FDA FAERS Dataset
About the Dataset
The FDA FAERS (FDA Adverse Event Reporting System) Dataset is a comprehensive data product providing normalized, cleaned, and analytically-ready access to pharmaceutical drug adverse event reports. This dataset encompasses over 18 million adverse event records, 65 million drug records, 54 million reaction records, and 18 million patient records, with advanced data quality enhancements and 4 pre-aggregated analytical reporting models for drug safety surveillance.
Quick Access
Base Tables: Adverse events, drugs involved, patient demographics, reported reactions, and report sources Aggregate Models: 4 pre-built analytical views for drug safety analysis, risk profiling, and trend monitoring Update Frequency: Quarterly from FDA FAERS database
Overview
The FDA FAERS Dataset provides comprehensive access to pharmaceutical drug adverse event data including:
Core Event Data (drug__events) - Main table with cleaned adverse event reports
Drug Information (drug__events_drugs) - Detailed information about drugs involved in events
Patient Demographics (drug__events_patients) - Normalized patient data with age and demographic information
Adverse Reactions (drug__events_reactions) - MedDRA-coded reactions and outcomes
Report Sources (drug__events_report_sources) - Information about who reported the events
Analytical Reports - Pre-aggregated intelligence for drug safety analysis, demographic risk profiling, and trend monitoring
Dataset Structure
The FDA FAERS dataset is organized around adverse event reports, with connected tables containing detailed information about the drugs involved, patient demographics, reported reactions, and report sources.
Base Tables
drug__events (Main Events Table)
Core adverse event reports containing event-level information. Each record represents a single adverse event report identified by a unique safety_report_id.
Key Features:
Over 18 million adverse event records
Clean, standardized data ready for analysis
Complete event details including dates, countries, and seriousness criteria
Reporter organization information for source tracking
drug__events_drugs (Drug Details Table)
Drug-specific information for each adverse event. Each record represents one drug involved in an adverse event report.
Key Features:
Over 65 million drug records with detailed information
Drug characterization (Suspect, Concomitant, Interacting)
Dosage information and administration routes
Indications for use and actions taken after the event
drug__events_patients (Patient Information Table)
Patient demographic information with comprehensive data cleaning and normalization.
Key Features:
Over 18 million patient records
Standardized age data in years and FDA age group categories
Weight data normalized to kilograms
Age at onset information for temporal analysis
drug__events_reactions (Adverse Reactions Table)
Reported adverse reactions and outcomes using standardized MedDRA terminology.
Key Features:
Over 54 million reaction records
MedDRA Preferred Terms and Lowest Level Terms
Reaction outcomes (Recovered, Fatal, etc.)
Links to specific patients and drugs through sequence numbers
drug__events_report_sources (Report Sources Table)
Information about the sources of adverse event reports.
Key Features:
Reporter qualifications and credentials
Primary source country information
Report source types for quality assessment
Links to main event records
Aggregate/Reporting Models
drug__agg_drug_reaction_associations
Pre-aggregated drug-reaction associations with statistical significance measures for safety signal detection.
Use Cases:
Drug safety signal detection and monitoring
Adverse reaction frequency analysis
Comparative safety assessments across drugs
Pharmacovigilance reporting and alerts
Key Metrics:
Association strength scores between drugs and reactions
Occurrence counts and statistical significance
Serious event and death rates by drug-reaction pair
Percentage of drug reports with specific reactions
drug__agg_manufacturer_adverse_events
Manufacturer-level adverse event analytics with portfolio safety metrics and competitive benchmarks.
Use Cases:
Manufacturer safety performance benchmarking
Portfolio risk assessment across drug products
Competitive intelligence for pharmaceutical companies
Regulatory compliance monitoring
Key Metrics:
Total reports by manufacturer with YoY comparisons
Serious event and death rates by company
Number of unique drugs with adverse events
Trend analysis and growth metrics
drug__agg_demographic_risk_profiles
Demographic risk analysis segmented by patient age groups and gender for targeted safety monitoring.
Use Cases:
Identifying high-risk patient populations
Age and gender-based safety signal detection
Pediatric and geriatric drug safety monitoring
Demographic-specific risk assessment
Key Metrics:
Risk scores by demographic segments
Event counts and rates by age group and gender
Comparative risk levels across populations
Statistical significance thresholds (10+ events)
drug__agg_safety_trends
Time-series safety trends with statistical trend detection and forecasting capabilities.
