Dear HHS Officials,

We are a dedicated research team based in Seattle, Washington; Wisconsin; and Chicago, IL, committed to enhancing the integrity and transparency of the Vaccine Adverse Event Reporting System (VAERS). Our mission is to address critical gaps in vaccine adverse event tracking, particularly during the COVID-19 mRNA vaccine rollout, by leveraging advanced analytical tools to uncover hidden patterns and ensure accurate public health oversight.

At the heart of our approach is a powerful, Python script we designed specifically for dissecting the "deltas"—the precise changes and discrepancies—between consecutive VAERS data drops. This script, available at https://chart.vaersdata.org, automates the comparison of snapshots from VAERS archives dating back to its inception, enabling rapid identification of additions, deletions, modifications, and anomalies that could otherwise go unnoticed. By preserving all changes in their original rows within a single, streamlined file, it empowers data scientists, regulators, and researchers to work more efficiently and reliably, fostering greater accountability in vaccine safety monitoring.

Our independent investigation into VAERS drop archives, cross-referenced against trusted collections like Dr. Steven Rubin's comprehensive medalerts.org repository (maintained by a computer scientist with over 40 years of experience), has already revealed concerning inconsistencies suggestive of potential data handling issues. For instance, follow-up reports—some confirming deaths—have been altered or entirely removed, with symptoms like "autopsy" deleted and vaccine types changed (e.g., from MNQ to MEN), effectively vanishing up to 10 death reports from official CDC tallies. These deltas highlight underreporting risks, such as uncounted follow-up cases (262 unique IDs identified via medalerts.org's tools) and discrepancies in batch numbers or severity classifications that persist across drops. Such findings underscore the urgent need for robust, verifiable delta analysis to safeguard data integrity and prevent undercounting of adverse events.

We believe adopting our Python script could transform HHS's VAERS auditing processes, providing an indispensable tool for proactive anomaly detection, streamlined compliance reviews, and evidence-based policy decisions. Early results from our delta-focused analyses, shared in 2023–2024 articles at https://deepdots.substack.com (attached as Exhibits A–F; authored by Gary Hawkins), demonstrate its value. Example exhibits are outlined below. More could be forthcoming.

So far, our work-in-progress shows clear areas that need attention:

• A. Detailed patient stories (narrative fields) were removed from reports outside the U.S., limiting full review (Exhibit A);
• B. Follow-up reports, including some that confirmed deaths, were later deleted (Exhibit B);
• C. Batch numbers (lot data) for Pfizer and Moderna are missing, blocking the ability to follow specific groups (Exhibit C);
• D. Serious blood clots in the lungs (pulmonary embolisms) appear under-reported due to weak labels (tagging) (Exhibit D);
• E. Many batch code mistakes (lot code errors) were fixed over time, pointing to early tracking problems (Exhibit E);
• F. Some severe reactions were labeled as mild, softening their real impact (Exhibit F).

We are still building this project and believe these early signals could lead to stronger VAERS practices and better public health protection. We would value your interest and input at this stage.

We invite you to:

Review the attached exhibits and let us know your thoughts; Share any insights on how VAERS data is managed; Explore ways we might work together to improve reporting.

Thank you for considering this developing work. We are ready to provide updates, raw data, or a live walk-through at your convenience over your preferred tele-communication platform.

Sincerely,

Jason Page

Exhibits

Exhibit A: Removal of Free-Text Fields

Source: "New VAERS Flat File: Easy Data Mining" (June 12, 2023)
Issue: Free-text fields (e.g., symptom narratives, lab data) for foreign VAERS reports were removed, limiting global adverse event analysis.
Impact: Obscures safety signals by reducing contextual data.

Exhibit B: Deletion of Follow-Up Reports

Source: "Undeleting a Gigabyte of Data Purged" (October 3, 2023)
Issue: Follow-up reports confirming deaths were deleted, with only initial reports retained.
Impact: Underreports fatalities, skewing safety analyses.

Exhibit C: Missing Pfizer & Moderna Lot Data (two articles)

Source 1: "Where are the missing 1,290 lots/batches in the Pfizer FOIA request response?" (August 8, 2023)
Issue: 1,290 Pfizer lot codes, linked to thousands of adverse event reports each, were absent from FOIA data.
Source 2: "Further info on the 958 missing lots/batches from the Moderna FOIA request" (August 15, 2023)
Issue: FOIA Request by ICAN provided a list of Moderna COVID vaccine lot codes and expiration dates. However, 958 lots were absent compared to a comprehensive list of 1,343 known Moderna lot codes. These missing lots are linked to thousands of adverse event reports in the Vaccine Adverse Event Reporting System (VAERS), including 19,659 harm reports tied to 103 lots without expiration dates.
Impact: Hinders tracking of batch-specific adverse events.

Exhibit D: Underreporting Pulmonary Embolisms

Source: "Far More Pulmonary Embolisms in VAERS From Covid Vaccines Than Others Are Reporting" (January 11, 2023)
Issue: 408 pulmonary embolism cases (8% sample of 4,976 reports manually reviewed) were not tagged.
Impact: Systematic underreporting of severe adverse events.

Exhibit E: Correction of Lot Code Typos

Source: "Overview: AI Fixed 150,000 Lot Numbers" (April 2, 2024)
Issue: Approximately 150,000 VAERS reports contained lot code typos (e.g., misspellings, incorrect formats) for COVID vaccines, which were corrected using AI-driven analysis.
Impact: Uncorrected typos could fragment data, leading to underreporting of adverse events associated with specific lots and complicating batch-specific safety analyses.

Exhibit F: Misclassification of Serious Reports

Source: "85% Are Serious in VAERS Reports Not Tagged as Serious" (December 4, 2023)
Issue: 85% of VAERS reports not tagged as “serious” contain symptoms meeting serious criteria (e.g., life-threatening conditions, hospitalizations).
Impact: Misclassification underestimates the severity of adverse events, potentially masking critical safety signals.