Finding Insights in Real-World Data with AI

It’s no secret that healthcare generates massive volumes of valuable, real-world data (RWD). The challenge is extracting meaningful insights from this trove of information. 

Most data in electronic health records (EHRs) is unstructured – clinician notes, radiology reports, images, etc. – and doesn’t fit neatly into standardized formats or spreadsheets that can be readily analyzed or queried. 

With the advance of artificial intelligence (AI), machine learning (ML) and natural language processing (NLP) models make it possible to curate and harmonize unstructured RWD to improve clinical research and patient treatment. 

The critical factor in realizing these advances, however, is ensuring high-quality, properly validated data curation. 

Ensuring a Quality AI Advantage

As the healthcare industry increasingly adopts AI-driven insights from unstructured RWD, the movement is supported by FDA guidance and a growing range of use cases. Yet ML and NLP models are only as good as the data that powers them.

Underlying Verana Health’s success with AI-extraction of RWD is a deliberate approach with a clinically informed process at each step—from ingestion to curation to application.

Verana Health partners with leading medical associations to transform clinical data into real-world evidence (RWE). 

A team of clinicians, nurses, clinical informaticians, data scientists, epidemiologists, biostatisticians, and engineers work together to decide how to curate and standardize the data while retaining its original clinical context. Data is harmonized (integrating structured and unstructured data) and models are continuously refined to prevent bias and maintain accuracy. 

Verana Health works directly with practices and health care professionals to regularly review data quality and enable faster, more accurate implementation of AI in data curation.

Real World Results

In this way, Verana Health is able to provide robust, rich, specialty-specific datasets of de-identified medical information. These research-ready datasets help drive business insights and inform better research outcomes. For example:

• To support prostate cancer research, Verana Health used AI-powered models to analyze patterns of language in EHR unstructured notes and identify metastatic progression not coded but noted in clinician free-form text. The effort identified a five-fold increase in metastatic patients and captured important variables, including Gleason Scores and PSA levels, offering a more complete, accurate view of the patient population. 

• To advance ophthalmic research on geographic atrophy (GA), Verana Health used AI-powered NLP models to extract disease signals from unstructured clinical notes and images. The solution expanded a GA cohort by over 476,000 patients, improving insights into disease progression and treatment outcomes.

• To support research aimed at earlier treatment of bladder cancer, Verana Health overcame variability in EHR documentation of Non-Muscle Invasive Bladder Cancer (NMIBC) and defined intermediate-risk patients by applying AI-assisted patient segmentation based on details including tumor grade and stage. The results enhanced clinical trial recruitment and improved research on treatment effectiveness.

Supporting Research into the Future

AI-curated data is helping life sciences companies unlock critical insights hidden in unstructured data. Utilizing Verana Health’s AI-powered tools, researchers gain deeper disease understanding, improve trial efficiency, and accelerate timelines. Some applications include:

• Optimizing site selection by analyzing historical recruitment, patient demographics and disease burden.
• Driving patient recruitment by matching eligible patients to trials with predictive analytics that identify optimal inclusion criteria adjusted for real-world populations.
• Improving study design and reducing complexity through AI-assisted protocol optimization.
• Enhancing patient safety by identifying potential safety signals and assessing adverse event patterns across diverse populations.

Beyond the structured insights of AI models and curated datasets, emerging generative AI shows further potential to create content dynamically in response to changing situations, a critical factor in shortening time-to-market entry and speed of patient treatment. 

With the right expertise, therapeutic-specific knowledge, and experience, AI is rapidly opening a world of healthcare advances. 

To learn more about how to accelerate clinical trials and gain insights with real-world evidence, download our white paper.