Analytical Wizards Identify Underdiagnosed Patients

Harness the predictive power of AI, NLP for richer insights from medical notes

A pharmaceutical company planned to launch a first-in-class treatment for an ultra-rare disease that recently became medically identifiable as a unique illness (no ICD-10 code). The company needed to identify and size the eligible patient pool to inform launch planning (e.g., Product supply stock, physician outreach, patient education, etc.) The company asked Analytical Wizards (AW) to estimate the size of the eligible patient universe, identify care sites and physicians to call on at launch and develop a forward-looking prediction on the emergence of new patients.

AW analyzed unstructured EMR clinical notes using NLP and medical expertise.

The NLP analysis of unstructured EMR clinical notes integrated with data available through claims data uncovered new learnings. Likelihood scores were then assigned to the annual universe of eligible patients based on predictive analytics and given the probability of illness (highly likely, likely, unlikely).

Business Results:

Identified ~4,400 patients highly likely to have the rare condition

Identified physicians critical to reach with product information and disease education at launch

Enabled the pharma company to create an accurate product supply forecast

IZE Patient Finder

Create dynamic, trigger-based, targeting

Predict eligible patients

Use in combination with KOL Intelligence to identify physicians with largest eligible patient pool