VAERS (U.S., 2006–2024): 74,945 symptom mentions (16.22/100k doses). Largest share: Other / Unclassified (49%), General / Systemic (non-local) (9%), Neurological (8%).
Canada Vigilance (CV Online extract): 1,010 reaction mentions in 283 unique reports (70.7% serious (200 of 283 reports)). Largest share: Neurological (20%), General / Systemic (non-local) (11%), Gastrointestinal (8%).
JADER (PMDA public CSV extract): 7 reaction mentions in 4 unique reports (0.0% serious (0 of 4 reports)). Largest share: Allergic / Anaphylactic (43%), Gastrointestinal (29%), General / Systemic (non-local) (29%).
VigiAccess (WHO): 150,931 reaction-term mentions · search: hepatitis A. Largest share: Other / Unclassified (21%), Injection-site / Local reaction (17%), General / Systemic (non-local) (13%).
Lareb (Netherlands): 1,539 reaction-term mentions · search: Hepatitis A vaccine. Largest share: General / Systemic (non-local) (24%), Injection-site / Local reaction (14%), Neurological (14%).
DAEN (Australia): 3,877 reaction-term mentions · search: hepatitis A. Largest share: Other / Unclassified (16%), General / Systemic (non-local) (15%), Neurological (15%).
Medsafe (New Zealand): 488 reaction-term mentions · search: hepatitis A. Largest share: Gastrointestinal (28%), General / Systemic (non-local) (25%), Injection-site / Local reaction (18%).
Cross-database note: All systems are passive and unverified; reporting rates are not directly comparable across countries (different populations, reporting incentives, and lack of dose denominators for Canada/Japan/EU). top VAERS: Other / Unclassified; top Canada Vigilance: Neurological; top JADER: Allergic / Anaphylactic.
SurVigilance note: VigiAccess, Lareb, DAEN, DMA, and Medsafe counts are live-scraped MedDRA PT mention totals (not deduplicated individual cases). Data via
SurVigilance (GPL-3.0; pip install SurVigilance). Category assignment uses keyword matching on reported reaction terms — approximate and exploratory. Neither database establishes causality.
Pharmacovigilance Lot Signal Detection — Hypothesis-Generating Only
Multi-system context below. VAERS (U.S.) supports lot-level volume z-scores and seriousness flags by product and lot (2006–2024). Each lot links to a summary with report count, seriousness %, adverse-event pie chart, U.S. state map, and timeline. A signal flag means a statistical threshold was exceeded — not that a lot is unsafe. Full dashboard →
VAERS flags:
VOL high report volume (z ≥ 3) ·
BURST clustered in <90 days ·
SER serious reports >50%.
Lot numbers are voluntary/incomplete in VAERS. Location data is U.S. state only (no postal codes in the public extract).
VAERS (United States) — all lots by product
16,588 reports with usable lot across 1,604 lots · 163 flagged
Other Pharmacovigilance Systems
Lot-level analysis is only possible where reporters supply batch/lot numbers in the public extract. Canada Vigilance, JADER (PMDA, Japan), and most other national systems publish product-level spontaneous reports without lot fields.
Canada Vigilance (Health Canada)
283 unique reports · 1,010 reaction mentions · 70.7% serious (200 of 283 reports). Top categories: Neurological (20%), General / Systemic (non-local) (11%), Gastrointestinal (8%).
Canada Vigilance spontaneous reports are unverified temporal associations. The public CV Online data extract does not include lot or batch numbers, so lot-level signal detection is not possible for this system — only product-level reaction patterns are shown here. No Canadian dose denominators are available. Extract 2026-03-31.
Search Canada Vigilance →
JADER (PMDA, Japan)
4 unique reports · 7 reaction mentions · 0.0% serious (0 of 4 reports). Top categories: Allergic / Anaphylactic (43%), Gastrointestinal (29%), General / Systemic (non-local) (29%).
JADER (Japanese Adverse Drug Event Report database) spontaneous reports are unverified temporal associations; PMDA has not assessed causality per case. The public CSV extract does not include lot or batch numbers, so lot-level signal detection is not possible — only product-level reaction patterns are shown here. Reaction terms in source data use MedDRA/J Preferred Terms. JADER CSV extract pmdacasereport202606 (2026-06). JADER reference (PDF)
Search JADER / PMDA adverse reactions →
No EudraVigilance (EU) summary is mapped for this page.
VigiAccess (WHO)
150,931 MedDRA PT mentions · search: hepatitis A. Top categories: Other / Unclassified (21%), Injection-site / Local reaction (17%), General / Systemic (non-local) (13%).
