MACHINE LEARNING AND

Identifying Patients for Clinical Trials

Big data offers an opportunity to dramatically enhance the patient recruitment process. The cloud contains terabytes of data that companies can draw upon to target their recruitment efforts with far greater precision:

  • For years now, Tudor Reilly Health, a patient recruitment firm, has relied on what it describes as a “sophisticated algorithm that layers both public and privately-held data” to reach those who are most likely to participate in a clinical trial (source).


  • Public data may include a variety of statistics that suggest whether an area is likely to yield participants, such as hospital admissions, prevalence of certain diagnoses, the smoking rate, the drug abuse rate and a variety of demographic data that can help guide recruitment towards populations most likely to include those who would qualify or be willing to participate in trials.


  • The next stage of the data revolution will likely focus just as much on harnessing the wealth of private data that people are constantly generating and sharing online, including with mobile health applications and the Internet of Things (IoT, aka FitBits, Apple Watches, etc) and using large sets of unstructured data based on conversations from social networks (“Why isn’t there anything that works for my migraines!!??”).