openFDA Scraper — FDA Recalls, 510(k), NDC & Adverse Events

openFDA Scraper — FDA Recalls, 510(k) Clearances, Drug NDC & Adverse Events

Extract official, structured regulatory data from the US Food and Drug Administration's openFDA database. Retrieve clean, standardized records across drug, medical device, and food recalls (Enforcement Reports), 510(k) medical device premarket clearances, the National Drug Code (NDC) Directory, and FAERS post-marketing drug adverse event reports.

Execute full openFDA Lucene queries with Boolean filters, field scoping, and custom sorting. Every record is normalized into tidy columns with unified summary fields, unpacked nested clinical arrays, and standardized ISO dates.

No API keys needed, no browser overhead, and zero complex infrastructure. Built for life sciences regulatory analysts, medical device competitive intelligence teams, pharmacovigilance specialists, clinical researchers, mass-tort / product-liability legal teams, and pharmaceutical supply chain monitors.

---

⚡ At a glance

Feature	Specification
Data Source	Official US FDA Public API (api.fda.gov)
Authentication	Keyless & public (optional free API key supported to increase rate limits)
Datasets Supported	Drug Recalls, Device Recalls, Food Recalls, 510(k) Clearances, NDC Directory, FAERS Adverse Events
Search Syntax	Full Apache Lucene search queries (field:value, AND, OR, NOT, wildcards)
Batch Efficiency	Up to 1,000 records per individual request (ultra-fast 2–5s execution)
Output Formats	Dataset (JSON, CSV, XML, Excel), Live API, Webhooks, Google Sheets
Scale & Pagination	Automated skip pagination supporting up to 25,000 records per query
Date Normalization	Automatic conversion from FDA YYYYMMDD strings to ISO YYYY-MM-DD
Proxy Requirement	Apify Datacenter Proxy with automated session rotation

---

⚖️ Comparison: openFDA Scraper vs Alternatives

Criteria	openFDA Scraper (This Actor)	Raw openFDA API	FDA FOIA Requests
Setup & Coding	🟢 Zero code, web UI + direct API	🔴 Complex custom REST integration	🔴 Manual paperwork & legal filings
Data Normalization	🟢 Tidy flat columns + unified summary	🔴 Deeply nested JSON / arrays	🔴 Scanned PDFs and paper copies
Date Formatting	🟢 ISO standard (YYYY-MM-DD)	🔴 Raw integer strings (YYYYMMDD)	🟡 Varies by document
Pagination Handling	🟢 Auto-paging up to 25,000 records	🔴 Manual skip/limit looping	🔴 Not applicable
Export Formats	🟢 Instant CSV, Excel, JSON, XML	🔴 JSON only	🔴 Static PDFs / paper records
Turnaround Time	🟢 Instant (seconds)	🟢 Instant (seconds)	🔴 Weeks to months

---

🎯 Key Use Cases

1. Drug, Device & Food Recall Surveillance

Set up automated compliance pipelines to monitor urgent Class I (high risk of serious injury or death), Class II, and Class III recalls. Track specific pharmaceutical manufacturers (e.g. recalling_firm:Pfizer), therapeutic classes, or affected states to safeguard hospital supply chains and retail inventory.

2. Medical Device Competitive Intelligence & Clearance Tracking

Track competitor 510(k) premarket clearance submissions (e.g. applicant:Medtronic, applicant:Stryker) and clearance dates (decision_date:desc). Monitor medical device market entry timelines, product codes, predicate device classifications, and advisory committee reviews.

3. Pharmacovigilance & Adverse Event Analysis

Analyze real-world post-marketing safety data from the FDA Adverse Event Reporting System (FAERS). Query reported patient adverse reactions (MedDRA preferred terms), concomitant suspect medications, patient demographics, and seriousness indicators (hospitalization, life-threatening, death).

4. Pharmaceutical Directory & National Drug Code (NDC) Enrichment

Build or enrich master pharmaceutical catalogs using authoritative NDC directory records. Extract proprietary brand names, non-proprietary generic names, active ingredient strengths, dosage forms, administration routes, and labeler/distributor corporate details.

