Artificial Intelligence in Clinical Trials: The Current State of Global Regulatory Frameworks

---

Contributing Expert: Tina Caruana, Director of eClinical Solutions

Artificial intelligence (AI) has enormous potential to revolutionize clinical trial research. But first, the industry must agree on how to regulate its use. Industry bodies, federal agencies, lawmakers, and even investors are weighing in on how to regulate AI within clinical research. 

The stakes are high for effectively regulating AI in clinical trials. If done well, AI could unlock new drug discoveries at a previously unimagined pace. 

However, if done poorly, applying AI in clinical trials could introduce unprecedented risk. Therefore, regulators must find a way to mitigate risk while supporting innovation and progress. 

In this article, we’ll look at the current state of published regulatory frameworks as of November 2024. 

We will cover the following:

• Evolution of AI guidelines in clinical research
• AI regulatory frameworks by country

Please note this article is not comprehensive but showcases a selection of documents available as of November 2024. 

Evolving Artificial Intelligence Regulations in Clinical Research

As regulatory bodies race to keep up with AI development, many have developed stringent requirements to prioritize patient safety and ethical conduct in clinical trials.

In general, global regulatory guidelines emphasize data:

• Integrity
• Accuracy
• Transparency 

AI is increasingly integrated in other areas under regulatory bodies’ jurisdiction, such as Digital Health Technologies (DHTs) and Real-World Data (RWD) analytics. Therefore, it is crucial for clinical researchers to stay informed of all relevant angles as they contemplate incorporating AI into a trial.

Consider the current state of:

• USA’s AI regulatory guidance for clinical trials
• EU and UK’s AI regulatory guidance for clinical trials
• Canadian AI regulatory guidance for clinical trials
• China’s AI regulatory guidance for clinical trials 

Visit our blogs about Digital Health Technologies (DHTs) and Real-World Data (RWD) to learn more about each topic.

USA’s AI Regulatory Guidance for Clinical Trials

The United States Food and Drug Administration (FDA) has taken a flexible, risk-based approach to regulating AI within clinical research. 

On February 7, 2020, the FDA announced its approval of the first cardiac ultrasound software, which uses artificial intelligence to guide users. Since then, the FDA “has accelerated its efforts to create an agile regulatory ecosystem that can facilitate innovation while safeguarding public health.” 

As part of its efforts, the agency has developed a glossary of relevant digital health and AI terms. 

US drug development: AI/ML guidance

The FDA “recognizes the increased use of AI/ML throughout the drug development life cycle and across a range of therapeutic areas.” The FDA also reports a significant increase in drug and biologic application submissions using AI/ML components in recent years.

Current AI/ML drug development documentation includes:

• AI/ML for Drug Development Discussion Paper | May 2023 | Discussion paper
• CDER Framework for Regulatory Advanced Manufacturing Evaluation (FRAME) Initiative, including discussion papers:
  • Distributed Manufacturing and Point-of-Care Manufacturing of Drugs | October 2022 | Discussion paper
  • Artificial Intelligence in Drug Manufacturing | 2023 | Discussion paper

US medical devices: AI/ML guidance 

The FDA states, “The complex and dynamic processes involved in the development, deployment, use, and maintenance of AI technologies benefit from careful management throughout the medical product life cycle.” With this in mind, they have developed several documents, including draft guidance. 

Current AI/ML medical device documentation includes:

• Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) | April 2019 | Discussion paper
• Artificial Intelligence and Machine Learning Software as a Medical Device Action Plan | January 2021 | Action plan
• Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions | April 2023 | Draft guidance
• Artificial Intelligence and Medical Products: How CBER, CDER, CDRH, and OCP are Working Together | March 2024 | Paper

Of note, the FDA also has several comprehensive web pages dedicated to Software as a Medical Device (SaMD).

European Union and United Kingdom AI Regulatory Guidance in Clinical Trials 

There is no current legislation specifically about the use of AI in clinical trials in the European Union (EU) or the United Kingdom (UK). There are, however, several factors sponsors need to consider. 

EU guidelines: AI/ML

Currently, any AI used within a clinical trial in the EU needs to comply with the EU AI Act adopted by the European Parliament in 2023.  The AI Act was established to comprehensively regulate AI systems across industries in the EU.

