The Danish Medicines Agency (DKMA) is working to define criteria for the application of artificial intelligence (AI) and machine learning (ML) across GxP-regulated areas.
To understand industry’s position, it has been gathering information using questions posed as though asked by an inspector assessing the quality of an AI/ML algorithm to control and evaluate product quality.
To form an industry response, BioPhorum gathered a cross-Phorum collaboration under the leadership of its Regulatory Governance Team across a series of meetings.
The response has now been published as Industry Feedback on DMKA Questions to Critical GxP AI-ML Applications, and submitted to the DKMA.
The questions considered the requirements and quality of data used to build, test and validate an algorithm, and how it would respond to biases and deviations in results. This situation may soon arise, especially with the introduction of new platform technologies (e.g. continuous manufacturing) that will increasingly use these algorithms.
“Any effort to align within industry and with regulatory agencies is key to enabling and fast-tracking the adoption of AI/ML,” said David Lauri Pla, Senior Data Scientist/AI at Pfizer. “We all agree on the benefits of these technologies and are internally working on how to leverage them more, so this type of forum that brings us together to address agencies’ questions is key for the industry.”
The team comprised senior data scientists, statisticians, automation and digitalization developers, and regulatory experts. The meetings were a great learning experience for all involved.
The response is a critical step in asserting BioPhorum’s position as the partner of choice for the Danish and other European agencies for the global implementation of AI and ML in the manufacture of biologics.