Glossary

Predetermined Change Control Plan (PCCP)

How the FDA pre-authorizes future updates to AI-enabled medical devices — what a PCCP contains, where the idea came from, and how many devices actually carry one. As of July 2026.

By Jonas WeirReviewed by Jonas Weir · editorial reviewUpdated

The short version

  • A predetermined change control plan (PCCP) is FDA-reviewed documentation describing what modifications will be made to a device and how each will be assessed — so pre-specified updates can ship without a new marketing submission.
  • The FDA finalized its PCCP guidance for AI-enabled device software functions on December 4, 2024; the current edition is dated August 2025.
  • A PCCP has three parts: a description of the planned modifications, a modification protocol for developing and validating them, and an impact assessment.
  • Uptake is real but early: through May 30, 2025, 26 AI/ML-enabled devices carried authorized PCCPs — 92% cleared via 510(k), with model retraining the most common authorized change (69% of devices).

A predetermined change control plan (PCCP) is documentation, reviewed by the FDA within a device's marketing submission, that describes what modifications will be made to the device and how each modification will be assessed — so that pre-specified updates can be implemented without a new submission for every change 1.

Why it matters in healthcare

AI-enabled devices are built to be updated: retrained on new data, tuned for new scanners, extended to new inputs. Under the traditional device framework, many of those updates would each trigger a fresh regulatory submission — months of delay per improvement, multiplied across a product's life. The PCCP is the FDA's structural answer. The agency reviews the plan once, as part of the marketing submission, "to ensure the continued safety and effectiveness of the device without necessitating additional marketing submissions for implementing each modification described in the PCCP" 3. For health systems buying AI, a device's PCCP is the document that says how the product they validated today is allowed to differ a year from now.

How it works in practice

The mechanism gained a statutory footing when the Food and Drug Omnibus Reform Act of 2022 added section 515C to the FD&C Act, and the FDA then built out guidance 1. The AI-specific final guidance was announced on December 4, 2024 2; the current edition on the FDA's site is dated August 2025 3.

The guidance recommends three components 3:

  1. Description of modifications — the specific, planned changes (for example, retraining on new data or adding a compatible input source).
  2. Modification protocol — the methods that will develop, validate, and implement each change, with acceptance criteria set in advance.
  3. Impact assessment — an analysis of the benefits and risks each change introduces, individually and in combination.

Once authorized, the plan travels with the device: the FDA reviews it inside the 510(k), De Novo, or PMA submission, and later modifications that follow the protocol and meet its acceptance criteria can be deployed on the plan's authority alone 3.

The approach is internationally coordinated: the FDA, Health Canada, and the UK's MHRA jointly identified five guiding principles for PCCPs for machine-learning-enabled devices in October 2023 5.

Where it appears today

Uptake is measurable and still early. A cross-sectional analysis of the FDA's public list found 26 AI/ML-enabled devices with authorized PCCPs through May 30, 2025 — 92% cleared via 510(k), all classified as moderate risk, with model retraining the most commonly authorized modification (69% of devices), ahead of logic updates (42%) and expansion of input sources (35%) 4. The same analysis flagged the soft spots: preapproval testing was limited, with seven devices prospectively evaluated; subgroup analyses were reported for eleven devices and none included patient outcomes data; and user manuals could be found online for 54% of devices, many silent on the PCCP itself 4. The FDA's own AI-enabled device list now flags such authorizations directly, with entries like "Fibresolve (with PCCP)" appearing among radiology clearances as of its late-2025 updates 6.

Common misunderstandings

A PCCP permits open-ended change. It authorizes only what is written in the plan. Modifications outside the plan — above all, changes to intended use — still require a new submission 3.

A PCCP means the model learns continuously in the wild. The authorized plans observed so far cover controlled, batched changes such as periodic retraining with pre-set validation gates 4, a discipline that exists because unmanaged updates invite model drift.

A PCCP replaces monitoring. It presumes monitoring. The modification protocol's acceptance criteria only mean something if the manufacturer and deploying health system watch real-world performance — the work of algorithmovigilance.

Related terms

See Software as a Medical Device for the regulatory category PCCPs attach to, model drift for the failure mode planned updates manage, and algorithmovigilance for the post-deployment surveillance that makes pre-authorized change safe.

Questions & answers

  • Does a PCCP let a manufacturer change an AI device however it wants?

    No. A PCCP authorizes only the specific modifications written into it, each with a pre-agreed validation method and impact assessment. Changes outside the plan — including changes to the device's intended use — still require a new marketing submission.

  • How many devices actually have PCCPs?

    Through May 30, 2025, an analysis of the FDA's public device list identified 26 AI/ML-enabled devices with authorized PCCPs. All were moderate-risk devices, 92% cleared through the 510(k) pathway, and the most commonly authorized modification was model retraining.

Sources

  1. Federal Register. Predetermined Change Control Plans for Medical Devices; Draft Guidance for Industry and Food and Drug Administration Staff; Availability. August 22, 2024. www.federalregister.gov/documents/2024/08/22/2024-18828/predetermined-change-control-plans-for-medical-devices-draft-guidance-for-industry-and-food-and-drug
  2. Federal Register. Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions; Guidance; Availability. December 4, 2024. www.federalregister.gov/documents/2024/12/04/2024-28361/marketing-submission-recommendations-for-a-predetermined-change-control-plan-for-artificial
  3. US Food and Drug Administration. Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions. Guidance for Industry and FDA Staff, August 2025 edition. www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial-intelligence
  4. Regulating Flexibility for Artificial Intelligence: FDA Experience with Predetermined Change Control Plans. medRxiv preprint. 2025. doi.org/10.1101/2025.08.26.25334477
  5. US Food and Drug Administration, Health Canada, Medicines and Healthcare products Regulatory Agency. Predetermined Change Control Plans for Machine Learning-Enabled Medical Devices: Guiding Principles. October 2023. www.fda.gov/medical-devices/software-medical-device-samd/predetermined-change-control-plans-machine-learning-enabled-medical-devices-guiding-principles
  6. US Food and Drug Administration. Artificial Intelligence-Enabled Medical Devices (AI-Enabled Medical Device List). www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-enabled-medical-devices