How to Survive FDA’s Predetermined Change Control Plan

It’s rare to hear the private sector applauding a regulatory agency, but the FDA’s Predetermined Change Control Plan (PCCP) deserves recognition. This initiative could be one of the most transformative in the agency’s history, driving innovation in medical devices while enhancing patient safety through continuous improvement.

However, a shakeout is on the horizon that will expose those companies with IT environments that were originally built on 1990s infrastructure. These IT systems never envisioned the cloud or AI, today’s volume of users, or the demands of modern applications. Over the coming years, it will become clear which manufacturers have successfully transformed their aging infrastructures with the right FDA compliance software to fully leverage the benefits of the PCCP. The difference will be immediately obvious when you compare the number of annual releases or versions on the app store. Companies that adapt will release updates frequently and incorporate AI/ML. However, most companies will struggle to make this transition, and, as a result, won’t be able to harness the PCCP. These companies will find themselves left behind by more agile manufacturers.

This article offers a primer on the PCCP and a strategy for companies to design their products with PCCP in mind, enabling them to leapfrog competitors still bound by traditional regulatory processes. PCCP is transformative because it empowers medtech companies to release safer products more rapidly, with adaptability at their core. By establishing Continuous Integration and Continuous Delivery (CI/CD) as the foundation for AI/ML-enabled devices, PCCP facilitates the development of these advanced technologies without the need for resubmission each time the model is retrained. The FDA has released an update to the draft PCCP guidance that extends beyond AI/ML applications, encompassing a broader range of medical devices and software modifications, demonstrating the agency's intent to apply PCCPs more broadly across the medical device industry.

Why the PCCP is so transformative

The PCCP empowers medical device manufacturers to implement updates and improvements without the traditional, lengthy and costly process of submitting new premarket applications. This flexibility is crucial for keeping pace with rapid technological advancements, especially in AI and ML. Given the increasing number of AI devices approved in recent years, with 106 so far this year alone, this guidance couldn’t have come at a better time.

The PCCP greatly simplifies change management for manufacturers, significantly reducing the documentation time required to navigate regulatory hurdles. By leveraging the PCCP to continuously enhance product performance, companies can make regular updates based on new data, leading to improved diagnostic accuracy and better patient outcomes.

Other enhanced outcomes include:

Connecting AI to the PCCP

Now, let’s explore how a PCCP can accelerate the development of AI/ML-powered devices. Imagine a manufacturer with an already approved and successful product, backed by several years of reliable patient data. The manufacturer wants to use this existing data along with AI/ML to develop a new version of the product that delivers more accurate and personalized patient treatments.

The first step is to analyze the data to ensure it accurately represents the target population. Next, the manufacturer should segregate the data into training and test sets, confirming the new model still works with the updated architecture or third-party input. Once the model is trained, it must meet acceptance criteria and pass an impact assessment in alignment with the PCCP.

This process ensures that the previously approved model infrastructure remains validated (does what it was intended to do) and that the model is retrained with the latest data. Manufacturers can run this process multiple times, continuously improving a product with the latest data or scientific advances, without having to resubmit to the FDA — as long as the changes are in accordance with their PCCP. While high-risk applications may require more cautious release schedules, most medical devices — especially Class I and II — can be updated weekly or even daily. Continuous improvements in these scenarios are crucial to successfully bringing hospital treatments into the home.

Designing AI/ML-powered products

One of the key reasons behind the FDA’s creation of the PCCP was to leverage AI/ML. When using AI/ML in software design, it’s important to think beyond just the software itself. Successful AI/ML integration involves not only the design of data models but also the testing of the data and training of models to ensure they meet PCCP requirements. To design AI successfully, you must focus on three key areas:

Creating a survival playbook requires a foundation built for rapid change

Many medtech companies' risk being left behind as they struggle to adapt to the PCCP.  Burdened by aging infrastructure, these companies still rely on manual documentation, testing, and approval processes.

To effectively leverage the PCCP, it’s essential to move beyond traditional compliance methods and embrace modern approaches that enable rapid change. CI/CD is at the heart of modern software development with leading companies like Google, Amazon, and Target making frequent, incremental code changes — sometimes hundreds per day — to continuously enhance their products and improve the customer experience.

The PCCP empowers medical device companies to achieve a similar pace of innovation, but manufacturers must follow three key steps to build a strong foundation for this approach:

Maintaining a successful implementation

A proven approach for maintaining successful implementation involves establishing an AI Center of Excellence, where experts in PCCP, validation, and AI work collaboratively. In this model, the AI Center of Excellence serves as the hub, working closely with various business units across the organization. These units, each with their unique applications and needs, receive support from the center to develop and apply models tailored to their specific requirements. As business units create and modify these models, they integrate them into different programs, which are then linked to a PCCP and an AI validation strategy or set of metrics.

A governance committee plays a central role in this process. They are responsible for approving the PCCP and AI validation strategies, ensuring that each program aligns with organizational standards and regulatory requirements. Over time, as the governance committee and the AI Center of Excellence continue to work together, they build expertise and efficiency in managing PCCP and AI validation. This continuous collaboration leads to more refined strategies, enabling broader deployment across business units and the successful execution of multiple AI programs.

Strong AI program governance and collaboration is necessary for organizations to successfully manage PCCP and AI validation. Graphic courtesy of Erez Kaminski.

Only the agile will survive

The PCCP marks a pivotal moment in medtech, driving rapid innovation while potentially reducing the overall costs of healthcare over time. By enabling continuous updates and improvements without lengthy regulatory delays, the PCCP ensures that medical devices evolve in real-time, leading to better patient outcomes and higher-quality care.

However, this transformation also sets the stage for a significant industry shakeout. Companies stuck with outdated infrastructures and manual processes will quickly fall behind. To survive, they must adopt CI/CD practices, design flexible systems, and integrate risk management throughout their product lifecycle.

The FDA’s PCCP isn’t just a regulatory shift, it’s a chance to innovate faster and deliver safer, more effective products. Those who adapt will thrive, those who don't will be left behind. The time to act is now.

Erez Kaminski is the Founder and CEO of Ketryx, the life sciences industry's first and only connected application lifecycle management software designed to improve quality and release software faster. Erez founded Ketryx after working at Amgen as the head of AI/ML for their medical device division. Prior to Amgen, he served as the special assistant to the CEO of Wolfram Research, the builders of Mathematica and Wolfram|Alpha, and as a visiting researcher at Princeton Plasma Physics Labs. Erez holds a Master of Science in Electrical Engineering and Computer Science and a Master of Business Administration from the Massachusetts Institute of Technology.