Clinical researchers have begun adopting predictive analytics to anticipate a patient’s engagement in a clinical trial, including if and how a patient will adhere to their treatment plan. As these models are fed more patient-level audio and visual data, they have the potential to open up a whole new world of a patient’s lived experience – from how a patient will respond to their medication, to the likelihood of developing side effects, to the broader trajectory of their disease. The continued development of this technology means we can not only anticipate how a clinical trial will progress, but also enable clinicians to provide more proactive, personalized care and truly move the needle on precision medicine.
Predicting dosing behavior
Having a rich dataset of dosing behavior means we can accurately predict a patient’s adherence and open up a number of opportunities to mitigate risk in a clinical trial. Knowing a patient’s past behavior is indicative of their future behavior, we can build models to anticipate the likelihood they will adhere to their treatment plan and work to improve patient engagement in a more targeted, personalized way. Because dosing regimens often become habitual, the behaviors that contribute to these routines are commonly observable from the outset of a clinical trial. By implementing a pre-trial screening using a two-week placebo period, we can observe participants’ behavior, make highly accurate predictions about the support they may need once enrolled, and even the likelihood that they will drop out.
Taking this precautionary measure is crucial to understanding the patients one is working with and optimizing a trial’s cohorts. These models can utilize rich information beyond if a dose was taken, including how long it takes a patient to dose, variations in the time of day they dose, how long after their alarm time they dose, and more. This predictive insight allows clinicians to take proactive rather than reactive measures and offer patient-specific support, and helps sponsors ensure balance across treatment arms.
Predicting a patient’s health journey
With enough high-quality data, predictive insights have the potential to extend far beyond adherence. By leveraging patient-level video and audio biomarker data, we can make significant strides in predicting how a patient will respond to medication and course-correct in a timely manner if a patient isn’t reacting as anticipated. Let’s take depression, for example. When developing a selective serotonin reuptake inhibitor (SSRI) therapy, participants often must wait four to six weeks before they are reassessed. With the accumulation of enough data, we could, in theory, develop a predictive model that allows clinicians to know in two weeks whether a person is likely to respond to a particular treatment and intervene much sooner. Predictive analytics could also help predict the entire course of a disease. For example, multiple sclerosis and Crohn’s disease often have a waxing and waning presentation over long time periods. With effective, long-term longitudinal data collection of symptoms over time, clinicians may someday predict flare-ups and disease progression to inform the future of a clinical trial and a patient’s care.
Easing burden through predictive insights
The use of predictive analytics has the potential to significantly reduce the burden on patients and clinicians. Patients can get answers about their health and response to treatment sooner, while clinicians can use their time and effort efficiently and in a more targeted way to work with those most in need of assistance. If properly implemented and managed, these insights can help sponsors better understand diverse patient populations and the nuances of an individual’s trajectory sooner, helping to accelerate trial timelines and bring potentially life-saving drugs to those that need them sooner.