5. Product Liability & Legal Litigation Research

Construct authoritative evidence bases for medical device defect investigations, pharmaceutical mass-tort cases, and corporate regulatory compliance audits.

---

📊 Extracted Data Fields by Dataset

Every record includes standardized top-level fields: recordType, title, date, summary, and scrapedAt. Dataset-specific fields are detailed below:

🚨 1. Recall Enforcement Reports (drug-enforcement, device-enforcement, food-enforcement)

Field	Type	Description	Coverage
recallNumber	String	Official FDA recall tracking number (e.g. D-0084-2027)	🟢 High (100%)
classification	String	Recall hazard level (Class I, Class II, Class III)	🟢 High (100%)
status	String	Current status (Ongoing, Completed, Terminated)	🟢 High (100%)
recallingFirm	String	Full corporate name of the firm initiating the recall	🟢 High (100%)
productType	String	Regulated product category (Drugs, Devices, Food)	🟢 High (100%)
productDescription	String	Detailed description of the recalled product and packaging	🟢 High (100%)
reasonForRecall	String	Official medical/regulatory reason for the recall	🟢 High (100%)
distributionPattern	String	Geographic distribution territory (e.g. US Nationwide, CA, NY)	🟢 High (95%)
voluntaryMandated	String	Whether voluntary or FDA-mandated	🟢 High (95%)
city / state / country	String	Corporate headquarters location of the recalling firm	🟢 High (95%)
recallInitiationDate	String	Date the firm first initiated the recall (YYYY-MM-DD)	🟢 High (95%)
reportDate	String	Date the recall was officially published by the FDA (YYYY-MM-DD)	🟢 High (100%)
brandNames	Array	Brand name aliases extracted from the record	🟡 Conditional (75%)

🩺 2. Medical Device 510(k) Clearances (device-510k)

Field	Type	Description	Coverage
kNumber	String	Premarket 510(k) clearance number (e.g. K231450)	🟢 High (100%)
deviceName	String	Trade name of the medical device	🟢 High (100%)
applicant	String	Device sponsor / manufacturing company name	🟢 High (100%)
contact	String	Primary corporate contact person	🟢 High (90%)
productCode	String	Three-letter FDA product classification code	🟢 High (98%)
decisionCode	String	FDA review outcome (e.g. SESE - Substantially Equivalent)	🟢 High (100%)
decisionDate	String	Date the clearance decision was issued (YYYY-MM-DD)	🟢 High (100%)
dateReceived	String	Date the FDA originally received the 510(k) application	🟢 High (98%)
clearanceType	String	Application type (Traditional, Special, Abbreviated)	🟢 High (95%)
advisoryCommittee	String	Review panel (e.g. Cardiovascular, Orthopedic, General Hospital)	🟢 High (95%)
deviceClass	String	Regulatory risk classification (1, 2, 3)	🟢 High (95%)

💊 3. National Drug Code (NDC) Directory (drug-ndc)

Field	Type	Description	Coverage
productNdc	String	Standard National Drug Code identifier (e.g. 50580-449)	🟢 High (100%)
brandName	String	Proprietary commercial brand name (e.g. TYLENOL)	🟢 High (95%)
genericName	String	Active pharmaceutical ingredient generic name	🟢 High (95%)
labelerName	String	Name of the pharmaceutical distributor or manufacturer	🟢 High (100%)
dosageForm	String	Physical form (e.g. TABLET, CAPSULE, INJECTION)	🟢 High (98%)
route	Array	Administration route (e.g. ORAL, INTRAVENOUS)	🟢 High (95%)
marketingCategory	String	Regulatory pathway (NDA, ANDA, OTC MONOGRAPH FINAL)	🟢 High (95%)
activeIngredients	Array	Active substance ingredients with strength measurements	🟢 High (95%)
packaging	Array	Commercial packaging units and NDC package codes	🟢 High (95%)

⚠️ 4. FAERS Adverse Event Reports (drug-event)

