mm7107e1 — Anomalies

Central finding #

Paper (Discussion): “V-safe does not capture diagnoses associated with hospitalization; however, registrants can include supplemental text for each health check-in.”

Paper (VAERS): adult booster seriousness is framed mainly through VAERS—3,004 serious reports; 37 myocarditis cases meeting the case definition; one myocarditis-associated death under investigation.

v-safe Results emphasize that boosters look milder than dose 2 and that medical care is rare. The hospitalization-diagnosis sentence admits that free-text exists, then still treats v-safe as silent on what was wrong—while free text and MedDRA in this same window record diagnosed myocarditis, MI/stents, stroke/clot, and death.

Checkbox care rates are published; the coded and typed diagnoses next to those checkboxes are not.

Cohort: booster 22 Sep 2021–6 Feb 2022, day-0–7 check-ins, mRNA primary (≥150 days to dose 3) or Janssen primary (≥60 days to dose 2), excluding mRNA→Janssen—about 1,037,186 people (paper ~721k; later freeze). Table 1 rates match; the gaps are diagnoses, coding, cardiac/clot silence, and omitted v-safe deaths. Full write-up: vsafe-analysis/mmwr/mm7107e1/anomalies_report.md.

A. “Does not capture diagnoses associated with hospitalization” — contradicted #

Paper (Discussion): “V-safe does not capture diagnoses associated with hospitalization; however, registrants can include supplemental text for each health check-in.”

When a registrant checks Hospitalization and types a diagnosis—or when MedDRA codes that diagnosis—v-safe has captured a diagnosis. Free text in this cohort includes “Was diagnosed with myocarditis,” “Myocarditis elevated troponin,” brain-clot surgery, heart bypass, STEMI with stents, and anticoagulant drips— sometimes coded as Myocarditis or Acute MI, sometimes dumped to Unevaluable event.

Why this matters: In a paper whose SAE story is VAERS myocarditis (37 confirmed; 1 death), telling readers that v-safe does not capture hospitalization diagnoses—while free text says “diagnosed with myocarditis”—is a misrepresentation of the surveillance system. It keeps active surveillance quiet exactly where VAERS is loud.

All day-0–7 hospitalization responses (388).

Myocarditis / pericarditis free text in the coding window (sample of 120 / 136 registrants).

B. Hospitalization checkboxes heavily under-coded #

Paper: “All registrants who indicated that medical care was sought after vaccination were contacted…”

Day-0–7 hospitalization responses: 388 / 289 registrants. Only 77 (~20%) carry MedDRA Hospitalisation; 83 (~21%) have no MedDRA at all. Roughly four in five hospitalization-flagged responses lack the standard hospitalisation preferred term.

Among these 289 hospitalization reporters, call-center / GDIT outcomes are roughly Successful 128 vs Unsuccessful 153—Unsuccessful exceeds Successful.

Same hospitalization grid as above (filter on Hosp PT = N).

C. Emergency visits under-coded #

ER checkbox responses: 1,467 (1,370 registrants). Only 251 (~17%) carry an emergency-care– or hospitalisation-like MedDRA term; 408 (~28%) have no MedDRA row at all.

Sample of under-coded ER responses (300 of 1,216 without ER-like PT).

D. Myocarditis / chest pain / clotting — unpublished next to VAERS #

Paper: “37 reports to VAERS of myocarditis among adults aged ≥18 years…” (Table 2 manufacturer headers sum to 36).

The public SAE narrative is VAERS myocarditis. v-safe Results never publish myocarditis, pericarditis, chest-pain, or thrombosis MedDRA counts for adult boosters. In this cohort’s coding window: 4,809 Chest pain, 41 Myocarditis, 72 Pericarditis, 197 Thrombosis, 109 Pulmonary embolism, 86 DVT.

Free text mentioning myocarditis/pericarditis in the coding window: 136 registrants; 75 receive a myocarditis/pericarditis-class PT. Weak dumps are huge (13,846 Unevaluable event; 26,329 No adverse event).

Why this matters: 41 Myocarditis-coded v-safe registrants (plus 72 Pericarditis and 4,809 Chest pain) go unpublished next to 37 VAERS confirmed cases—and next to a sentence that v-safe does not capture hospitalization diagnoses.

Myocarditis / pericarditis / myopericarditis MedDRA (112 registrants with any of those PTs).

Chest pain MedDRA sample (250 of 4,809).

Thrombosis / PE / DVT MedDRA sample (250).

Serious free text coded Unevaluable (sample).

E. Deaths reported in v-safe — omitted while one VAERS death is discussed #

Paper (Discussion): “One death was reported [to VAERS among myocarditis cases]; investigation is ongoing…”

Under inclusion, MedDRA through 15 March 2022 shows 43 registrants coded Death and 10 Cardiac arrest (53 in the death/arrest union)—including stroke-after-booster, fatal coronary, first-person “I passed away,” and cardiac-arrest free text. The paper never states that any death was reported through v-safe.

Caveats: MedDRA is not a death certificate; many entries are proxy notes; some cite cancer, Parkinson’s, COVID, or “unrelated” causes. Even so, dozens of Death-coded reports sit inside the inclusion window and are omitted while death discussion is limited to one VAERS myocarditis case.

F. Call-center follow-up vs paper wording #

Paper: “All registrants who indicated that medical care was sought after vaccination were contacted and encouraged to complete a VAERS report.”

Among 289 hospitalization registrants, GDIT final outcomes are roughly Successful 128 / Unsuccessful 153 / priority 2/3: 8. A majority are not Successful—softer than “all… were contacted” implies to a lay reader.

Summary #

• Misrepresentation: “does not capture diagnoses associated with hospitalization” — free text/MedDRA include myocarditis, MI, stroke/clot, death
• Major omission: 43 Death + 10 Cardiac arrest (53 union) — never mentioned as v-safe deaths (only 1 VAERS myocarditis death)
• Reporting gap: 41 Myocarditis + 72 Pericarditis + 4,809 Chest pain (+ thrombosis/PE/DVT) unpublished next to 37 VAERS confirmed myocarditis
• Anomaly: VAERS myocarditis text 37 vs Table 2 18+18=36
• Coding failure: ~20% of Hospitalization responses get PT Hospitalisation; ER-like PT on ~17% of ER responses
• Anomaly: GDIT Unsuccessful > Successful among hospitalization reporters
• OK: Table 1 rates and VAERS severity arithmetic check out

Bottom line: Compared to the v-safe dataset, adult booster reactogenicity rates verify—but the MMWR tells readers v-safe does not capture hospitalization diagnoses even when enrollees typed “diagnosed with myocarditis,” leaves thousands of chest-pain and dozens of myocarditis/pericarditis codes unpublished beside a VAERS myocarditis story, under-codes hospitalisations, and omits dozens of Death-coded v-safe registrants while discussing only one VAERS death.