What is a Predetermined Change Control Plan (PCCP) and Who Needs One?

In April 2023, FDA published a new draft guidance, Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning (AI/ML)-Enabled Device Software Functions.

But if you’re short on time, here’s what you should know about the Predetermined Change Control Plan (PCCP) guidance and how FDA is thinking about machine learning algorithms in medical devices from a regulatory standpoint.

BONUS RESOURCE: Click here to download your free PDF copy of our Ultimate Guide to Software as a Medical Device (SaMD).

Why has FDA published this draft guidance on a PCCP?

Artificial intelligence (and especially machine learning) are having a bit of a moment in the spotlight right now, to put it lightly. But FDA recognizes that aside from the hype, there is enormous potential for the use of AI/ML in the medical device industry.

NOTE: Before I go any further, I want to mention that while MedTech is famous for the number of acronyms it employs, FDA has decided we need at least one more. So, you will see the acronym ML-DSF used in the draft guidance. That stands for Machine Learning-Enabled Device Software Functions. 

The reason FDA has put out this draft guidance about a Predetermined Change Control Plan is because, by definition, a machine learning algorithm is learning from the data that is put into the system. It then has the potential to change its output based on what it’s learned from those inputs. 

For instance, a diagnostic tool may get better at detecting a certain type of cancer as it sees more images of both cancerous and benign cells and is able to determine which is which. 

The catch here is that changes like this to an ML-DSF’s algorithm that result in iterative improvements are technically changes to the device. And traditionally, a change like that could result in the manufacturer being required to submit these changes as a Special 510(k), De Novo submission, or PMA supplement, depending on the device’s regulatory pathway.

To avoid a situation where manufacturers are constantly re-submitting to FDA as their algorithm learns, the agency is saying that MedTech companies may instead use a Predetermined Change Control Plan.

What is a Predetermined Change Control Plan (PCCP)?

In short, a Predetermined Change Control Plan is a document that describes what modifications will be made to software using machine learning and how you’ll assess those modifications. You’ll submit it as a standalone section within your marketing submission, whether that’s the 510(k), De Novo, or PMA pathway. 

The PCCP essentially allows you to set out the boundaries of any modifications to the ML-DSF due to the machine learning component. By predetermining these limits, you won’t have to submit a new marketing submission when those modifications occur (as long as they stay within the limits!) 

I recently had FDA regulatory expert and long-time guest Mike Drues on the Global Medical Device Podcast, and he described it as “pre-validating” the changes (Mike does a great job of contextualizing the PCCP in relation to medical device regulations, so I’d encourage you to give that episode a listen).

However you want to think about it, the PCCP gives makers of devices with ML-DSF a way to avoid onerous and unnecessary regulatory hurdles in getting their devices to market and keeping them there. 

As FDA puts it, “This draft guidance describes an approach that would often be least burdensome and would support the ability to modify an ML-DSF while continuing to provide a reasonable assurance of safety and effectiveness across relevant patient populations.”

What are the components of a PCCP?

The PCCP consists of three major components: 

1. Description of Modifications

2. Modification Protocol

3. Impact Assessment

I’m going to describe each component briefly here, but you should definitely take some time to read through them all carefully.

1. Description of modifications

As the title suggests, this section is for describing any planned modifications to your ML-DSF. However, you’ll also need to include a detailed description of any changes to the device characteristics and performance that would result from implementing those modifications. 

FDA stresses here that you should limit the number of modifications (for efficient review) and make sure that any you list are specific and can be verified and validated within your existing quality system. Other questions that you should address here include:

• Will these modifications be implemented automatically by software? Or will the modification be done manually, requiring human action and input?

• Will these modifications be implemented uniformly across every device on the market? Or will modifications happen locally, based on characteristics unique to a specific clinical site or patient?

Also, keep in mind that FDA clearly states “All modifications included in a PCCP must maintain the device within the device’s intended use. At this time, FDA expects that modifications included in a PCCP should also maintain the device within the device’s indications for use.”

2. Modification Protocol

The Modification Protocol will need to include a description of the methods for developing, validating, and implementing the modifications in your Description of Modifications section. 

It will include the verification and validation activities, “and is intended to provide a step-by-step delineation of how the modifications proposed in the PCCP will be implemented while assuring the device remains safe and effective.”

The draft guidance includes four primary components of a Modification Protocol that manufacturers will need to address:

• Data management practices

• Re-training practices

• Performance evaluation

• Update procedures

3. Impact Assessment

Your impact assessment is defined as the assessment of the benefits and risks of implementing a PCCP for your ML-DSF, as well as mitigations of those risks. 

Again, here FDA points to the importance of a great QMS, noting that “the manufacturer’s existing quality system should be used as the framework in which to conduct an Impact Assessment for the modifications set forth in the PCCP.”

Here, FDA lists five items that the Impact Assessment should discuss:

1. A comparison of the device with each modification implemented to the original version of the device

2. The benefits and risks, including risks of social harm, of each modification

3. How the activities in the Modification Protocol will continue to reasonably ensure safety and effectiveness of the device

4. How the implementation of one modification will affect the implementation of another

5. The collective effect of implementing all the modifications

While this is still a draft guidance, it does represent the agency’s current thinking on how to approach devices that include machine learning components. I’ll update this when the final draft guidance comes out. But for now, this is a great document to familiarize yourself with.

BONUS RESOURCE: Click here to download your free PDF copy of our Ultimate Guide to Software as a Medical Device (SaMD).

Choose a QMS solution that’s purpose-built for MedTech

There’s a lot of change and uncertainty going on right now when it comes to the use of AI/ML in MedTech. But what hasn’t changed (and won’t be anytime soon) is the regulatory requirements around your QMS. 

Whether you’re working on SaMD with ML-DSF or just a good old fashioned catheter, the best way to stay on the right side of regulations is to use a QMS that’s built specifically and exclusively for the MedTech industry. 

With Greenlight Guru’s QMS software, you get a scalable quality management system with end-to-end traceability, helping you stay audit ready at all times. Our QMS comes pre-validated per FDA best practices, so you can ensure compliance with regulatory requirements from the very start. 

If you want to see how a purpose-built QMS can help you deliver high quality medical devices to patients, then get your free demo of Greenlight Guru today ➔