P227: In AI Do We Trust: Workforce Perspectives on AI Readiness and Integration in Clinical Trials
Poster Presenter
Aarthi Iyer
Senior Corporate Counsel
Cogent Biosciences, Inc. United States
Objectives
Conduct pilot research exploring clinical trial workforce perspectives on trust and accountability, regulatory clarity, appetite, and readiness in employing Artificial Intelligence as a collaborative tool to enhance the: clinical trial process, efficiency in research, and equity for participants.
Method
We conducted a literature review and employed a mixed-methods anonymous online survey questionnaire featuring both qualitative and quantitative questions as well as an AI tool use case, querying research participants across distinct clinical trials workforce populations: study staff and IRB members.
Results
A total of 47 participants (out of 62 screened) completed the survey, with 35 clinical research personnel and 9 IRB members. While most respondents indicated that they were either neutral or not very familiar with the diverse potential applications of AI in clinical trials, with a majority of clinical research personnel self-reporting to never having worked on a clinical trial that used AI tools, all respondents did highlight that AI could aid or improve the following clinical trial challenges: recruitment efficiency, study design complexity, data integrity, and regulatory compliance. Further, most respondents expressed comfort with an AI chatbot assisting with administrative workflow automation (e.g., such as contacting the research team about a patient’s concerning lab values). All respondents believe that use of AI tools should be subject to regulatory guidance and oversight but lacked confidence that current regulations are sufficient to manage the ethical use of AI in clinical trials. While all IRB respondents agreed that the integration of AI tools into clinical trial protocols is increasing with belief that integration of AI tools will become standard practice in the next 5 years, they were split on their respective IRBs: a) having dedicated policies, SOPs, or guidance to support review of AI tools in clinical trials and b) implementing specific language in informed consent forms when AI tools are being used in clinical trials. The majority of respondents expressed a need for more information on whether the benefits of using AI tools outweigh the risks, identifying potential for bias in algorithms and making inaccurate predictions or recommendations as the most concerning risks associated with using AI tools. While no respondent indicated that AI tools should make clinical decisions independent of human oversight, all respondents did agree that AI tools should be used alongside human researchers working in tandem on recruitment, monitoring, and data analysis.
Conclusion
While the use of AI tools in clinical trials is not a foreign concept, our preliminary results indicate it is still in its early stages and there are varying levels of familiarity, knowledge, comfort level with and readiness for widespread integration of AI tools across the clinical trials workforce. Our research revealed a need for more information, training, transparency, evidence generation, and creation of uniform regulatory guidance in order for adoption of AI tools in clinical trials becoming a best practice. In considering practical pathways and ethical considerations for implementing AI tools, priority should be placed on AI tools being able to work alongside humans and be subject to human oversight and decision-making, as having humans in the loop will increase trust and transparency for adoption of AI tools while improving efficiency within the existing clinical trial design framework. As respondents agree on the ability of AI tools to potentially aid with or improve traditional challenges encountered in clinical trials, there is also ample opportunity for AI tool developers to increasingly engage in collaborative pilots and research partnerships to jointly work with sponsors and the clinical trials workforce to educate, design, develop, and progress the responsible use of AI tools in clinical trials.
