The AI-activated clinical trial.


A new, targeted treatment could hold the answer for people born with rare diseases. Our client needed our help building and optimizing a clinical trial for this specific rare population, identifying and mapping investigator performance to relevant sites and patients, alongside the geographic distribution of patient clusters, investigators and identified sites.

Knowing that 80% of clinical trials are delayed due to inadequate site performance and recruitment, our client needed to understand the gap between the status quo and the optimal network of potential sites for the trial.

+ idea

By first identifying key performance indicators and metrics to ensure success, we utilized a highly sophisticated AI-driven model to determine the ideal patient profile characteristics, estimate the patient population size, map the patient journey and identify key investigators as well as optimal site locations.

The core of the patient, provider, and site profiling was driven by our proprietary Identity Graph, a unique, interconnected repository of data that covers all 6.2 million healthcare providers and includes over 4,000 data variables to choose and define audiences from.

As a last step, we came up with a ‘fit and recruitment’ scoring model to rank each site and investigator, based on their historical performance relative to clinical trial experience, indication expertise, diagnosed patient volume, site enrollment rate, time to report results and trial design complexity.

= impact!

Using this process, we were able to identify the size of the patient population, various patient hot-spots across the US, and ‘match-make’ patients to the best providers and trial site locations.

Through our process, our client identified more high-value investigators and more diagnosed patients in relation to the top thirty site locations. We were able to match a list of top sites with the top investigators and perform in-depth analysis, based on physician specialty, location, and diagnosed patient count.

We then generated geographic ‘heat maps’ through visualizing site data, overlayed with patients and providers within a 100-mile radius of an identified site.

Get in touch.