FDA Adverse Events

FDA adverse-event records from FAERS (drugs) and MAUDE (devices) — for pharmacovigilance signal detection.

Public regulatory/health data — not medical advice.

📊 Sample Output

report_id	report_type	received_date	seriousness_flags	drug_names	manufacturers
26931376	drug_event	2026-06-28	['hospitalization']	['TRULICITY', 'OZEMPIC', 'WEGOVY']	['Eli Lilly and Company', 'Novo Nordisk Pharmaceutical Industries, LP…
26931835	drug_event	2026-06-28	['hospitalization', 'other']	['WINREVAIR', 'WINREVAIR', 'WINREVAIR', 'WINREVAIR', 'TREPROSTINIL',…	['Merck Sharp & Dohme LLC', 'Sandoz Inc', 'Alembic Pharmaceuticals In…
26932111	drug_event	2026-06-28	['other']	['OZEMPIC']	['Novo Nordisk Pharmaceutical Industries, LP']
26932113	drug_event	2026-06-28	['other']	['OZEMPIC']	['Novo Nordisk Pharmaceutical Industries, LP']
26932116	drug_event	2026-06-28	['hospitalization', 'other']	['OZEMPIC', 'OZEMPIC', 'OZEMPIC']	['Novo Nordisk Pharmaceutical Industries, LP']
26932121	drug_event	2026-06-28	['hospitalization', 'other']	['OZEMPIC', 'OZEMPIC', 'OZEMPIC']	['Novo Nordisk Pharmaceutical Industries, LP']

Real rows from run d1nnUk8KfqGaIW209 on build 0.0.8 (2026-09-25), unedited apart from masked emails/phones and shortened long text; fields the source does not publish are empty.

⚙️ Sample inputs

This is the Store example input — the same input Apify's automated check runs — so it is proven to return rows on the current build. Paste it into the input form, or send it through the API, CLI or a schedule:

{
  "event_type": "drug",
  "search_term": "ozempic",
  "serious_only": true,
  "days_back": 90,
  "max_events": 100
}

🔧 Input reference

Field	Type	Default	What it does
event_type	string (drug, device, both)	"drug"	Which adverse-event database(s) to query. 'drug' = FAERS (drug events); 'device' = MAUDE (medical device events); 'both' = combined.
search_term	string	"ozempic"	Free-text Lucene query against drug/device names — matches medicinalproduct, brand_name, generic_name (FAERS) and device.brand_name / device.generic_name (MAUDE). Examples: 'ozemp…
serious_only	boolean	true	If true, filter to only serious reports — FAERS serious=1 and MAUDE event_type Injury / Death. Recommended for signal detection.
days_back	integer	90	Look-back window from today in days. FAERS uses receivedate; MAUDE uses date_received. Default 90 days.
max_events	integer	100	Hard cap on total events returned across both sources. When event_type='both', budget splits 50/50.

🧾 JSON sample record

One real record from run d1nnUk8KfqGaIW209 (emails/phones masked, long text shortened):

{
  "source": "FAERS",
  "report_id": "26931376",
  "report_type": "drug_event",
  "received_date": "2026-06-28",
  "occur_country": "EU",
  "reporter_country": "EU",
  "serious": "Y",
  "seriousness_flags": [
    "hospitalization"
  ],
  "patient_age": null,
  "patient_age_unit": null,
  "patient_sex": "male",
  "drug_names": [
    "TRULICITY",
    "OZEMPIC",
    "WEGOVY"
  ],
  "device_names": [],
  "manufacturers": [
    "Eli Lilly and Company",
    "Novo Nordisk Pharmaceutical Industries, LP"
  ],
  "meddra_pt_terms": [
    "Pancreatic injury",
    "Lipase increased",
    "General physical health deterioration",
    "Drug ineffective"
  ],
  "event_type": null,
  "narrative_summary": "Reactions: Pancreatic injury; Lipase increased; General physical health deterioration; Drug ineffective. Drugs: TRULICITY; OZEMPIC; WEGOVY. Seriousness: hospitalization",
  "source_url": "https://api.fda.gov/drug/event.json?search=safetyreportid:26931376",
  "sender_organization": "ELI LILLY AND CO",
  "scraped_at": "2026-09-25T00:05:28Z",
  "prr_signal": {
    "drug_or_device": "TRULICITY",
    "meddra_pt": "Pancreatic injury",
    "prr": 99,
    "chi_square": 24.01,
    "co_occurrence_count": 1,
    "total_in_set": 100,
    "interpretation": "insufficient_cases"
  }
}

What the PRR score is — and is not

prr_signal is a Proportional Reporting Ratio computed over the reports this run retrieved (your drug/device/date filters and the row cap). It is a screening cue on that sample: basis: "retrieved_sample" and denominator_note say so on every record. It is not a population-level disproportionality analysis over the whole FAERS/MAUDE database, duplicates/follow-up versions of a case are not collapsed, and a reported association is never evidence of causation. Lead with the source-linked event records; treat the score as a pointer for where to look next.

💰 Pricing

Event	Price (USD)	When it is charged
Actor Start (apify-actor-start)	$0.005	Charged when the Actor starts running. Number of events charged depends on Actor memory (one event per GB, minimum one event).
Adverse event (adverse-event)	$0.05	Per FAERS drug event or MAUDE device event extracted + signal-scored

Pay-per-event: you pay only for what the run delivers. A run that delivers nothing bills no result events (only the actor-start event, when the actor defines one). Example: a run that delivers 100 results costs 100 × $0.05 = $5.00 plus the start fee.

🔗 Related Actors

More from the NexGenData Healthcare & FDA family:

• CMS Open Payments — Pharma-to-Doctor Records API
• Consumer Product Recall & Safety Intelligence — CPSC
• Doctor Directory Lead Finder — Provider Contacts API
• FDA Drug Approvals Tracker — NDA/BLA/ANDA API
• FDA Drug Shortages Monitor — Current & Resolved API
• FDA Orange Book — Drug Patent & Exclusivity API

15 more in this family on the NexGenData Store page.

📊 What you get

Clean JSON, one record per report (full output has 21 fields):

• source — Source
• report_id — Report id
• report_type — Report type
• received_date — Received date
• occur_country — Occur country
• reporter_country — Reporter country
• serious — Serious
• seriousness_flags — Seriousness flags
• patient_age — Patient age
• patient_age_unit — Patient age unit

Each record also carries a prr_signal object (Proportional Reporting Ratio + Yates-corrected chi-square) computed across the records returned by your run.

Pricing: $0.05 per report (Pay-Per-Event) — about 20 reports per $1. Actor start is charged separately by the platform at $0.005 per GB of run memory.

Rows are charged only when they are delivered: if openFDA returns no matches, the run finishes successfully with 0 rows and nothing is charged; if openFDA cannot be reached, the run fails loudly rather than returning an empty dataset.

🤖 Use with AI agents

Point Claude, the OpenAI Agents SDK, an n8n flow or any MCP-aware client at it and pull official records on demand.

Agentic payments (x402): Supports agentic payment via x402 — agents can call this actor with USDC, no API key required.

🔗 Related regulatory actors

FDA & public-health records family:

OpenFDA Drug · Drug Approvals · Drug Shortages · Purple Book · Warning Letters/483 · Pharma Pipeline · Medical Devices

Product-safety recalls: NHTSA Recalls · NHTSA Defects · NHTSA Safety · CPSC Recalls

Agent front door — Healthcare & FDA MCP: Healthcare & FDA MCP

---

Public official records and enforcement data, assembled for research and monitoring.