Use Cases:
Safety signal trend monitoring over time
Emerging risk detection and early warnings
Seasonal pattern analysis in adverse events
Post-market surveillance and monitoring
Key Metrics:
Multiple time granularities (Daily, Weekly, Monthly, Quarterly)
Moving averages and trend indicators
Serious event and death rate trends
Statistical trend direction classifications
Data Quality Improvements
Patient Information Standardization
Age Normalization
All patient ages converted to years for consistent analysis
FDA age group categories applied (Neonate, Infant, Child, Adolescent, Adult, Elderly)
Handles various reporting formats and units
Weight Standardization
All weights normalized to kilograms
Automatic unit detection and conversion
Validation for realistic weight ranges
Demographic Enhancement
Gender categories standardized (Male, Female, NULL for unknown)
Missing data clearly identified for transparency
Age at onset calculations for temporal analysis
Drug Information Enhancements
Drug Name Standardization
Medicinal product names cleaned and normalized
Brand names, generic names, and ingredients handled
Consistent formatting for analysis
Characterization Categories
Clear classification: Suspect, Concomitant, Interacting
Action taken standardization (Withdrawn, Dose reduced, etc.)
Route of administration normalization
Reaction Coding
MedDRA Standardization
Preferred Terms (PT) and Lowest Level Terms (LLT) included
Hierarchical coding system for reaction analysis
Outcome standardization for severity assessment
Getting Started
Basic Query Examples
Recent Adverse Events Overview
-- Search for adverse events involving a specific drug
SELECT
e.safety_report_id,
e.receipt_date,
e.occur_country,
e.serious,
d.medicinalproduct,
d.drug_characterization,
d.drug_indication,
d.drug_dose_text
FROM drug__events e
INNER JOIN drug__events_drugs d
ON e.safety_report_id = d.safety_report_id
WHERE d.medicinalproduct ILIKE '%ozempic%'
LIMIT 100;Top Adverse Reactions Analysis
-- Count top adverse reactions for a specific drug
SELECT
r.reaction_meddra_pt as adverse_reaction,
COUNT(*) as report_count,
COUNT(CASE WHEN e.serious = true THEN 1 END) as serious_cases,
ROUND(COUNT(CASE WHEN e.serious = true THEN 1 END) * 100.0 / COUNT(*), 2) as serious_percentage
FROM drug__events e
INNER JOIN drug__events_drugs d
ON e.safety_report_id = d.safety_report_id
INNER JOIN drug__events_reactions r
ON e.safety_report_id = r.safety_report_id
WHERE d.medicinalproduct ILIKE '%rinvoq%'
AND d.drug_characterization = 'Suspect'
AND r.reaction_meddra_pt IS NOT NULL
GROUP BY r.reaction_meddra_pt
ORDER BY report_count DESC
LIMIT 10;Geographic Distribution Analysis
-- Analyze adverse events by country and seriousness
SELECT
occur_country,
COUNT(*) as total_reports,
COUNT(CASE WHEN serious = true THEN 1 END) as serious_reports,
COUNT(CASE WHEN seriousness_death = true THEN 1 END) as fatal_reports,
ROUND(COUNT(CASE WHEN serious = true THEN 1 END) * 100.0 / COUNT(*), 2) as serious_percentage
FROM drug__events
WHERE receipt_date >= CURRENT_DATE - INTERVAL '2 years'
AND occur_country IS NOT NULL
GROUP BY occur_country
ORDER BY total_reports DESC
LIMIT 15;Advanced Analytics Examples
Demographic Risk Analysis
-- Analyze patient demographics for serious adverse events
SELECT
p.patient_sex,
p.patient_age_group,
COUNT(*) as patient_count,
AVG(p.patient_age_years) as avg_age,
COUNT(CASE WHEN e.seriousness_death = true THEN 1 END) as fatal_cases
FROM drug__events e
INNER JOIN drug__events_patients p
ON e.safety_report_id = p.safety_report_id
WHERE e.serious = true
AND e.receipt_date >= '2022-01-01'
AND p.patient_sex IS NOT NULL
AND p.patient_age_group IS NOT NULL
GROUP BY p.patient_sex, p.patient_age_group
ORDER BY patient_count DESC;Manufacturer Safety Performance
-- Top pharmaceutical manufacturers by adverse event volume
SELECT
e.sender_organization_name as manufacturer,