Live-scraped public portal data via SurVigilance (GPL-3.0). Counts are reaction-term mentions, not deduplicated individual cases. No lot/batch field.
Search VigiAccess (WHO) →
Lareb (Netherlands)
1,539 MedDRA PT mentions · search: Hepatitis A vaccine. Top categories: General / Systemic (non-local) (24%), Injection-site / Local reaction (14%), Neurological (14%).
Live-scraped public portal data via SurVigilance (GPL-3.0). Counts are reaction-term mentions, not deduplicated individual cases. No lot/batch field.
Search Lareb (Netherlands) →
DAEN (Australia)
3,877 MedDRA PT mentions · search: hepatitis A. Top categories: Other / Unclassified (16%), General / Systemic (non-local) (15%), Neurological (15%).
Live-scraped public portal data via SurVigilance (GPL-3.0). Counts are reaction-term mentions, not deduplicated individual cases. No lot/batch field.
Search DAEN (Australia) →
All global data sources → · Data schemas →
Active Pharmacovigilance (Defined-Population Surveillance)
Curated findings for Hepatitis A vaccines from active systems (not VAERS). Page inventory last reviewed: 2026-07-10.
ⓘ Active vs. passive — why this pane is separate
The VAERS / multi-system charts above are passive surveillance: spontaneous, unverified reports without a fixed denominator.
Active surveillance starts from a defined, enumerated population (EHR/claims or structured post-vaccination surveys), applies pre-specified statistical tests, and asks whether an outcome occurs
more often than expected in a risk window versus a comparison window or group.
These are not two flavors of the same evidence — active findings are the harder tier that can confirm, refute, or leave under investigation a signal first hinted in passive data.
Do not add VAERS report counts to active incidence rates.
○ No signal detected ◐ Signal under investigation ◑ Investigated — not confirmed ● Signal confirmed (true association) – Not currently under active surveillance
CDC Vaccine Safety Datalink (VSD)
Outcome: Pre-specified serious outcomes in pediatric/adult cohorts
Tier 2
○ No signal detected
Inactivated hepatitis A vaccines are among the least reactogenic routine products. Active post-licensure work has not identified a confirmed rare serious risk that changed routine childhood recommendations.
PopulationVSD HepA vaccinees
Risk intervalStudy-specific
ComparisonControl intervals
Evaluation periodPost-1990s licensure era
MethodPost-licensure observational monitoring
Sources: CDC VSD
Record last reviewed: 2026-07-10
Update cadence: Tier 1: check AusVaxSafety monthly when public pages update. Tier 2/3: quarterly review around ACIP meetings and PubMed/MMWR; set lastReviewed per record.
Source tiers: Tier 1 = public near-real-time dashboards (e.g. AusVaxSafety);
Tier 2 = VSD / Sentinel / PRAC-type findings released via ACIP slides, MMWR, or papers (no public VSD raw dashboard);
Tier 3 = regulator label/safety communications.
Detecting a signal and later classifying it as not confirmed is normal system behavior — not an anomaly to hide or amplify.
Rank-aggregated VAERS signal detection (rankv)
The table below lists vaccine–event pairs that were detected as disproportionality signals by all four base methods used in rankv (GPS, PRR, ROR, BCPNN) on multi-decade VAERS data, then ordered by rank aggregation (Borda average rank; related to the Spearman/GA top-list approach in the rankv paper).
| Agg. rank |
VAERS product |
Preferred term (event) |
N |
Method ranks (GPS / PRR / ROR / BCPNN) |
| 228 |
HEP A (VAQTA) |
Hepatitis A
clinical-coded PT
|
17 |
GPS rank 253 (EBGM=14.89); PRR rank 159 (PRR=34.3359); ROR rank 159 (ROR=34.3532); BCPNN rank 261 (IC_LB=2.6484) |
Showing up to 15 pairs for this page (clinical-coded terms listed first).
Full processed tables: rankv_signals.json.
- Methods combined: BCPNN (IC), GPS/EBGM, PRR, ROR — then rank aggregation.
- Data: ~30 years of public VAERS (rankv processed tables).
- Origin: precisionFDA “Gaining New Insights by Detecting Adverse Event Anomalies” challenge solution.
- Caveat: Disproportionality signals are statistical associations in spontaneous reports. They do not establish causality, incidence, or product defect. Many top pairs reflect administration/product-use coding rather than clinical injury.
Source: nanx.me/rankv
· Code: github.com/nanxstats/rankv (MIT)
· Built: 2026-07-30.