Chemular Regulatory Sciences
Literature reviews built for PMTA submissions.
A purpose-built workspace that walks scientists through the 1.10.1 review protocol — search, screen, appraise, and synthesize — and produces submission-ready 1.10.2 narrative drafts and 1.10.3 evidence tables.
Protocol-aligned
Every step maps to the Chemular 1.10.1 review protocol. Scientists drive the include / exclude decisions; the AI handles the volume.
Audit trail by default
Each decision, each appraisal, and each adjudicated conflict is timestamped and attributed. Outputs are reproducible from the database alone.
Built around regulatory deliverables
The pipeline ends in PMTA-shaped artifacts: PRISMA flow counts, the 1.10.3 Evidence Table, the 1.10.2 narrative scaffold, and screening workbooks.
How a review runs
The workspace orchestrates the literature flow end-to-end so the scientist's time is spent where it matters — judgment on the studies that survive screening.
- 1
Search & dedupe
The system runs your PubMed queries against the configured update window, normalizes the records, and removes duplicates against the prior review packet.
- 2
AI title-and-abstract screening
Every record is scored against the inclusion criteria with structured rationales. Low-confidence and exception records are flagged for the scientist.
- 3
Scientist abstract review
The scientist confirms includes from the AI's ranked queue. Records they don't include are out of scope — no extra step required.
- 4
Full-text retrieval & AI read
Open-access PMC articles are downloaded and parsed. The AI re-reads each at the full-text level, recommending Include or Exclude.
- 5
Scientist full-text review & evidence extraction
The scientist verifies AI full-text decisions and signs off on the AI-drafted appraisal — methods, key findings, limitations, PMTA domains, materiality.
- 6
Build outputs
The pipeline assembles the PRISMA counts, 1.10.3 Evidence Table, 1.10.2 narrative draft, and audit workbooks into a downloadable bundle.