Knowledge Center

Key Learnings from the FDA’s Draft Guidance on AI in Drug Development

Katalyze AI

February 3, 2025

min read

The FDA’s draft guidance on Artificial Intelligence (AI) in regulatory decision-making for drugs and biological products provides critical insights into leveraging AI while maintaining stringent standards for safety, quality, and regulatory compliance. Below are some brief takeaways from the document:

AI as a Tool for Enhancing Drug Development

AI can reduce reliance on traditional methods (e.g., animal studies) and integrate data sources to improve the understanding of diseases and outcomes.

“Reducing the number of animal-based pharmacokinetic, pharmacodynamic, and toxicologic studies;”
“Using predictive modeling for clinical pharmacokinetics and/or exposure-response analyses;”
“Integrating data from various sources (e.g., natural history, clinical studies, genetic databases, clinical trials, social media, registries) to improve understanding of disease presentations, heterogeneity, predictors of progression, recognition of disease subtypes; “
“Processing and analyzing large sets of data (e.g., data from real-world data sources or data from digital health technologies) for the development of clinical trial endpoints or assessment of outcomes;”
“Identifying, evaluating, and processing for reporting postmarketing adverse drug experience information;”
‘Facilitating the selection of manufacturing conditions”

(The reference of the 6 above-mentioned points is the FDA Draft Guidance, page 4)

AI can reduce reliance on traditional methods (e.g., animal studies) and integrate data sources to improve the understanding of diseases and outcomes.

“Reducing the number of animal-based pharmacokinetic, pharmacodynamic, and toxicologic studies;”
“Using predictive modeling for clinical pharmacokinetics and/or exposure-response analyses;”
“Integrating data from various sources (e.g., natural history, clinical studies, genetic databases, clinical trials, social media, registries) to improve understanding of disease presentations, heterogeneity, predictors of progression, recognition of disease subtypes; “
“Processing and analyzing large sets of data (e.g., data from real-world data sources or data from digital health technologies) for the development of clinical trial endpoints or assessment of outcomes;”
“Identifying, evaluating, and processing for reporting postmarketing adverse drug experience information;”
‘Facilitating the selection of manufacturing conditions”

(The reference of the 6 above-mentioned points is the FDA Draft Guidance, page 4)

AI can reduce reliance on traditional methods (e.g., animal studies) and integrate data sources to improve the understanding of diseases and outcomes.

“Reducing the number of animal-based pharmacokinetic, pharmacodynamic, and toxicologic studies;”
“Using predictive modeling for clinical pharmacokinetics and/or exposure-response analyses;”
“Integrating data from various sources (e.g., natural history, clinical studies, genetic databases, clinical trials, social media, registries) to improve understanding of disease presentations, heterogeneity, predictors of progression, recognition of disease subtypes; “
“Processing and analyzing large sets of data (e.g., data from real-world data sources or data from digital health technologies) for the development of clinical trial endpoints or assessment of outcomes;”
“Identifying, evaluating, and processing for reporting postmarketing adverse drug experience information;”
‘Facilitating the selection of manufacturing conditions”

(The reference of the 6 above-mentioned points is the FDA Draft Guidance, page 4)

AI can reduce reliance on traditional methods (e.g., animal studies) and integrate data sources to improve the understanding of diseases and outcomes.

“Reducing the number of animal-based pharmacokinetic, pharmacodynamic, and toxicologic studies;”
“Using predictive modeling for clinical pharmacokinetics and/or exposure-response analyses;”
“Integrating data from various sources (e.g., natural history, clinical studies, genetic databases, clinical trials, social media, registries) to improve understanding of disease presentations, heterogeneity, predictors of progression, recognition of disease subtypes; “
“Processing and analyzing large sets of data (e.g., data from real-world data sources or data from digital health technologies) for the development of clinical trial endpoints or assessment of outcomes;”
“Identifying, evaluating, and processing for reporting postmarketing adverse drug experience information;”
‘Facilitating the selection of manufacturing conditions”

(The reference of the 6 above-mentioned points is the FDA Draft Guidance, page 4)

A man in a lab coat and mask stands in a factory, surrounded by machinery and equipment, focused on his work.

Data Quality and Representativeness Are Key

The reliability of AI outputs hinges on the quality and representativeness of the training data, underscoring the need for diverse and accurate datasets.

Transparency and Credibility Framework

AI models must operate transparently, with a clear framework to assess their credibility and outputs. The FDA recommends a 7-step process to achieve this.

“Step 1: Define the question of interest that will be addressed by the AI model (see section IV.A.1 for details).

Step 2: Define the COU for the AI model (see section IV.A.2 for details).

Step 3: Assess the AI model risk (see section IV.A.3 for details).

Step 4: Develop a plan to establish the credibility of AI model output within the COU (see section IV.A.4 for details).

Step 5: Execute the plan (see section IV.A.5 for details).

Step 6: Document the results of the credibility assessment plan and discuss deviations from the plan (see section IV.A.6 for details).

Step 7: Determine the adequacy of the AI model for the COU (see section IV.A.7 for details). “

(The reference of the 6 above-mentioned points is the FDA Draft Guidance, page 5-6)