Photo By Paul Lagasse | A laboratory technician at the Center for Military Psychiatry and Neuroscience at the...... read more read more
FORT DETRICK, Md. – As drug developers rely increasingly on artificial intelligence to assist with the discovery, testing and manufacture of new medicines, researchers at the Defense Health Agency Research and Development-Medical Research and Development Command are learning how to apply the latest regulatory guidance related to AI to better carry out their mission to develop life-saving medicines that protect our Warfighters and enable them to maintain operational effectiveness even in chemically and biologically contaminated environments.
“The use of AI in the drug product life cycle continues to increase,” says Dr. Barima Frempong, a regulatory affairs analyst at the Defense Health Agency Research and Development-Medical Research and Development Command’s Office of Regulated Activities. “AI is transforming how drug development is conducted by accelerating discovery, improving accuracy and reducing costs.”
ORA plays a vital role in each of DHA R&D-MRDC’s strategic lines of effort to tightly integrate medical research and development with the DOD’s Warfighter requirements. The office’s regulatory affairs, compliance and clinical support professionals specialize in identifying and mitigating regulatory risks early in a product’s development life cycle, which helps them pass more smoothly – and quickly – through the U.S. Food and Drug Administration’s rigorous review process and get approved for licensing and manufacture.
Until recently, AI has mostly been used to accelerate the process of discovering promising new drug candidates, thanks to its ability to process vast amounts of data quickly and accurately. However, drug developers are increasingly using AI to assist with other data-intensive tasks such as assisting with patient diagnosis, improving the fidelity of augmented reality and imaging systems, and informing treatment decisions. To help provide meaningful guidance for these and other emerging uses of AI, the FDA is developing regulatory frameworks and policies to ensure AI is used safely and ethically.
In January, the FDA issued a new draft guidance document, “Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products,” which lays out a seven-step method for assessing an AI model’s ability to evaluate the safety, effectiveness, and quality of a new drug. The new approach lays out a risk-based assessment process that determines how an AI model will impact the regulatory decision-making and documentation requirements associated with a particular project.
Valeria Sena-Weltin, a regulatory affairs scientist at ORA, uses the analogy of a smart watch to describe how the risk-based credibility assessment works.
“The regulations that apply to a smart watch that can track your heart rate will be different than for a watch that can detect atrial fibrillation, because that watch would be considered a medical device,” says Sena-Weltin. “The more power you give to an AI model for decision making, the greater the documentation requirement will be to support its safe use.”
DHA R&D-MRDC researchers who develop vaccines, therapeutics, and devices for detecting and combating infectious diseases, biological agents and chemical weapons are increasingly incorporating AI into their new products. For example, the Automated Processing of the Physiological Registry for Assessment of Injury Severity Hemorrhage Risk Index, which uses vital-sign data from trauma patients to assess their risk of hemorrhage, and SeptiBurnAlert, which identifies whether burn patients are at risk of experiencing life-threatening complications from sepsis, both use AI to help clinicians and first responders make timely and life-saving health care decisions.
As the use of AI in military medicine R&D increases, researchers need to take precautions to ensure the resulting products are safe and effective, says Sena-Weltin. To avoid the introduction of bias that could skew the accuracy of a readout, researchers need to ensure the data sets used to train the algorithm are reliable, accurate, and complete. Because AI relies on complex computational and statistical methodologies, researchers must provide the FDA with detailed and transparent explanations of the methods they use to train the algorithm, so that they can be independently verified. And to ensure the data sets remain trustworthy over time, they should be monitored and evaluated regularly throughout its life cycle, especially after they are put into use and new data are added over time.
Frempong and Sena-Weltin discussed the use of AI to support regulatory decision-making for drug and biological products during a recent presentation to DHA R&D-MRDC staff. The presentation, which was a follow-up to ORA’s previous session on the use of AI in drug development, was part of a series of bimonthly “lunch and learn” sessions conducted by ORA for MRDC staff as a way to share the latest regulatory guidelines from the FDA related to drugs, biologics, and medical devices, and how they apply to medical research and development efforts at MRDC.
| Date Taken: | 04.23.2025 |
| Date Posted: | 04.23.2025 15:39 |
| Story ID: | 496015 |
| Location: | FORT DETRICK, MARYLAND, US |
| Web Views: | 111 |
| Downloads: | 0 |
This work, New FDA Guidance on Artificial Intelligence Will Help MRDC Field Better Drugs More Quickly, by Paul Lagasse, identified by DVIDS, must comply with the restrictions shown on https://www.dvidshub.net/about/copyright.
No keywords found.
