In a study published in Nature Medicine, a team of researchers highlighted the transformative potential of artificial intelligence (AI) in augmenting the patient experience within decentralised clinical trials (DCTs).
Also referred to as virtual or direct-to-participant trials, DCTs represent a paradigm shift in healthcare studies, enabling patients to participate in clinical trials from the comfort of their homes, according to the research. Highlighting AI’s role in revolutionising clinical research, the study suggests CROs adopt DCTs to accelerate the process of bringing life-improving therapeutics to the market.
The authors of the study emphasise the importance of driving innovation in DCT research models to maximise their potential and augment trial outcomes. The remote or hybrid nature of the trials places greater responsibility on trial participants, potentially leading to reduced adherence to protocols and participant dropouts.
The COVID-19 pandemic has been marked as an enabler of the fast adoption rate of DCT and hybrid study models, allowing CROs to keep their clinical studies running despite the global health crisis. However, transitioning away from on-site traditional trials has made researchers depend more on digital technologies such as sensors and wearables, telemedicine, patient-centric virtual trial interfaces, and more.
A critical challenge highlighted by the authors is the need to improve digital health interfaces. According to them, integrating AI automation in trial design can significantly augment user interfaces, ultimately benefitting the trial participants.
Another significant obstacle is the escalating volume of data collected during these trials. Even though eCOA, ePRO, or eConsent can mitigate trial shortcomings by offering flexibility to the patients and empowering sponsors with data integrity, they can sometimes lead to errors and reduced protocol adherence. This is because participants often make mistakes while inputting their health data through eCOA, diaries, or surveys. According to the study, by sending customised notifications to the trial participants, thus reducing the number of notifications they get, reinforcement learning (RL), an area in machine learning (ML), can help accelerate the completion of data entry tasks.
Again, in some DCTs—for example, dermatological research—participants are required to take images or videos. Adhering to specific image guidelines while taking these pictures is critical and may raise challenges for the participants. The authors urge CROs to implement computer vision—a branch of AI—in trial design that can streamline the process by automatically assessing and correcting images and videos.
Finally, the study highlights the need for real-time mobility tracking in some DCTs as another leading challenge for the trial participants. Sensors and wearables and sensors are critical for gathering data on mobility for assessing the moving patterns of patients. However, they can be cumbersome to apply, often requiring live guidance. According to the study, implementing AI models designed for temporal data can limit the use of wearables and sensors while enabling more accurate movement measurements.
With the response and exposure of over 20% of therapeutics showing variations with racial and ethnic differences, the FDA recognises minority inclusion in trials as a dire need. Citing a study published in the ASCPT Journal, the authors pointed out that DCT models can encourage a much higher patient recruitment and retention rate. High-end DCT platforms that include eCOA, ePro and eConsent such as ObvioHealth enable participants to contribute to clinical trials remotely. Such study models can effectively fill the patient inclusivity gaps in clinical research while also speeding up the long-overdue changes in trial design.
Sohela is an electrical engineer and a self-professed writer with a keen interest in all things tech. When she’s not writing killer content pieces, you’ll find her enjoying tempting foods in her favourite restaurants.
