently, how I might refine the queries. The change dramatically affected myocarditis counts, for example, so it seemed prudent to refine the query suitably. To date, we are continuing these refinements by aiming to include only diagnostically-definitive reports; ie: to not count those that indicate a requirement for follow-up. This is extremely labor-intensive and it will take some time. Over the next few weeks some movement in numbers will occur that are due to changes / refinements in queries.
In addition, we will be adding a new level of enquiry to some of the reports whereby we will be counting the reports that required follow-up, but where this was not done (or at least not put in the public-facing data as per VAERS’ own guidelines). This has always been an issue when parsing the data and why the narrative field, in part, was used. When we have this in place we will be writing more about this. This is a significant issue with the VAERS export.
The recent change to the European data and our subsequent house-cleaning has spurred me to revisit an idea suggested to me by a friend last year to make OpenVAERS ‘open source', and to show all the queries behind the numbers and allow others to improve the algorithm. Once we have reviewed our queries we will embark on the road to making our query engine open source. We are not yet sure what form this will take—it may be a GitHub repository, it may be a page with our algorithms available for view and possibly a changelog. At the moment I am leaning toward the latter since we lack the infrastructure to make a git repository work the way it should.
OpenVAERS has always been a work-in-progress and we are committed to transparency. We walk a difficult line trying to mine the data to give meaningful information and continually improve, while at the same time needing to maintain public trust. I hope that adding open sourcing to the site will allow both to continually occur. This is both the beauty and the conundrum of the internet.
Once we open our algorithms there will be a method for directly contacting us to suggest improvements to, or if you have questions about, our queries. In the meantime, we ask you to contact me or the site directly with any suggestions or issues you may have. Both my direct email and the public email are readily available on this thread.
I wish everyone a Happy New Year! May it be filled with productive work, more joy than sorrow, and JUSTICE.
As Aravind pointed out, her 306 reports they labeled "batshit crazy" are not all 100% crazy when one reads the full writeups but just assuming they ALL are, with ~1.46 million reports of harm from a covid vaccine, there are a whopping 4,777 she could not criticize for every 1 of her 306.
On another note, the scientist in the video has successfully shown that CDC's 50 employees (I recall reading that somewhere) who are vetting reports are sleepy. Or perhaps were told to leave those reports in there to make the system criticizable? Personally as director of the CDC, I would set them aside in a separate file publicly available labeled "Questionable and/or Batshit Crazy".
Since they often DO delete reports, so long as the 306 remain in place, allowing the system to be mocked, those sit as ongoing indication that CDC is in wilful violation of its duty in the 1986 VAERS law "to achieve optimal prevention against adverse reactions to vaccines".
You can see the delight in her emotions as she picked 4 favorites of the 12 pranks involving a penis out of the 306 as her way of devalidating the 1.4 million destroyed lives, unless that was all just a parody skit to make the critics of VAERS look silly, which would not be very nice.
>>On another note, the scientist in the video has successfully shown that CDC's 50 employees (I recall reading that somewhere) who are vetting reports are sleepy
Indeed, it is remarkable, because some of those "crazy" reports are from many YEARS back.
Great great article! Now I have a quick link to slap the kool-aid drinkers with, when they point to the "Hulk" and snake bite and "bent penis" reports! There is another "type" of report that seems to discredit VAERS and used by the kool-aid drinkers. These reports will say "I saw on social media" hundreds of people died or have myocarditis. Basically there are some people but very few in the grand picture that have filed a report based on something they saw in the "news". I have seen two other reports filed against my Alaskan Baby death report that was never published story as a direct example. See here: https://www.vaersaware.com/post/this-is-the-most-infamous-deleted-vaers-report-of-all-24-000-deleted-reports . There was a spot in this piece where you mention "false positives", which was great! So much philosophy and logic in those two words that industries have sprung up around it and in the context we are using it here. Unsupervised Extraction of Diagnosis Codes from EMRs Using Knowledge-Based and Extractive Text Summarization Techniques is such an endeavor and beaten path: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5524149/. Speaking of "false positives" have you noticed how much the "myocarditis red box" has come down in OpenVAERS? I'm talking above and beyond from the supposed lose of "signal" from the foreign data scrubbing. An additional ~9K have been corrected because of this "false positive" you speak of. I have lamented for over three months these openvaers myocarditis stats and any stats Liz produces that rely on data "extraction" from the summary narrative and I finally called her and Rose to task and behind the scenes with all the VIP's in the email thread audience. This was my original exchange with Liz when I questioned how she was arriving at her then 52K myocarditis reports? https://i.imgur.com/22P2JcR.jpg
You should consider putting your blog posts front and center on the VaersAware website.
There are a lot of dashboards - which the technical folks might appreciate - but I think a much larger group want to see static posts interspersed with a few summary images. You can always link to the dashboard from the summary image's caption if you like.
Edit:
Also another great help would be to add a giant VAERS FAQ page and answer all these questions, with the answers linking to more detailed articles where it makes sense.
I think your website will automatically become a link magnet from the "anti-vaxxers" :-)
Although I have a feeling that won't help with any SEO, given search engines are now more or less actively hostile to such content.
Here was Dr. Rose's response to Liz about one minute after Liz's admission: 100% :D Happy New Year you genius.
_______________________________________________________________________-
Here was Dr. Lyons-Wieler's response to Liz: We should build our own vaccine adverse events reporting system that captures HIPAA permission to
check parity on a small % of reports and give people the option of making their own medical data public
if they care to.
