README.md

---
license: other
license_name: public-domain
task_categories:
- tabular-classification
- tabular-regression
tags:
- healthcare
- pharmacovigilance
- drug-safety
- fda
- adverse-events
- united-states
pretty_name: FDA FAERS Drug Adverse Events 2023
size_categories:
- 1M<n<10M
---

# FDA FAERS Drug Adverse Events 2023

FDA Adverse Event Reporting System (FAERS) data for 2023 — cleaned, deduplicated,
and structured for pharmacovigilance and drug safety research.
1.5M+ adverse event reports across 7 relational tables, with normalised drug names
and MedDRA preferred reaction terms.

**This repository contains a 1,000-row sample of the Demographics table (Public Domain).**
Full dataset (all 7 tables in CSV + Parquet) available at
[claritystorm.com/datasets/fda-faers](https://claritystorm.com/datasets/fda-faers).

## Quick Start

```python
from datasets import load_dataset
ds = load_dataset("claritystorm/fda-faers-drug-adverse-events")

import pandas as pd
df = pd.read_csv("sample_1000.csv")
print(df["sex_label"].value_counts())
print(df["age_years"].describe())
```

## Schema (DEMO table, selected fields)

| Field | Type | Description |
|-------|------|-------------|
| primaryid | string | Unique report identifier |
| caseid | string | Case ID (groups versions of same case) |
| caseversion | int | Case version (deduped: latest version kept) |
| fda_dt | string | FDA receipt date (YYYY-MM-DD) |
| rept_dt | string | Report date (YYYY-MM-DD) |
| age_years | float | Patient age in years (normalised) |
| sex_label | string | Male / Female / Unknown |
| reporter_type | string | Physician / Consumer / Pharmacist / etc. |
| report_type | string | Expedited / Periodic / Direct / Voluntary |
| wt_kg | float | Patient weight in kg (normalised) |
| _quarter | string | Source quarter (e.g. 2023Q1) |

## Tables in Full Dataset

- **demo** — 1.5M+ rows: one row per deduplicated adverse event report
- **drug** — 7.4M+ rows: drugs involved in each report
- **reac** — 5.8M+ rows: MedDRA preferred reaction terms per report
- **outc** — 1.2M+ rows: serious outcome codes per report (Death, Hospitalization, etc.)
- **ther** — 2.6M+ rows: drug therapy start/end dates per report
- **indi** — 4.5M+ rows: drug indication (reason for use) per report
- **rpsr** — 52K+ rows: report source per report

All tables join on `primaryid`.

## Source

US Food and Drug Administration (FDA), Adverse Event Reporting System (FAERS).
FDA FAERS data is a US federal government work in the **public domain** (17 U.S.C. 105).

Processed by [ClarityStorm Data](https://claritystorm.com).