
Modern clinical trials face an enrollment challenge. Over 80 percent of clinical trials conducted in the United States fail to meet their recruitment timelines, contributing to delays in therapeutic development, higher trial costs, and slower patient access to innovative treatments. Enrollment inefficiencies remain one of the most resource-intensive and time-consuming aspects of the clinical trial process. Despite growing access to real-world data (RWD), traditional recruitment methods have not evolved quickly enough to capitalize on these new information sources.
To move clinical research forward, the industry must rethink how it identifies eligible participants and deploys recruitment strategies.
Structured data alone misses critical clinical signals
Most recruitment efforts rely heavily on structured data fields such as claims, lab values, and ICD codes to identify potential participants. While this approach offers consistency and ease of querying, it often fails to capture the complexity of a patient’s health status or the nuanced criteria required by modern protocols. As a result, many potentially eligible individuals are missed, especially when eligibility depends on indicators that are not typically coded, such as functional status, treatment response, or progression captured through imaging.
These overlooked patients are frequently documented in unstructured parts of the electronic health record (EHR). This includes free-text physician notes, radiology reports, pathology narratives, and other clinically rich documentation. By focusing solely on structured data, recruitment teams risk bypassing a large subset of patients who could qualify for a trial based on their clinical history, but whose eligibility is not reflected in coded fields.
EHR unstructured data holds untapped potential
The majority of clinically relevant information in an EHR is unstructured. These text-based fields capture a physician’s impressions, reasoning, and context that often do not map neatly to dropdown menus or checkboxes. For example, disease progression may be noted as “increasing lesion size” in a scan interpretation, or a physician may describe a patient as “failing to respond to initial therapy.” These types of insights are vital for trial inclusion but are not captured by standard coding systems.
Unstructured EHR data provides a more holistic view of the patient journey. However, accessing it at scale has historically been a barrier. Advances in artificial intelligence (AI) and natural language processing (NLP) are now changing that reality.
How AI-powered tools unlock recruitment insights
Modern NLP platforms trained on clinical language can analyze unstructured text and extract key data points relevant to trial eligibility. These tools use rule-based models, machine learning classifiers, and terminology mapping to identify mentions of specific symptoms, disease stages, biomarker results, or response to prior therapies. Unlike keyword searches, these systems can interpret context and flag when a clinical term indicates progression, severity, or treatment failure.
For example, instead of relying on a diagnosis code for a condition like geographic atrophy (GA), AI tools can scan ophthalmology notes for references to visual acuity decline, lesion characteristics, or treatment plans. These data points can then be combined with structured EHR data to create a more complete profile of the patient.
To ensure the accuracy of these insights, successful implementations pair AI models with expert clinical validation. This process often involves training algorithms on annotated datasets, regularly reviewing flagged terms and extracted variables, and calibrating the system based on input from practicing physicians. Once validated, these models can operate across thousands of EHRs, enabling real-time identification of patients who meet complex inclusion and exclusion criteria.
Bringing structure and meaning to the entire EHR
To be effective, AI models must process both structured and unstructured data in a harmonized and standardized format. This includes ingesting EHR data from multiple sources, de-identifying and normalizing formats, and applying curation rules to ensure completeness and quality. Platforms designed for clinical development often integrate these capabilities, enabling researchers to define eligibility criteria with greater specificity and translate those criteria into search parameters across large, diverse datasets.
The result is a more dynamic, real-time approach to cohort discovery that supports faster feasibility assessments, smarter site selection, and earlier patient identification.
Building smarter, more inclusive trials with AI
By tapping into the full depth of the EHR, AI-driven recruitment strategies improve both precision and reach. These tools enable sponsors to find patients earlier in their disease journey, identify underrepresented populations, and better match trial design to real-world conditions. This contributes not only to faster enrollment but also to higher data quality and greater generalizability of trial results.
In an environment where speed, equity, and scientific rigor are all imperative, modernizing patient recruitment is no longer a future goal. It is a present necessity.
Real-world data, real-time impact
Artificial intelligence is no longer theoretical in clinical development. It is actively helping to reshape how trials are designed, launched, and executed. By transforming the EHR into a research-ready resource through advanced AI techniques, clinical oversight, and data standardization, the industry has an opportunity to fundamentally reimagine what is possible in trial recruitment.
Modern trials require modern infrastructure. Unlocking the full value of real-world data begins with understanding where the information resides, how to extract it responsibly, and how to convert it into insights that accelerate innovation and improve patient outcomes.
Photo: Andriy Onufriyenko, Getty Images
Sujay Jadhav is the Chief Executive Officer at Verana Health where he is helping to accelerate the company’s growth and sustainability by advancing clinical trial capabilities, data-as-a-service offerings, medical society partnerships, and data enrichment.
Sujay joins Verana Health with more than 20 years of experience as a seasoned executive, entrepreneur, and global business leader. Most recently, Sujay was the Global Vice President, Health Sciences Business Unit at Oracle, where he ran the organization’s entire product and engineering teams. Before Oracle, Sujay was the CEO of cloud-based clinical research platform goBalto, where he oversaw the acquisition of the company by Oracle. Sujay is also a former executive for the life sciences technology company Model N, where he helped to oversee its transition to a public company.
Sujay holds an MBA from Harvard University and a bachelor’s degree in electronic engineering from the University of South Australia.
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.