Spontaneous Reporting of Adverse Drug Events in Brazil: VigiMed Dataset (2018–2025)

Abstract/Description: This dataset comprises Individual Case Safety Reports (ICSRs) extracted from VigiMed, the Brazilian national database for pharmacovigilance (ANVISA). It covers spontaneous reports of adverse drug events from 2018 to 2025. Technical Specifications: The data is provided in Parquet format, optimized for high-performance analytical queries. The extraction was automated via GitHub Actions, with data processing and consolidation performed using Python, SQL, and DuckDB. Data Content: The dataset includes patient demographics, suspected drug information (active ingredients, dosage), and adverse reaction details (MedDRA terminology, seriousness, and clinical outcomes). Sensitive personal data has been removed to ensure de-identification and compliance with data privacy regulations. Usage and Reproducibility: Ideal for research in pharmacovigilance, drug safety surveillance, and data mining applications. The Parquet files are ready for direct ingestion into DuckDB or Python (pandas/polars) environments. The source code for the ETL pipeline is available at: https://github.com/mlab-inf-ufrgs/vigimed