We studied where VAERS reporting breaks — then built the flow that fixes it
VAERS is a passive surveillance system: its value depends on the reports people manage to finish and how cleanly those reports are captured. The current experience works against both. Here's the evidence, and the redesigned flow it drove — the same one you just walked through.
20 min
Inactivity timeout on the current form — no autosave, no resume
Current vaers.hhs.gov
<1%
Share of vaccine adverse events an AHRQ pilot estimated are ever reported
Lazarus / AHRQ, 2011 ¹
13
Distinct data-quality problems documented in VAERS free-text processing
Du et al., 2023 ²
−30%
Abandonment reduction CDC asks the modernized app to achieve
CDC's stated goal
What's broken today
Each problem hurts both whether the report gets finished and whether its data is usable — and each maps to a concrete redesign move.
The 20-minute timeout wipes work in progress
The form times out after 20 minutes of inactivity and discards what was entered — punishing exactly the reporters who need time to find a lot number or pull a chart. No autosave, no resume.
Fix →Autosave and resumable sessions.autosave & resume
One dense form serves two very different users
Public reporters and clinicians face the same clinical fieldset — laypeople meet untranslated terminology; providers wade through fields that don't apply. Both suffer, and the data arrives mixed.
Fix →Fork on role; two right-sized paths over one shared data model.role-aware paths
Free-text capture degrades the data at the source
Reports arrive incomplete, with typos, vague text, generalized terms, and missing time stamps — and downstream coding inherits every ambiguity. CDC itself notes that because all reports, even incomplete ones, are accepted and anyone may submit, "the accuracy and amount of information varies between reports." A peer-reviewed study catalogued 13 such problems and recommended, in its words, "a structured VAERS interface using pre-defined terminology selections" — precisely this redesign.
Five fixed sections show every field regardless of relevance. A provider filing a vaccine error with no adverse event still confronts the entire reaction apparatus — one of the most common situations the current form handles badly.
Fix →Branching that asks only what applies; the no-AE path skips every reaction field.ask only what applies
Not built for the device people actually use
A desktop table-form adapted downward — small targets, wide layouts, long scrolls — despite much of the public arriving on a phone.
Fix →Mobile-first responsive layout on every screen.mobile-first
Fragmented entry, off-page help, bolt-on Spanish
Online form, writable PDF, provider portal, and eFax compete before a single field is entered; guidance lives on a separate page; and "en Español" swaps to a parallel page rather than a first-class Spanish experience carried through the flow.
Fix →One guided path, in-context help and FAQ, full English/Spanish parity.one guided path · EN/ES
A recognized, ongoing effort
This isn't a new concern. CDC and FDA modernized the VAERS reporting form once before — the VAERS 2.0 form, rolled out in 2017 — and CDC openly documents the system's limits: underreporting, and report completeness that varies with the severity of the event and how the report is made. The redesign here continues that recognized effort, and focuses it on the front door where reporters actually are.
The flow, before and after
The redesign turns a fragmented, timeout-prone path into one guided decision tree. Every node is tagged with the requirement it satisfies.
Current-state submission path — hazards marked in red.
Reading it: four entry paths converge on one form, gated by a timeout that can erase everything. The form is identical for a first-time parent and a clinician, and its help lives elsewhere.
Redesigned reporting decision tree — decisions in blue, the no-reaction skip in green.
The one edge that matters: the "no — vaccine error only" branch removes the entire reaction section — the single change that matters most for both reporter burden and data quality. Both roles run off one shared data model, so the fork changes wording and which fields show, never the underlying VAERS record.
Potential improvements to explore
Directions we're exploring for future phases — beyond the committed base scope. Each is chosen because it makes reporting easier and the captured data cleaner at the same time. ★ marks the three signature bets.
These are exploratory concepts, shown to signal where the platform can go. They are not claims of built functionality in the base offer.
Capture coded data at the source
Voice & tap → confirmed chips ★ signature
The reporter describes the reaction out loud; we surface what we heard as tappable chips they confirm or correct — each one a coded value, not free text.
EasierBetter data
Plain-language symptom tiles
The public picks "red and swollen where I got the shot"; the system stores the clinical term behind the scenes.
EasierBetter data
Body-map tap for site
Tap where it happened on a simple figure to capture precise localization most free-text reports never include.
EasierBetter data
Fix the two data black holes — lot numbers and timing
Snap the vaccination card ★ signature
Photograph the card; OCR pulls lot, manufacturer and date and validates the lot — the highest-value field for batch-level signal, and the one people most often guess at or skip.
EasierBetter data
Visual event timeline ★ signature
Place events on a small timeline — shot given, symptom onset, care sought, resolved — capturing the temporal metadata research says is usually missing.
EasierBetter data
Cut the effort, so more reports finish
Minimum-viable report first
Capture the core signal in about three questions, then invite enrichment — so even an unfinished report has banked the essential safety information.
EasierBetter data
Magic resume link
A secure link to continue on another device or after finding records — start in the pharmacy, finish at home, come back days later.
Easier
Cleaner data — and more trust
Duplicate & follow-up detection
Gently flag when a report for this vaccination may already exist, or offer to link a follow-up — attacking a known dataset problem.
Better data
Kind cross-field checks
Catch a reaction dated before the shot, or dose 3 with no prior doses, at entry — fixed in one tap, not in coding.
Better data
Voice-only / read-aloud path
The whole report navigable by voice for low-vision, low-literacy or motor-impaired reporters — widening who can report at all.
EasierBetter data
"What happens next" + add records later
Clear post-submit guidance and a way back in — answering the documented complaint that reporters feel unheard, and drawing in fuller follow-up.
EasierBetter data
A self-studying form
Instrument where reporters struggle and drop off, feeding CDC a live view so the form keeps improving through the configuration layer.