Field	Type	Description	Coverage
safetyReportId	String	Unique FDA safety report identifier	🟢 High (100%)
serious	String	Seriousness code (1 = serious event, 2 = other)	🟢 High (100%)
seriousnessDeath	String	Indication if the adverse event resulted in patient death	🟡 Conditional (when reported)
receiveDate	String	Date report was logged into the FDA database (YYYY-MM-DD)	🟢 High (100%)
patientSex	String	Patient biological sex (1 = Male, 2 = Female)	🟢 High (90%)
patientOnsetAge	String	Numerical age at adverse event onset	🟡 Conditional (75%)
reactions	Array	MedDRA preferred terms for adverse reactions	🟢 High (100%)
drugs	Array	Suspect and concomitant medications administered	🟢 High (100%)
indications	Array	Medical indications for which suspect drugs were taken	🟢 High (85%)
primarySourceCountry	String	ISO country code of the reporting healthcare provider	🟢 High (95%)

---

⚙️ Input Parameters

Parameter	Type	Default	Description
dataset	String	"drug-enforcement"	Target endpoint: "drug-enforcement", "device-enforcement", "food-enforcement", "device-510k", "drug-ndc", or "drug-event".
search	String	""	openFDA Apache Lucene query string. Supports field qualifiers and Boolean operators.
sort	String	""	Sort expression (field:desc or field:asc), e.g. report_date:desc, decision_date:desc.
maxItems	Integer	0 (prefill: 100)	Maximum records to extract (up to 25,000 per query). Set 0 for default 100.
apiKey	String	""	Optional free openFDA API key to increase rate limits. Not required for standard runs.
proxyConfiguration	Object	{ "useApifyProxy": true }	Apify proxy settings. Standard datacenter proxy handles all endpoints cleanly.

---

🔍 Proven Lucene Search Query Recipes

Use these high-impact query templates in the search input field:

Objective	Dataset	Lucene Query Expression
Urgent Class I Drug Recalls	drug-enforcement	classification:"Class I" AND status:Ongoing
Recalls by Specific Pharma Firm	drug-enforcement	recalling_firm:Pfizer
Medical Device Clearances by Company	device-510k	applicant:Medtronic
Cardiovascular Device Clearances	device-510k	advisory_committee:CV
Drug Product Search by Brand Name	drug-ndc	brand_name:Tylenol
Prescription Only Injectable Drugs	drug-ndc	dosage_form:INJECTION AND marketing_category:NDA
Adverse Events for Specific Medication	drug-event	patient.drug.medicinalproduct:aspirin AND serious:1
Pediatric Adverse Drug Reactions	drug-event	patient.patientonsetageunit:801 AND patient.patientonsetage:[1 TO 12]

---

📦 Output Sample

Here is a representative JSON record output from the drug-enforcement dataset:

{
  "recordType": "drug-enforcement",
  "title": "Greenwich Rx",
  "date": "2026-10-28",
  "summary": "Microbial Contamination of Sterile Products",
  "recallNumber": "D-0084-2027",
  "status": "Ongoing",
  "classification": "Class I",
  "recallingFirm": "Greenwich Rx",
  "productType": "Drugs",
  "productDescription": "Glutathione Injection, 200 mg/mL, 10 mL multi-dose vial, Rx only",
  "reasonForRecall": "Microbial Contamination of Sterile Products identified during sterility audit",
  "distributionPattern": "US Nationwide",
  "voluntaryMandated": "Voluntary: Firm initiated",
  "city": "Tomball",
  "state": "TX",
  "country": "United States",
  "recallInitiationDate": "2026-10-01",
  "reportDate": "2026-10-28",
  "brandNames": [
    "Glutathione"
  ],
  "eventId": "94812",
  "scrapedAt": "2026-11-10T14:40:00.000Z"
}

---

⚠️ Understanding FDA Recall Hazard Classifications

The FDA classifies recalls into three numeric classes based on the relative degree of health hazard:

Hazard Classification	FDA Statutory Definition	Example Scenarios	Action Urgency
Class I	Dangerous or defective products that could cause serious adverse health consequences or death.	Non-sterile injectable drugs, endotoxin contamination, undeclared lethal allergens, defective life-supporting defibrillators.	🚨 Urgent public warnings & immediate product removal
Class II	Products that might cause temporary or medically reversible adverse health consequences, or where the probability of serious injury is remote.	Subpotent prescription drugs, mislabeled dosage instructions, particulate matter in oral solutions.	🟡 Standard commercial recall & patient notification
Class III	Products that violate FDA labeling or manufacturing regulations but are unlikely to cause adverse health consequences.	Minor container defects, minor labeling omissions, technical compliance issues without therapeutic compromise.	🟢 Administrative product return or relabeling

