Top 3 Approaches using Machine Learning of EHRs to Increase Clinical Trial Recruitment

It is estimated that six million people per year are needed to fulfill US clinical trial recruitment goals, yet only two million people per year participate.  Low recruitment levels are due to a variety of different reasons and traditionally it has been difficult to address each issue on a broad scale.  However, with the push to implement artificial intelligence and machine learning in drug development processes, it is believed that deep learning and predictive modeling obtained from scanning electronic health records (EHRs) will help with the patient recruitment process.



Here are our top 3 approaches where we believe machine learning of EHRs will increase clinical trial recruitment:


Identifying candidate patients

Many patients, especially those with rare or chronic disease are never connected with clinical trials that they may have interest in participating in.  Many patients are not fortunate enough to have a clinician who is well-versed and connected with clinical trials network.  Therefore, the potential to reach this population of patients will dramatically increase enrollment.


Increase speed of data reporting

Keeping a close eye on contraindications and adverse events on a rolling and fast scale will undoubtedly help to ensure candidate patients that safety is closely monitored.  This may narrow patient pools, yet at the same time open the door to opportunities for more suitable clinical trial participants that is more likely to result in success. 


Decrease data silos

EHRs have the potential to bridge information gaps and reduce error prone data reporting within and between clinicians, clinical trials sites and any other stakeholder.  Transparency and sharing of data will certainly lead to an increase awareness of clinical trials and their respective patient candidate profiles.