Improving Clinical Trials IQ with the Launch of Jeryl on CenterWatch iConnect

We are excited to release Jeryl, version 4.0 of iConnect on CenterWatch (more on why and who is Jeryl here).  This version is our biggest ever release and includes a major overhaul of our clinical trial search and comprehension features. We believe that the Jeryl release builds on the work CenterWatch has done since 1994 (yes, that is 25+ years as the oldest online clinical trials listing and search service)  to help patients find clinical trials near them.

Going Beyond for the next decade

The critical problem that we have attempted to address in the Jeryl release is providing ways for consumers to improve clinical trial comprehension or what we have termed their “Trial IQ”.  As an industry, we continue to  grapple with the issue of technical jargon that exists in the clinical trials content submitted to the ClinicalTrials.gov registry, as it serves as an important source of trials for us and also other clinical trial finders. And it is not just the trial content (i.e., title, description) which is difficult for potential participants to understand, but even the metadata, or the ABCs of a trial such as phase, type, therapeutic area, that are daunting for potential participants to comprehend or further, use these data elements to filter trials relevant to them.

With the Jeryl release, we are #GoingBeyond our previous efforts over more than a decade of building tools to empower patients to find trials matching their medical condition in integration with Personal Health Records and later the NCI awarded AI enabled guided trial finder, Dory that is live on MJ Fox, Alzheimer’s Association and the Rett Syndrome. We have employed data science, machine learning and several UI innovations  in the Jeryl release to enable patients to more easily filter through such complex content and hopefully to better understand it and act upon it.

Key Innovations to aid Trial IQ in Jeryl 

Contextual Keywords/labels

To improve on our search capabilities, potential participants can now narrow results using specific keywords that have been derived by a machine learning algorithm that surfaces contextual keywords relevant to the user’s search query. These contextual terms are derived from the trial content and are categorized  into  terms that may indicate what a patient “has/had” such as past medical conditions, treatments or procedures and what they may be “looking for” such as treatments/procedures. These terms are weighted using information gain algorithms to pick the most discriminating terms specific to the user search query.

We think these patient-centric terms will provide an easier to understand context than offering usual filters such as trial phase, etc or even asking a st of questions as we do in our decision engine. Dory.

Infoimages as visual ABCs of clinical trial

We have also introduced “infoimages”, images that provide a visual summary of the ABCs of a clinical trial.  We believe that this is the first such attempt to create a unique thumbprint of a trial based on a data-driven approach but which can provide one a quick overview of the trial at a quick glance.

Aids to Improve Study Comprehension 

Several innovations have been implemented on the study detail pages.

Each page now has  been given a look and feel to improve its readability. Summary keywords are displayed to provide a quick way for a user to ascertain if the trial is relevant to them or not. We have also made it easy for patients to understand technical terms by including definitions which have been linked using NLP to determine terms of interest and link them to semantically equivalent terms from wikipedia. 

Furthermore, we recognize that user’s may still have questions about the trial content, hence we have introduced the ability for user’s to make private notes by simply highlighting the text. All private notes are saved in a private workspace that can also be used to access save trials or messages sent to sites. 

We have also added the ability for participants to further check their eligibility – each inclusion /exclusion criteria has been parsed and check boxes provided. The user can also find a site near them and send a message to the study team closest to them. The clinical trial journey describing what happens next, is provided as an aid to help users understand what lies ahead.

Finally, we have used machine learning to suggest “similar trials” so that the user can view other trials of interest to them.

The development Process 

The Jeryl release was  planned and executed  over a period of almost a year. The main ideas emerged from several brainstorming sessions starting  October 2019 which were then sketched into wire frames and interactive prototypes by February 2020. The prototype was made with inputs from domain experts and development began in June 2020. Given the complexity and novelty of several new features, we employed a strategy of quick POC (proof of concept) development sprints to assess technical feasibility of some of the innovative features we had envisioned. This strategy was critical in ensuring we met our planned 4 month period for developing all these amazing and exciting features. Finally, before go-live we enlisted 8 patient advocates to give us feedback on the new site and used it to make suitable changes. We hope that these tools can serve to improve the TRIAL IQ of an average health consumer looking for a clinical trial. 

Stay tuned as we monitor site usage to understand how the new features are being used. And feel free to give Jeryl a spin and send us any feedback.