The AI Act takes a risk-proportionate approach with four levels of AI systems, ranging from minimal to unacceptable. Of note, medical devices are classed as “high risk.” This is the highest level of acceptable risk and may be subject to a more stringent set of requirements.

While the use of AI in clinical research is not yet specifically governed by the European Medicines Agency (EMA), they have begun to weigh in.

The European Medicines Agency (EMA) published:

• European Medicines Agency: Reflection paper on the use of Artificial Intelligence (AI) in the medicinal product lifecycle | July 2023 | Draft reflection paper
• Artificial Intelligence Workplan | December 2023 | Workplan

UK guidelines: AI/ML

In 2023, the UK government published a policy paper on regulating AI, taking a pro-innovation stance on regulating AI. This approach leaves the Medicines and Healthcare products Regulatory Agency (MHRA) discretion over how the principles apply to clinical research. 

Current AI/ML documentation:

• Software and artificial intelligence (AI) as a medical device | June 2024 | Guidance document 
• Impact of AI on the regulation of medical products | April 2024 | Policy paper 

Canadian AI Regulatory Guidance for Clinical Trials

So far, Health Canada has only addressed AI as it pertains to devices using machine learning. The agency has opted to take a total product lifecycle approach. A final guidance document is expected to be released soon. 

• Pre-market guidance for machine learning-enabled medical devices | September 2023 | Draft guidance

China’s AI Regulatory Guidance for Clinical Trials 

China is a major producer and consumer of medical devices and a hub for much healthcare–related software development. Therefore, authorities recognized the need for comprehensive regulatory guidance, particularly for international manufacturers. 

The National Medical Products Administration (NMPA), China’s regulatory body for clinical research, generally takes “a more cautious approach towards AI-empowered medical devices” than the FDA. 

• Guideline for Artificial Intelligence Medical Devices | June 4, 2022
• Key Points on Clinical Evaluation of AI-assisted Detection Medical Devices (Software)
• Key Points on Review of Imaging Ultrasound AI Software (Process Optimization)
• Key Points on Performance Evaluation Pathological Image AI Analysis Software
• Key Points on Clinical Evaluation of Pathological Image AI Analysis Software
• Key Points on Performance Evaluation of AI Analysis Software for Blood Disease Flow Cytometry
• Key Points on Review of Magnetic Resonance Imaging System AI Software Functions

According to a 2024 Nature article, “The NMPA review guideline places special emphasis on an evidence-based approach relying on data, highlighting the significance of data sufficiency, diversity and representativeness, and addressing data bias in AI algorithm development.”

Joint Efforts to Develop Guidance for the Use of AI in Research

As regulatory bodies strive to make progress in regulating AI, some groups are combining forces. 

In 2020, SPIRIT-AI and CONSORT-AI were created as part of an international collaborative effort to improve the transparency and completeness of clinical trials evaluating interventions involving AI. 

These documents are extensions of the existing SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) and CONSORT (Consolidated Standards of Reporting Trials). Both documents sought to provide minimum guidelines for protocols and reporting for randomized trials, respectively. 

Meanwhile, the FDA, Health Canada, and the MHRA have teamed up to develop guidance documents, including:

• Ten International Guiding Principles on Good Machine Learning in Medical Devices | October 2021 | Guidance document
• Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles | October 2023 | Guidance document
• Transparency for Machine Learning-Enabled Medical Devices: Guiding Principles | June 2024 | Guidance document

Into the Future of AI

The world of AI is moving fast. The intersection of large language models (LLMs), machine learning (ML), and evolving algorithms presents thrilling new possibilities within clinical research. 

While some may argue that regulating it is too difficult or may even stifle innovation, the clinical research industry must find a way forward that includes AI in clinical trials. 

In the next few years, the clinical trial industry will be in dire need of additional guidance documents to navigate the new world of AI. We expect to see a surge in the release of such documents to aid researchers in this evolving landscape. 

Medrio’s experts have decades of experience navigating the changing clinical trial environment amidst evolving regulatory guidelines. Our experts are equipped with innovative and creative solutions to help you keep pace. Connect with us at medrio.com/contact-us/.