Unstructured EHR Data Expands Cohort Identification by 1,500%

When a biopharma manufacturer began developing a new therapy to treat a rare disease, its leadership team realized that finding diagnosed patients would be a challenge. The company wanted to conduct feasibility research on the indication to map the patient journey, identify patients with a specific subtype, and quantify the patient population that would potentially be eligible for its pipeline therapy. Partnering with Norstella, the company:

Expanded rare disease cohort identification with unstructured EHR data
Identified eligible patients beyond claims data alone
Supported drug development with longitudinal patient insights

Key Results

1.5K%

Expanded Patient Identification

Challenge

The manufacturer had already approached several vendors to complete preliminary research using claims and registry data, but the analyses returned unexpectedly low patient counts. The company turned to Norstella for help verifying these counts, specifically requesting the use of its comprehensive closed claims dataset. After Norstella’s analysts confirmed that less than 200 patients were identifiable via claims, the company was ready to abandon its research efforts—and potentially its new therapy.

Solution

Norstella’s data experts urged the company to expand the scope of its research. They explained that although the rare disease was not new, it had not been assigned an ICD-10 diagnosis code until 2022. The coding classification system also failed to distinguish between the various sub-types of the disease, which made identifying those patients virtually impossible without additional datasets. Instead, physicians used an ‘unspecified’ placeholder code for all diagnosed patients, regardless of disease progression or type.

The biopharma company agreed to try a broader search, and Norstella’s experts began analyzing two additional real-world datasets powered by NorstellaLinQ: Norstella’s lab data and structured/unstructured EHR data. By leveraging curated data to pinpoint relevant terms from unstructured clinical notes, the data analysts were able to identify a much broader cohort of more than 3,000 patients with the specific sub-type of interest.

Outcome

The team extracted this deidentified patient data into structured fields, tokenizing and linking them to deidentified lab data as well as Norstella’s deidentified claims data. As these massive datasets encompass years of test results, procedures, diagnostic notes and claims, the biopharma company now has access to a rich longitudinal history for all identified patients with the rare disease.

Using unstructured EHR data to supplement the initial patient counts identified via claims alone proved to be a successful approach. As the biopharma company now has a confirmable cohort for future study, its leadership team is moving forward with drug development.

Get Started

Get Closer to the Patient Than Ever Before

NorstellaLinQ links claims, EHR, lab, and electronic prescription data into a single patient view — decision-grade and AI-ready. See what it can do for your team.

Frequently asked questions

Everything You Need to Know

How can unstructured EHR data improve patient cohort identification?

Unstructured EHR data captures valuable clinical information found in physician notes that may not exist in claims data alone. By analyzing these notes alongside structured healthcare data, researchers can identify significantly larger and more accurate patient cohorts for clinical research and drug development.

Why isn't claims data alone enough for rare disease research?

Claims data can be limited when rare diseases lack specific diagnosis codes or when disease subtypes are grouped under generic classifications. Supplementing claims with unstructured EHR and laboratory data provides a more complete view of the patient population and improves cohort identification.

How does integrated real-world data support drug development?

Integrating claims, laboratory, and structured and unstructured EHR data creates a longitudinal view of the patient journey. This enables biopharma organizations to conduct feasibility research, identify eligible patients, and generate stronger evidence to support clinical development.

What was the outcome of using unstructured EHR data in this case study?

By expanding beyond claims data and analyzing unstructured clinical notes, Norstella identified more than 3,000 patients with the targeted rare disease subtype—expanding the identifiable cohort by 1,500% and allowing the manufacturer to continue its drug development program.

How does NorstellaLinQ help identify rare disease patients?

NorstellaLinQ links claims, EHR, laboratory, and prescription data into a unified, longitudinal patient view. This integrated approach enables researchers to uncover patients who may be missed by traditional datasets, providing richer insights for evidence generation and clinical development.