[August 2023 Issue] AI and Machine Learning: FDA Spells Out Specifics for Manufacturing, Another Perspective on PCCPs for AI/ML Devices + Warning Letter Breakdown

Welcome to Insider issue #16.

And a very warm welcome to the 187 new life science professionals who joined last month. If you haven’t already graduated to a paid subscription to unlock issues like this one, you can do so here.

This month, for our pharma crowd, we examine an FDA paper that begins to look at how to apply machine learning models in the pharmaceutical manufacturing space, expanding on it with some considerations if your team decides to pursue their application itself.

In our device feature, we look again at the FDA's latest draft guidance for Predetermined Change Control Plans for Artificial Intelligence/Machine Learning-Enabled Device Software (AI/ML). As the healthcare landscape becomes increasingly digital and machine learning continues finding its way into all kinds of tech, managing changes in devices and software is becoming more complex. We discuss the implications of this development and provide a detailed breakdown of the PCCP as part of the device submission—an area many device and digital health firms will no doubt need to pay attention to soon if they aren’t already.

Lastly, our warning letter breakdown covers a multitude of good manufacturing practice (GMP) violations tied to the production of sterile eye drops, including inadequate contamination safeguards and lax process controls. We examine the specifics of this case and provide key takeaways that manufacturers can learn from to avoid similar compliance pitfalls.