Better data
Sources & notes
1 · Lazarus R, et al. Electronic Support for Public Health–VAERS (ESP:VAERS), AHRQ grant final report, 2011 — the widely cited pilot estimating fewer than 1% of vaccine adverse events are reported; used here to frame reporting burden, not to quantify vaccine risk. digital.ahrq.gov
2 · Du J, et al. Data-quality challenges in free-text VAERS processing; recommends a structured interface with pre-defined terminology. BMC Medical Informatics & Decision Making, 2023. link.springer.com
3 · CDC. Surveillance Manual, Chapter 21 (VAERS) — CDC's own account of the system's limits, including underreporting and that "the accuracy and amount of information varies between reports." cdc.gov
4 · Implementation of the VAERS 2.0 Reporting Form. MMWR, 2017 — the last modernization of the reporting form. cdc.gov
5 · Current-form behavior (timeout, fragmented entry, single linear form, off-page help, bilingual bolt-on) observed on the live vaers.hhs.gov flow. This is an illustrative design study on synthetic data — no real reports or protected data are represented.
❤
Keep your report?
Reports are easy to leave unfinished — especially when you're not feeling your best, or filling this out for someone who isn't. Your progress is saved; you can finish now or come back to it.
You're not alone in reporting: people have submitted more than 2 million VAERS reports since 1990, most describing mild reactions. Each one — including yours — helps the safety picture.
We'll never suggest what to report — what happened is yours to describe.
✓
Your progress is saved
You can pick up where you left off. In the delivered app, a secure link would let you continue on any device — no report lost to a timeout.
Vaccine Adverse Event Reporting System
Report a reaction after a vaccination
A faster, guided way to file a VAERS report — on any device, in English or Spanish. We ask only what applies to you, save your progress as you go, and help at every step.
⚠️
Medical emergency?If you are experiencing a medical emergency, call 9-1-1 or seek care immediately. VAERS is for reporting, not medical advice.
💾
Never lose your work
Progress saves automatically — no 20-minute timeout wiping your report.
🔀
Only what applies
Branching questions skip fields that don't fit your situation.
💬
Help in context
Plain-language tips and assistance right where you need them.
♿
Built for everyone
Mobile-first, Section 508 accessible, English and Spanish.
Who is filing this report?
We tailor the questions and wording to you. You can change this later.
Saved just now
✓Patient
2Vaccine
3Reaction
4Review
The vaccination
Not sure of the exact name? Pick the closest — you can add the brand next. Your vaccination card lists it.
We list manufacturers for the vaccine you chose. "Don't know" is fine — it won't block your report.
Found on your vaccination card. Format is usually letters + numbers.
The lot number links your report to a specific batch — it makes vaccine-safety data far more useful. On your card it's near the vaccine name.
We check this against the reaction date for you.
Provider field — hidden on the public path to reduce burden.
Was there an adverse event (a reaction) after this vaccine?
Sometimes a vaccine is given in error with no reaction at all. If so, we skip every reaction question — you won't fill in what doesn't apply.
Talk it through in your own words — say what happened and when. You can edit the text after.
From your wordsWe captured these candidate details — review them before they enter your report:
✓
Reaction questions skipped.Because there was no reaction, we won't ask you to complete adverse-event fields. Branching logic — PWS Task 1.6.2.
✦
Need a hand?
Ask in plain words — "where's my lot number?" or "what counts as a reaction?" — and get an answer without leaving this page. This is a design mockup; in the delivered application, any assistant would run inside CDC's approved secured environment.
Saved just now
✓Patient
✓Vaccine
✓Reaction
4Documents
Documents & records
Attach medical or vaccine records related to the reaction. Optional — you can also add documents later.
⤒
Drag files here, or browse
PDF, JPG, or PNG · up to 25 MB each · Phase 1: medical & vaccine records
Design mockup — selecting adds a simulated example file.
Smart helperBased on this report, these documents would strengthen it:
✦
Only what's needed
Phase 1 accepts medical and vaccine records. The tool checks what you've told us and suggests documents that fill the gaps — nothing more.
✓
Report saved
This is a design mockup — no real report was submitted. In the live app, you'd get a confirmation number and clear next steps.
Reporter type
Time to complete
4 min 12 sec
Fields skipped by branching
0
Autosaves
7
About VAERS reporting
A quick, plain-language guide to what VAERS is, what to expect when you report, and the questions people ask most.
What VAERS is
VAERS is a national early-warning system, co-managed by CDC and FDA, that detects possible safety problems in U.S.-licensed vaccines. It's a passive system: it relies on people reporting health problems that happen after vaccination so scientists can watch for unusual or unexpected patterns.
What VAERS is not
A report is not proof a vaccine caused a problem — VAERS is not designed to determine cause — and it is not a source of medical advice. Each report is a signal to be reviewed, not a conclusion. In a medical emergency, call 9-1-1.
What to expect when you report
Common questions
Get VAERS data
Search de-identified, public vaccine-safety reports from CDC and FDA. Personal identifiers are removed to protect privacy. The anchor figures below are real; the filtered breakdown and sample rows are illustrative for this mockup.
2M+reports since 1990 (CDC / FDA)
4–6 wksuntil de-identified data is published
Publicfree to search & download
—matching reports (illustrative)
Most-reported reactions (illustrative)
Most VAERS reports describe mild reactions such as fever (FDA). The percentages here are illustrative.
Sample of matching reports (de-identified)
The anchor figures above are from public CDC and FDA VAERS data. The filtered counts, breakdown percentages and sample rows are illustrative and synthetic for this mockup — not real records. The delivered application searches CDC's public, de-identified vaccine-safety data.