Supercharging Data Mining to Match Patients for Clinical Trials

AI data mining can be a powerful tool to help clinicians and researchers make successful matches.

Read time: 3 minutes

By: Katya Hall

For large health systems, especially those with cancer and infusion centers, clinical trial matching is an important area that supports patient care in two ways: by giving participants access to promising new treatments before they are available to the general public, and by helping drug makers gather the data they need to bring these drugs to market — safely and, ideally, affordably. AI data mining can be a powerful tool to help clinicians and researchers make successful matches.

Slide discussing the challenges of finding the right patients for clinical trials, listing inclusion/exclusion criteria such as age, gender, and health status.

The complexity of the matching process starts with the trial protocols themselves. Not only does each clinical trial have its own unique set of criteria, but there is no single comprehensive registry of all U.S. trials. As a result, health systems must sift through mountains of disparate data published in various formats and platforms to identify trials relevant to their patient populations. Then, the greater challenge begins — to scour both structured and unstructured patient data to identify potential matches. The electronic medical record (EMR) may have relevant details expressed in different ways in different sections, making it difficult for traditional software to process. Additionally, unstructured data like notes, scanned documents, and voice recordings have historically required manual review by staff who are already stretched thin with other tasks and responsibilities.

This time-intensive approach translates into missed opportunities to help patients connect with medications that have the potential to enhance outcomes for trial participants as well as countless others once the drug is approved. With these obstacles in mind, it’s easy to understand how trial sites fail to meet enrollment targets more than 37% of the time across the industry as a whole. In fact, in 10% of clinical trials nationwide, sites are not enrolling a single patient. These shortfalls delay drug development by more than 40% and can also contribute to rising operational costs for sponsors.

Image with text discussing the use of AI in healthcare for interpreting unstructured data and comparing costs of manual review vs. AI-enabled matching.

To learn more about McKesson’s investments in AI to help your health system achieve more, download the full eBook by filling out the form below.

While AI supports our services, and while we strive for accuracy, outcomes may vary, and AI may occasionally make mistakes. We are committed to using AI responsibly, focusing on transparency, privacy, security, fairness, and continuous improvements to enhance reliability and accuracy.

 

Subscribe to our monthly Prescribed Perspectives newsletter to have our exclusive content delivered directly to your inbox!

top