Artificial intelligence (AI) promises efficiency, accuracy, and cost savings, but what can it actually deliver when it comes to clinical trials? Whether you’re in the design process of your clinical trial, working on regulatory submissions, recruiting participants, or collecting and analyzing data, there may be AI tools to help.
In June 2023, the first AI-developed drug entered phase II clinical trials, leading to a wave of excitement around the potential of AI in drug development and medical research, particularly in oncology trials. Many major pharma companies have forged partnerships with AI firms, using their tools throughout the research and development process.
Pharma companies expect AI to increase their efficiency: 40% of executives say expected savings due to AI are reflected in their budgets for 2024, and 60% have set cost saving targets.
But beyond the headlines, what is the reality of AI in clinical research today? And as a biotech or pharma company, how can you use AI to optimize your clinical trials?
So far, progress has been slow and steady; AI isn’t yet revolutionizing the way we work in clinical research, but it’s showing great promise. In this article, we look at some of the ways AI is already in use in clinical research, and we look at what’s needed to get the most out of the technology.
AI-discovered drugs accelerate clinical trial timelines
The first impact is that AI could be the source of the drug you’re testing in clinical research. According to Dave Lawtshaw II, Ph.D., CEO of BioPhy, “AI-native biotech companies and their pharmaceutical partners have entered 75 AI-discovered molecules into clinical trials since 2015, demonstrating a compound annual growth rate of over 60%.”Researchers are using AI tools to identify targets and synthesize pathways, create simulations, and prioritize candidates. The result is greater efficiency and accelerated timelines, which means clinical trials start earlier than they would in a traditional R&D setting.
What’s more, AI-discovered drug candidates have so far been more successful in phase I trials, at 80-90% compared to the 40-65% seen with traditionally developed drugs.
More efficient and informed trial design
AI can support the process of clinical trial design, with tools to help improve hypothesis generation and optimize protocol design. Harnessing the potential for AI to analyze data and spot patterns, the technology can provide predictions based on different scenarios, helping investigators adjust their approaches depending on the most likely outcomes.
Given its potential to create synthetic data, AI can also provide input that reduces the number of trial arms required. This can simplify the trial design and drastically reduce the costs and other resources needed to run the trial.
Generative AI for documentation
Over the past year, the use of generative AI tools like ChatGPT and Gemini has seen unprecedented growth across industries, including clinical research. Tools are emerging that support sponsors in producing regulatory documents, ranging from application before the trial starts to reporting when data collection is complete.
As with all generative AI tools, this technology can provide a starting point for experts, helping increase their efficiency.
More effective participant recruitment
Research has shown that more than half (55%) of trials that terminated did so because of low recruitment. Finding, enrolling, and keeping the right patients remains the biggest challenge in clinical trials. In a recent meta analysis published in Health and Technology, over half of the papers reviewed used AI in patient recruitment – more than double the number that used the technology for trial design and analysis.
In recruitment, AI technology is helping sponsors reduce sample sizes, for example. This is particularly helpful for rare disease trials, which face even greater recruitment challenges due to smaller numbers of suitable participants. AI tools are also supporting recruitment by automating eligibility analysis and participant screening, performing better than humans in this area. Examples include oncology and dementia trials, where AI is helping improve the chance of regulatory approval.
Improve clinical trial efficiency and success
The challenge doesn’t end with recruitment; patients leaving clinical trials can slow or even halt their progress. One of the factors affecting retention is the often-complex process of informed consent, specifically informed consent forms (ICFs). AI-powered eConsent tools are helping communicate the details of a trial to participants and ensure transparency and understanding.
AI can also help improve adherence through monitoring, including through AI-based sensors. AI can also identify potential issues such as adverse events and early withdrawal. In such cases, the AI tool could flag the issue to investigators, who could then check the data and follow up as needed to reduce risk of withdrawal.
Accurate and complete data
Data analysis is one of the areas that is most likely to benefit from added efficiency through AI: algorithms can analyze huge amounts of data fast, identifying patterns that would likely otherwise go unnoticed. AI can also support the analysis of medical images.
It’s also possible to solve the problem of incomplete data using AI, for example, filling in data from missing visits, which happened frequently during the COVID-19 pandemic. AI-driven digital twins can provide data to support the whole clinical research process, from helping improve patient recruitment and retention to providing input for additional trial arms.
Harnessing the potential of AI in clinical trials
AI is already proving to be a useful assistant in clinical trials, but it’s not yet fully reached its potential. The sensitive nature of medical data means questions around the reliability and privacy of generative AI in particular need to be answered. One step towards this is the proposed EU regulation, the Artificial Intelligence Act. AI is on the clinical regulation agenda too, and it’s included in the EMA’s Regulatory Science Strategy to 2025.
Given their use of data sets, AI models also need improved access to high-quality data that is clinically relevant and well annotated. And as the use of AI increases, it’s important to ensure tools are inclusive, supporting the move towards greater representation in clinical research.
AI has great potential in many areas, but it can’t replace real-world clinical research experience. At Siron Clinical, our expert CRAs each have 15 years’ experience across a range of trial types and phases. Find out how we can support your clinical trial. https://sironclinical.com/expertise/
0 Comments