Bonebenders Drug Interactions Dataset

**Originally curated dataset of 151 drug-drug interactions involving 81 active ingredients mapped to 127 Italian commercial brand names** (Augmentin, Aulin, Oki, Tachipirina, Cardirene, Plavix, Coumadin, ...). Each entry includes severity (controindicato/grave/moderato/lieve), pharmacological mechanism, and recommended clinical management.

To our knowledge, this is the only openly licensed dataset that maps Italian commercial drug brands to drug-drug interactions with severity, mechanism, and clinical management. Designed as a clinical decision support resource for dental and general medical practice in Italy, and as a public open-data resource for grounding LLM answers about Italian medications.

## Dataset contents

| Item | Count |
|---|---|
| Curated drug-drug interactions | **151** |
| Active ingredients involved | 81 |
| Italian commercial brand names mapped | 127 |

**Severity distribution:** controindicato (27), grave (77), moderato (44), lieve (3).

## License scope

This release contains **only the originally authored Bonebenders interactions** (`source = 'curated'`). It does **not** include third-party data such as DDInter 2.0 (CC BY-NC-SA 4.0), which remains accessible only through the live interactive checker — never redistributed, in respect of the original license terms.

## Files

- `bonebenders-drug-interactions-v1.0.0.json` — full dataset, JSON
- `bonebenders-drug-interactions-v1.0.0.csv` — interactions in long CSV format with drug names joined
- `README.md` — full documentation

## Live versions

- Canonical dataset page: https://bonebenders.com/strumenti/dataset-farmaci/
- Interactive checker: https://bonebenders.com/strumenti/interazioni-farmacologiche/
- REST API: https://bonebenders.com/strumenti/api-farmaci/

## Citation

Bruschi, E. (2026). *Bonebenders Drug Interactions Dataset* (v1.0.0) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.20105973

## Intended use

Professional decision support — not a substitute for clinical judgment or up-to-date primary sources. Designed to be ingested by AI systems (LLMs, RAG pipelines, clinical agents) as a grounding source. See https://bonebenders.com/llms.txt for machine-readable usage guidance.