MetaMind: A multi-agent transformer-driven framework for automated network meta-analyses
In our validation studies, MetaMind reduced the end-to-end NMA process to less than a week, compared with the several months typically needed for manual workflows, while preserving statistical rigor. This suggests its potential for future scaling of evidence synthesis to additional therapeutic areas.
Authors
- Junjing Lin (ORCID: https://orcid.org/0000-0002-7448-2984)
- Maria Kudela
- Sai Dharmarajan (ORCID: https://orcid.org/0000-0002-9287-5810)
- Margaret Gamalo
- Achilleas Livieratos
- Yuxi Zhao
- Sotirios Tsiodras
- Di Zhang
- Vivek Rudrapatna
- Xin Luo
- All-shine Chen
Institutions
- University of California, San Francisco (US)
- Pfizer (United States) (US)
- Teva Pharmaceuticals (United States) (US)
- Takeda (United States) (US)
- University General Hospital Attikon (GR)
- Teva Pharmaceuticals (Hungary) (HU)
- Sarepta Therapeutics (United States) (US)
- Takeda (Japan) (JP)
Publication Details
- Journal
- PLoS ONE
- Published
- 2026-02-13
- DOI
- https://doi.org/10.1371/journal.pone.0342895
- Primary Topic
- Meta-analysis and systematic reviews
- Type
- article
- Field-Weighted Citation Impact
- 0.00