To know your enemy you must become your enemy - Sun Tzu, The Art of War
JLW
__________________________________________________________
Here was my response to Liz and JLW:
Liz,
Thank you for your detailed explanation and assessment on this particular situation. I know as much as anyone the immense amount of sophistication that goes into creating sound, valid and cogent coding rules for “Unsupervised extraction of diagnosis codes from EMRs Using Knowledge-Based and Extractive Text Summarization Techniques”. Industries have sprung up going down this beaten path (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5524149/ ). It is an area I was very familiar with managing 12 EMR systems at once for an enterprise level MSO/HMO here is San Jose aka Verity Medical Foundation. I will tell you this however, it’s going to take more than an engineer to fine tune the logic but it’s a great start. Just focusing on the natural language within summaries to exclude the obvious “false positives” is a huge first step. Of course, there will be great challenges determining previous patient history and if it was myocarditis history from years ago, or history from the first round of covid jabs last year? Yikes. In the perfect world the algorithms could deduce lab results where applicable. That’s a whole other level, but something I’m not unfamiliar with. It’s something we achieved in my journey with Chris Riedel of Hunter Laboratories in the eventual creation of the HunterHeart Panel. Maybe Dr. McCullough is familiar with this advanced heart panel? After all it was issued a McKesson Moldx “Z” code by CMS and can be order by any physician in the country. https://casetext.com/case/hunterheart-inc-v-bio-reference-labs-inc . I digress.
Thank you Liz for taking the time to perfect your system. Coding and algorithms aside, I will be the first to say there is probably much more “carditis” than is being capture by VAERS summary narratives and symptom fields. The mere fact ~40K reports of “Chest Pain” that hadn’t been clinically diagnosed before reports were submitted. Hence no “carditis” symptoms could be extracted by CDC’s inhouse certified professional coders. How many of these people walking around today have since developed clinical “carditis”? How many more are now since dead? How many kids with reports only for “Inappropriate age” are now walking around with myocarditis? We will never know since ONLY INITIAL REPORTS ARE MADE PUBLIC! I don’t deserve a thank you, but I can tell you, your welcome. I did state back in September in no uncertain terms I was concerned with the statistics presented. I do now see your statistics dropped from 35,917 to 26,045 this week. I can state given my personal in-depth analysis there is an additional ~8-9K reports that will need additional scrutiny, and I’m sure you’ll address them on future iterations. Keep up the great work. Let me know if you need help. Having full transparency to your code or at least those reports in the gray area will be helpful. At the very least, it will create a discussion about the CDC’s coders what kind of malfeasance is going on. Looking forward future audits. God Bless.
James,
“Fate whispers to the warrior, ‘You cannot withstand the storm.’ The warrior whispers back, ‘I am the storm.” -Some Great Warrior
Sincerely,
Albert
________________________________________________________-
So young Aravind, what do you think of them apples? What are these implications going to do to any future musings of Dr. Rose substack articles on this subject? In hindsight this means Jessica's original assertion of "32K missing myocarditis signals/reports" was over inflated in the first place because the "false positives" were not addressed. This means Dr. Wolf's piece on the missing signal was also incorrect in hindsight. No worries at least our side is sharpening iron so we can be rock solid at Nuremburg 2.0. I have since met Dr. Henry Ealy and will be collaborating with him, he seems to have a little bit of resources and could help bring a better VAERS Interactive Dashboard forward and one that incorporates all the machine learning Python Probability spaCy Logic you speak of to better "cleanse" data all the while keeping track of every edit in effect creating a "before and after" version of VAERS which is what my current dashboards try to do! God Bless https://www.vaersaware.com/
This was openvaers response on Sunday New Years Day, keep in mind how disingenuous this admission is knowing that I was poking around the issue 3 months prior: Last fall OpenVAERS underwent its 3rd major revision. We brought in an engineer to redesign the database, make searches more efficient, and give me back my Fridays.
In the process of that revision, we rewrote most of our queries, including that of myocarditis. Since the change to the European data on November 18th, 2022, I have been reevaluating how that change affected the site counts, and subsequently, how I might refine the queries. The change dramatically affected myocarditis counts, for example, so it seemed prudent to refine the query suitably. To date, we are continuing these refinements by aiming to include only diagnostically-definitive reports; ie: to not count those that indicate a requirement for follow-up. This is extremely labor-intensive and it will take some time. Over the next few weeks some movement in numbers will occur that are due to changes / refinements in queries.
In addition, we will be adding a new level of enquiry to some of the reports whereby we will be counting the reports that required follow-up, but where this was not done (or at least not put in the public-facing data as per VAERS’ own guidelines). This has always been an issue when parsing the data and why the narrative field, in part, was used. When we have this in place we will be writing more about this. This is a significant issue with the VAERS export.
The recent change to the European data and our subsequent house-cleaning has spurred me to revisit an idea suggested to me by a friend last year to make OpenVAERS ‘open source', and to show all the queries behind the numbers and allow others to improve the algorithm. Once we have reviewed our queries we will embark on the road to making our query engine open source. We are not yet sure what form this will take—it may be a GitHub repository, it may be a page with our algorithms available for view and possibly a changelog. At the moment I am leaning toward the latter since we lack the infrastructure to make a git repository work the way it should.
OpenVAERS has always been a work-in-progress and we are committed to transparency. We walk a difficult line trying to mine the data to give meaningful information and continually improve, while at the same time needing to maintain public trust. I hope that adding open sourcing to the site will allow both to continually occur. This is both the beauty and the conundrum of the internet.
Once we open our algorithms there will be a method for directly contacting us to suggest improvements to, or if you have questions about, our queries. In the meantime, we ask you to contact me or the site directly with any suggestions or issues you may have. Both my direct email and the public email are readily available on this thread.
I wish everyone a Happy New Year! May it be filled with productive work, more joy than sorrow, and JUSTICE.