COUNT(DISTINCT e.safety_report_id) as total_reports,
COUNT(DISTINCT d.medicinalproduct) as unique_drugs,
COUNT(CASE WHEN e.serious = true THEN 1 END) as serious_events,
ROUND(COUNT(CASE WHEN e.serious = true THEN 1 END) * 100.0 / COUNT(*), 2) as serious_rate
FROM drug__events e
LEFT JOIN drug__events_drugs d
ON e.safety_report_id = d.safety_report_id
WHERE e.receipt_date >= '2023-01-01'
AND e.sender_organization_name IS NOT NULL
AND e.sender_organization_name != ''
GROUP BY e.sender_organization_name
HAVING COUNT(DISTINCT e.safety_report_id) >= 100
ORDER BY total_reports DESC
LIMIT 20;Temporal Trend Analysis
-- Monthly adverse event trends
SELECT
DATE_TRUNC('month', receipt_date) as report_month,
COUNT(*) as total_events,
COUNT(CASE WHEN serious = true THEN 1 END) as serious_events,
COUNT(DISTINCT sender_organization_name) as reporting_companies,
COUNT(DISTINCT occur_country) as reporting_countries
FROM drug__events
WHERE receipt_date >= CURRENT_DATE - INTERVAL '2 years'
AND receipt_date < CURRENT_DATE
GROUP BY DATE_TRUNC('month', receipt_date)
ORDER BY report_month DESC;Using Pre-Aggregated Analytics
Drug-Reaction Association Analysis
-- Use pre-built drug-reaction associations for quick insights
SELECT
drug_name,
reaction,
occurrence_count,
serious_rate,
death_rate,
association_strength_score
FROM drug__agg_drug_reaction_associations
WHERE occurrence_count >= 1000 -- Focus on statistically significant associations
AND serious_rate >= 50.0 -- High seriousness rate
ORDER BY association_strength_score DESC
LIMIT 25;Demographic Risk Profiling
-- Identify high-risk demographics for specific drugs
SELECT
drug_name,
age_group,
patient_sex,
event_count,
serious_rate,
risk_score,
risk_level
FROM drug__agg_demographic_risk_profiles
WHERE drug_name ILIKE '%metformin%'
AND event_count >= 10 -- Statistical significance threshold
ORDER BY risk_score DESC;Safety Trend Monitoring
-- Monitor monthly safety trends for a drug
SELECT
drug_name,
time_period,
period_type,
event_count,
serious_rate,
trend_direction,
moving_avg_30d
FROM drug__agg_safety_trends
WHERE drug_name ILIKE '%humira%'
AND period_type = 'Monthly'
AND time_period >= '2023-01-01'
ORDER BY time_period DESC;Why Choose This Dataset
Ready for Analysis: Clean, normalized data with standardized terminology
Comprehensive Coverage: Complete adverse event data with all related information connected
Pre-Built Intelligence: Aggregated analytics for immediate insights without complex queries
Statistical Significance: Built-in thresholds and association measures for reliable analysis
Performance Optimized: Denormalized keys and pre-aggregations for fast query performance
Data Update Frequency and Freshness
Update Schedule
FDA Updates: FDA releases new FAERS data quarterly
Our Updates: We check for new data twice daily and load it immediately when available
Typical Data Lag: Drug adverse event data is typically 3-6 months behind real-time due to FDA's quarterly release cycle
Understanding Data Timeliness
The FDA FAERS data has a significant lag due to the quarterly reporting cycle:
Adverse events are collected throughout each quarter
FDA processes and validates reports after quarter end
Data is typically released 3-4 months after the quarter closes
What This Means for You
Most recent adverse event data will be 3-6 months old
This is inherent to FDA's quarterly release schedule for ensuring data quality
Important Data Interpretation Guidelines
Adverse event reports are voluntary submissions and may not represent all occurrences
Reports do not establish causation between drugs and adverse events
Duplicate reports may exist despite FDA's deduplication efforts
Reporting rates can be influenced by media attention, new drug launches, and other factors
Statistical associations should be validated with clinical expertise
FDA Documentation Resources
Last updated