Artificial intelligence has great potential to transform clinical trials by enhancing retention, aiding in participant recruitment, personalizing engagement, detecting dropout risk, recommending intervention strategies, and monitoring adverse events to benefit participants, sites, and sponsors – all at reduced cost.

Clinical trials play a vital role in advancing medical research and developing innovative treatments, but often face significant logistic and financial burdens to ensure smooth operations and optimize for success. Participant attrition throughout a study represents one of the most significant financial burdens to clinical trials. This problem is so widespread that “up to 85% of clinical trials fail to recruit or retain a sufficient sample size, leading to recruitment failures in four out of every five trials, even though nearly $1.9 billion is spent on recruitment annually”. A recent poll of pharma executives showed that “Increasing clinical trial retention” was their highest priority (64% of participants).

Recent work in AI/ML has shown that the integration of AI algorithms holds great promise in addressing challenges faced in participant retention. Here we outline some ways in which AI can and is contributing to improving clinical trial retention.

1. Participant Recruitment and Screening

• Optimizing retention for the duration of a trial begins during recruitment efforts. Being able to identify optimal participants through the prescreening process can reduce attrition throughout the study.
• AI-powered algorithms can identify inclusion/exclusion criteria that not only optimize trial performance but also maximize retention across trial participants.
• Incorporating AI algorithms early in the recruitment process can identify ways to better communicate with participants, identify participants with higher risk factors of dropping out in the future, and help coordinators with personalized communication.
• By empowering study coordinators with AI powered approaches before a study even begins, we can create a better, more personalized interaction with participants and the larger community, thereby improving retention throughout the study.

2. Personalized Patient Engagement

• Of the 12 general retention strategy themes outlined by Robinson et al, four of the themes are specifically focused on personalizing interactions with participants, thus ensuring consistent and robust engagement between the site and each participant. Unfortunately, personalizing communication efforts in large studies can result in significant cost to sponsors and sites, thereby limiting the adoption of such practices.
• Advances in generative AI offer solutions to personalize communication and engagement while minimizing costs; tailoring approaches to each individual participant.
• Natural language processing algorithms can analyze patient data, including medical history, preferences, and demographics, to deliver personalized reminders, educational materials, and motivational messages.
• Advances in generative AI allow for conversational AI and virtual assistants to provide real-time support, answer participant queries, offer guidance, and foster a sense of connection and support throughout the trial. Such a conversational AI (e.g. chatbot) was rated to have better usability, interactivity, and dialogue compared to just using a website to understand eligibility (Chuan and Morgan).

3. Early Detection of Dropout Risk

• AI algorithms can also identify user behaviors and actions that are indicative of higher dropout risk. These risk factors can be identified from diverse data sources, including patient-reported outcomes, wearable devices, and electronic diaries to detect early signs of participant disengagement.
• AI guided approaches allow for targeted retention efforts, thereby minimizing the cost and burden on sites and coordinators by moving away from resource intensive human interactions (Gamble et al.)
• By identifying patterns and indicators associated with elevated dropout risk, AI models can alert researchers and healthcare professionals, enabling timely interventions and personalized support to address potential concerns.

4. Identifying Recurrent Trends Leading to Drop Out

• Beyond identifying specific individuals with a high drop out risk, AI can also be deployed to identify aspects of a protocol, study design, or site interactions that lead to elevated attrition across studies.
• By identifying aspects of the protocol or study that are likely to contribute to higher attrition it becomes possible to alter protocols to minimize patient burden, plan for increased retention efforts during these steps, or increase recruitment to compensate for expected attrition.
• These analyses can further guide the implementation of interventions throughout a study, such as adjusting dosing schedules or improving patient support to compensate for attrition.
• This valuable insight allows researchers to proactively develop targeted interventions and strategies to mitigate risk factors and enhance participant engagement.

5. Real-Time Adverse Event Monitoring

• AI can continuously monitor participant data, as it is collected, to detect and assess adverse events promptly. This data could take many forms including electronic health records, electronic clinical outcomes assessment (eCOA), wearable sensor readings, and/or social media feeds.
• By detecting and identifying adverse events early, sites and coordinators can take immediate action, ensuring participant safety and well-being.
• These approaches maintain participant trust, which reduces the likelihood of dropout due to safety concerns and ultimately results in higher trust throughout the community.

Conclusion

Poor patient retention is a leading factor in study failure and elevated costs. While there is a significant body of research identifying strategies that lead to increased retention, many of these approaches are time intensive, placing significant burden on the sites and coordinators to enact.

Increasingly, sponsors, sites, and CROs are leveraging artificial intelligence, generative AI, and machine learning to increase retention. The types of approaches are varied, with applications throughout the clinical life cycle. They encourage increased retention by enhancing the overall trial experience for participants, building trust, encouraging better communication, and identifying risk factors leading to attrition. Combined, these approaches lead to quicker, lower cost clinical trials with a higher likelihood of success.

As AI continues to evolve, it holds the promise of optimizing clinical trial processes, accelerating medical advancements, and improving patient outcomes. Our team combines experience with AI driven methods for improving clinical trial retention and expertise in NLP and risk/prediction modeling to create bespoke solutions for supporting clinical trial success.