Unstructured EMR Data Expands Cohort Identification by 1,500%

A biopharma manufacturer developing a rare-disease treatment needed to size its potential patient population, but standard ICD-10 coding "failed to unearth many patients or distinguish between disease sub-types." Panalgo's data experts used unstructured EMR data alongside lab results and claims data to build comprehensive longitudinal patient histories and identify a much broader eligible cohort of the specific disease subtype. This unstructured-data approach expanded cohort identification by 1,500% compared with ICD-10 coding alone, giving the manufacturer a far larger, more accurately characterized eligible patient population.

The customer is anonymised in Norstella’s own case study — not withheld by us.

Norstella is a holding company merging several pharma-intelligence brands — Citeline (trial intelligence), Evaluate (forecasting), MMIT (payer/formulary data) and Panalgo (real-world analytics) — into one data asset.

This sits in the Data Providers stage of the AI use-case lifecycle: Named examples from the real-world-data, claims and competitive-intelligence providers pharma AI systems are trained on, benchmarked against, or run inside of.