---

🏛️ Medical Device 510(k) Review Panels Reference

When filtering 510(k) medical device premarket clearances using advisory_committee: in your Lucene queries, use these standard two-letter panel codes:

Panel Code	Advisory Committee Review Panel	Common Cleared Device Types
AN	Anesthesiology	Ventilators, gas delivery systems, monitors
CV	Cardiovascular	Vascular stents, heart valve catheters, ECG devices
DE	Dental	Dental implants, orthodontic aligners, bone graft matrix
GU	Gastroenterology & Urology	Endoscopes, hemodialysis machines, urinary catheters
HO	General Hospital & Personal Use	Infusion pumps, hypodermic syringes, hospital beds
NE	Neurological	Neurostimulators, cranioplasty plates, EEG monitors
OP	Ophthalmic	Intraocular lenses, diagnostic lasers, contact lens solutions
OR	Orthopedic	Spinal fusion cages, hip/knee prosthetic implants
RA	Radiology	CT scanners, MRI systems, medical imaging PACS software

---

💻 API & Integration Examples

Integrate live FDA regulatory data directly into your Python analytical pipelines, Node.js services, or automation webhooks.

Python Example (Pharma Compliance Monitor)

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run_input = {
    "dataset": "drug-enforcement",
    "search": "classification:\"Class I\" AND status:Ongoing",
    "sort": "report_date:desc",
    "maxItems": 100
}

run = client.actor("haketa/openfda-scraper").call(run_input=run_input)

for recall in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(f"[{recall['classification']}] {recall['recallingFirm']} - {recall['title']}: {recall['reasonForRecall'][:80]}...")

Node.js Example (Device Clearance Tracker)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });

const input = {
    dataset: 'device-510k',
    search: 'decision_date:[20260101 TO 20261231]',
    sort: 'decision_date:desc',
    maxItems: 50
};

const run = await client.actor('haketa/openfda-scraper').call(input);
const { items } = await client.dataset(run.defaultDatasetId).listItems();

console.log(`Retrieved ${items.length} 510(k) clearances.`);
console.table(items.map(d => ({ kNumber: d.kNumber, device: d.deviceName, applicant: d.applicant, date: d.decisionDate })));

cURL Example

curl --request POST \
  --url "https://api.apify.com/v2/acts/haketa~openfda-scraper/runs?token=YOUR_APIFY_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
    "dataset": "drug-ndc",
    "search": "brand_name:Amoxicillin",
    "maxItems": 50
  }'

Webhook Alert Example for Compliance Teams

Set up an Apify webhook to instantly alert your regulatory affairs team on Slack, Microsoft Teams, or PagerDuty whenever a Class I recall is detected:

{
  "eventType": "ACTOR.RUN.SUCCEEDED",
  "actorId": "haketa/openfda-scraper",
  "datasetId": "YOUR_DATASET_ID",
  "recordsCount": 100,
  "dataset": "drug-enforcement",
  "alert": "Class I Recall Published",
  "recallingFirm": "Greenwich Rx",
  "hazardSummary": "Microbial Contamination of Sterile Products"
}

---

🚀 Tips for High-Volume Extraction

• Instant 1,000-Record Pulls: The actor requests 1,000 items per batch. Standard pulls under 1,000 items complete in a single lightning-fast HTTP call.
• Sorting for Freshness: Always include sort: "report_date:desc" (recalls) or sort: "decision_date:desc" (device clearances) to ensure your dataset prioritizes newly published actions.
• Targeted Lucene Filtering: Use exact phrase quotes (e.g. classification:"Class I") to avoid false positive term matching.
• Handling openFDA Limits: openFDA limits standard skip-based pagination to 25,000 records per search expression. For corpora larger than 25,000 records, partition your queries by year or geographic state.

---

❓ Frequently Asked Questions (FAQ)

Do I need to obtain an official openFDA API key?

