AMIA2026

Materials associated with AMIA 2026 submissions by Jeffery L. Painter and collaborators.

This repository contains the poster, code, and released data artifacts supporting work on interoperability between PVLens and SIDER, enabling PVLens-derived drug safety information to be expressed in the standard SIDER data format.

---

Repository contents

• poster/ Poster PDF and figure source files for the PVLens--SIDER interoperability poster submitted for review at AMIA 2026.

• pvlens_sider_update/ Code, documentation, and release materials for generating SIDER-compatible replacement files derived from PVLens data while preserving SIDER STITCH/CID identifiers for ATC-matched drugs.

---

Featured poster

Enabling Interoperability Between PVLens and SIDER:
A Pipeline for Generating SIDER-Compatible Drug--Adverse Event Data

Poster PDF:
poster/Enabling_Interoperability_Between_PVLens_Painter.pdf

---

Project overview

PVLens extracts drug safety information directly from FDA Structured Product Labeling (SPL) documents using NLP and ontology-based mapping.

This repository provides a pipeline that:

1. Extracts adverse events and indications from the PVLens database\
2. Maps terms to MedDRA concepts\
3. Matches PVLens drugs to SIDER STITCH identifiers using ATC code overlap\
4. Generates SIDER-compatible output files that can replace the original SIDER tables for matched drugs

The resulting files maintain the original SIDER drug identifiers while substituting PVLens-derived safety data.

Pipeline overview

---

Download PVLens–SIDER replacement dataset

The fully generated PVLens-derived SIDER-compatible tables used in the AMIA 2026 poster are included in this repository.

Location:

pvlens_sider_update/release/files/

The release includes the following compressed TSV files:

• meddra_all_label_indications.tsv.gz
• meddra_all_indications.tsv.gz
• meddra_all_label_se.tsv.gz
• meddra_all_se.tsv.gz

These files are drop-in replacements for the corresponding SIDER 4.1 tables for drugs that can be matched to PVLens via ATC codes.

Additional metadata files:

• matched_flat_cids.csv – SIDER STITCH flat identifiers matched to PVLens drugs
• matched_atcs.csv – ATC codes used for drug matching
• sider_atcs.csv – ATC codes extracted from SIDER

All files include SHA-256 checksums listed in:

pvlens_sider_update/release/files/checksums.txt

---

Citation

If you use these materials, please cite:

Anthony McDonald (Georgia State University, Undergraduate), Jeffery L. Painter. Enabling Interoperability Between PVLens and SIDER: A Pipeline for Generating SIDER-Compatible Drug–Adverse Event Data. Submitted for consideration to the AMIA Annual Symposium 2026.

---

Related work: PVLens

This repository builds upon the PVLens pharmacovigilance framework, which extracts drug safety information from FDA Structured Product Labeling (SPL) documents using NLP and ontology-based mapping.

The PVLens project repository is available at: https://github.com/GSK-Global-Safety/pvlens

If you use PVLens in your research, please cite:

Painter, J.L., Powell, G.E., & Bate, A. (2025). PVLens: Enhancing pharmacovigilance through automated label extraction. AMIA Annual Symposium Proceedings, 2025, Atlanta, GA. DOI: 10.48550/arXiv.2503.20639

@article{pvlens2025,
  author       = {Painter, J.L. and Powell, G.E. and Bate, A.},
  title        = {{PVLens: Enhancing pharmacovigilance through automated label extraction}},
  journal      = {AMIA Annual Symposium Proceedings},
  publisher    = {AMIA},
  volume       = {2025},
  month        = {11},
  address      = {Atlanta, GA},
  type         = {Paper},
  doi          = {10.48550/arXiv.2503.20639}
}