The EHR Dark Data Goldmine: Why Sentence-Level Provenance Is Pharma's New Real-World Evidence Moat
DDDr. Dereck Mush, MD, MBA
Sep 21
RespondHealth
Clinical NLP engine that turns unstructured clinician notes into research-grade real-world evidence.
What they do
RespondHealth extracts structured, sentence-level provenance data from clinical progress notes for drug surveillance and trial matching. A multi-center validation study published in Nature Medicine in September 2026 evaluated records from more than 16,000 GLP-1 patients.
The record
- Vertical
- Clinical AI & Documentation
- Sub-segment
- Clinical NLP and real-world evidence
- Product type
- Platform software
- Stage
- Series A
- Headquarters
- Bethesda, MD
- Region
- North America
- Founded
- Not disclosed
- Total funding
- Undisclosed
Also in Clinical AI & Documentation
Related on The Signal
How Much Weight Does Real-World Evidence Actually Carry at FDA?
DDDr. Dereck Mush, MD, MBA
Jul 18
Does AI Actually Improve Radiology Accuracy? What the Prospective Trials Show in 2026
DDDr. Dereck Mush, MD, MBA
May 21
Healthcare AI Has an ROI Problem: Here Is What the Evidence Actually Shows
DDDr. Dereck Mush, MD, MBA
Sep 13