No. The openFDA API is publicly accessible without an API key. For high-frequency automated polling, you can optionally provide a free API key from open.fda.gov in the apiKey field to lift rate limits.

How frequently are FDA datasets refreshed?

The FDA typically updates recall enforcement reports weekly on Wednesdays, medical device 510(k) clearances monthly, and FAERS adverse events on a quarterly basis.

Can I export results to Excel, CSV, or Google Sheets?

Yes. Every dataset can be downloaded immediately as Excel (XLSX), CSV, JSON, or XML from the Apify Console or integrated automatically using live dataset API endpoints.

Why are some brandNames empty in recall records?

The FDA's underlying recall records originate from enforcement field reports. When an official product matches the NDC directory, openFDA attaches an openfda annotation block containing brand names. When not matched, this field remains null, which is expected behavior.

How do I load FDA recall data directly into pandas?

Use the Apify Python client or load the dataset CSV endpoint directly:

import pandas as pd
df = pd.read_csv("https://api.apify.com/v2/datasets/YOUR_DATASET_ID/items?format=csv")
print(df[["classification", "recallingFirm", "reportDate"]].head())

Can I schedule weekly recall checks?

Yes. Create a scheduled task in the Apify Console with the cron expression 0 14 * * 3 (every Wednesday at 2:00 PM UTC) to align with the FDA's standard weekly enforcement report publishing cycle.

Does the scraper support wildcards in Lucene queries?

Yes. You can use standard Lucene single-character (?) and multiple-character (*) wildcards (e.g., brand_name:Amox* or recalling_firm:Pharma*).

📋 Regulatory Compliance & Legal Use Notes

• Authoritative Primary Source: Data is directly queried from api.fda.gov, reflecting official FDA database records without intermediary alteration.

• Audit Trail Compliance: Every extracted record contains an immutable ISO 8601 UTC timestamp (scrapedAt), supporting regulatory audit trail requirements for life sciences quality management systems (QMS).

• Public Domain Information: FDA recall enforcement reports, premarket 510(k) summary decisions, and NDC directory filings reside in the public domain under US government open data statutes.

• Data Freshness Guarantee: Live queries dynamically retrieve the most up-to-date entries in the FDA repositories at execution time.

• Deduplication Safeguards: Output records feature natural composite keys (such as recallNumber, kNumber, productNdc, or safetyReportId) ensuring seamless database ingestion.

• Automated Field Normalization: Nested clinical objects are unpacked into tidy array strings, eliminating parsing overhead for data science pipelines.

• Enterprise Cloud Scalability: Runs on resilient Apify cloud workers capable of streaming large historical cohorts into persistent datasets without timeouts.

---

🔗 Related Healthcare & Registry Scrapers

• ClinicalTrials.gov Scraper — Global clinical study protocols and trial results.
• NPPES NPI Registry Scraper — US National Provider Identifier doctor and clinic registry.
• Healthgrades Doctor Scraper — Physician ratings, patient reviews, and hospital affiliations.
• SEC EDGAR Scraper — Public biotech and pharmaceutical 10-K / 10-Q filing extractor.
• European Medicines Agency (EMA) Scraper — European drug regulatory decisions.

---

📋 Changelog

Version 1.2 (2026-11)

• Unlocked 1,000+ Deep Pagination: Removed restrictive pagination ceilings; streamlined skip-pagination engine to traverse up to 25,000 records per query cleanly.
• Fast Prefill Execution: Optimized default batch parameters to complete test runs in under 10 seconds while providing 100 rich sample records.
• Enhanced Clinical Array Parsing: Standardized flattening of FAERS adverse drug reactions, concomitant medication lists, and MedDRA clinical terms.
• Robust 429 Retry Engine: Integrated progressive exponential backoff on high-volume rate-limited requests.

Version 1.1 (2026-10)

• Added automatic ISO date formatting (YYYY-MM-DD) from raw FDA integer timestamps.
• Unified top-level title, date, and summary schema across all six distinct endpoints.

Version 1.0 (2026-09)

• Initial release supporting drug, device, and food recalls, 510(k) clearances, NDC directory, and FAERS adverse events.