{"slug": "pccps-best-practices-faqs-and-examples", "title": "PCCPs: Best Practices, FAQs, and Examples", "summary": "The U.S. Food and Drug Administration (FDA) has introduced Predetermined Change Control Plans (PCCPs) as a regulatory tool allowing companies to pre-clear certain device changes within a 510(k), De Novo, or PMA submission, potentially avoiding additional 60-120 day review times. While initially developed for AI/ML model updates, PCCPs have been used broadly, including for non-AI devices and hardware-only devices, and are expected to become increasingly important following Congress's grant of statutory authority. Innolitics notes that PCCPs cannot be added after an original submission and that manufacturers are not obligated to implement pre-cleared changes.", "body_md": "PCCPs, or Predetermined Change Control Plans, are a new regulatory tool that allow companies to “pre-clear” certain changes within a 510(k), De Novo, or PMA. While the unique regulatory challenges related to AI are what spurred FDA to develop this innovative approach, PCCPs can be used much more broadly. Here at Innolitics, we expect that PCCPs will become an increasingly important over time. The surge in their use since Congress provided FDA this statutory authority demonstrates this.\n\nA Predetermined Change Control Plan, or PCCP, describes planned changes that may be made to a device that would otherwise require another FDA submission. A PCCP for a device is reviewed as part of a 510(k), De Novo, or PMA.\n\nPCCP plans may allow you to avoid time-consuming and expensive regulatory submissions by pre-clearing your changes with FDA. If you have a PCCP plan you will have to do most of the work required for a 510(k), but you won’t need to wait the additional 60 - 120 days for the FDA to complete their review.\n\nThe focus has been to use PCCPs to allow updates to AI/ML models, however, the underlying laws and regulations are quite broad. In fact, they have been used in several devices without AI/ML and even hardware only devices already. (See the examples section below for some creative uses of PCCPs.)\n\nHere at Innolitics we’re still exploring possible use cases for PCCPs. We believe they may be used in a number of ways, including:\n\nThat said, you can’t use PCCPs for several types of changes, including\n\nA major use of PCCPs is to make it easier to update AI/ML functions, however, PCCPs have already been used for a number of other situations:\n\nSee the examples section below for in depth explorations of some of these.\n\nYes. PCCPs have been used in De Novos classification requests already.\n\nFor example, see [DEN220063](https://www.accessdata.fda.gov/cdrh_docs/pdf22/DEN220063.pdf) which established the QVD product code “\n\nSee the examples section for a particular De Novo and product code that was established.\n\nNo. Based on our experience with FDA, it has been required that you include the PCCP in your original submission. You can’t add a PCCP after you’ve already submitted a 510(k), for example.\n\nIn the early days of PCCP plans, we had FDA suggest to us that devices with a PCCP would not be considered substantially equivalent to devices without a PCCP. As a result, new types of devices with PCCPs would require a De Novo classification request to create a new product code.\n\nHowever, it seems clear this is not the case as there are several cleared 510(k)s for devices with PCCPs whose predicates did not have PCCPs.\n\nNo. The manufacturer is not obligated to make any changes described within a PCCP.\n\nNo.\n\nFDA may catch the inappropriate changes in an audit.\n\nFDA also will review marketing materials from your website and if they see something inappropriate they will ask to see your change control documentation for the changes.\n\nIf FDA deems the changes were inappropriate, they may recall the updated version of the software and issue a warning.\n\nYes. However, an FDA reviewer suggested to keep the changes split apart so that it would be easy for FDA to selectively reject certain changes but not others.\n\nAlso, “FDA recommends that a PCCP include only a limited number of modifications that are specific, and that can be verified and validated.”\n\nYes, one of the use-cases of a PCCP is to allow for site-specific tuning.\n\nHere’s a PCCP that was cleared for site.\n\nAlso, see [this example](https://innolitics.com/articles/fda-guidance-2024-ai-pccp/#3-ventilator-settings-software) from the FDA guidance regarding ventilator settings. the example from the FDA guidance:\n\n**Background:**\n\nThe device is an AI-DSF intended for use in the healthcare or home-use setting. The AI-DSF recommends the ideal ventilation parameters based on input data interpretation, which can then be programmed into the ventilator by a healthcare provider. The manufacturer proposes modifications to the AI-DSF to improve performance within the original indications. The AI- enabled medical device was authorized with a PCCP.\n\n**Brief Overview of Pre-Specified Modification:**\n\nThe manufacturer would like to re-train the AI model to optimize site-specific performance for a specific subset of patients with a particular condition, for whom sufficient data were not previously available. Specifically, the manufacturer would like to modify the AI model to improve its ability to optimize ventilator settings for minute volume and tidal volume to reduce the variability to ±x% within the specified range to improve treatment outcomes for that subset of patients at different sites.\n\n**Post-Authorization Modification Scenarios:**\n\n*Modification Scenario 1: Modification related to the device’s use and performance in a subset of the patient population, which was specified in the PCCP and implemented in accordance with the PCCP*\n\nThe manufacturer re-trained and re-validated the AI model on newly acquired data in a subpopulation of patients with a particular disorder. As evidenced by additional clinical performance data collected and analyzed per the Modification Protocol, the re-training on new data improved the reliability and precision of ventilator setting recommendations, showing improvements and specializations to improve site-specific ventilator operation. The updated recommendations were validated against patient outcomes and adverse events that may occur due to ventilator setting inaccuracies following the methods in the Modification Protocol. The adverse event rates and outcomes acceptance criteria were established in the Modification Protocol, and as such, were used to validate the updated AI model. The AI-DSF was updated to implement the re-trained AI model and the labeling was updated for clarity to inform users how the updated AI model accounts for local experience and prevalence. The implementation of this modification was done only at applicable sites. Because the device modification was specified in the PCCP, and it was implemented in conformance with the PCCP, the device modification would not require a new marketing submission. The manufacturer should document the modification that was specified in the PCCP in accordance with their quality system.\n\n*Modification Scenario 2: Modification related to device’s use and performance in a subset of the patient population, which was specified in the PCCP, but was not implemented in conformance with the PCCP*\n\nThe manufacturer re-trained and re-validated the AI model on newly acquired data, but was unable to fulfill the protocol because the manufacturer had to implement a reference standard that was different from the one described in the Modification Protocol. Even though the modification was specified in the PCCP, it was not implemented in conformance with the PCCP. Because this modification that was not implemented in conformance with the PCCP could significantly affect the safety or effectiveness of the device, a new marketing submission would be required.\n\nA PCCP should contain the following sections:\n\nYes. The labeling should note the PCCP and its scope.\n\nThey have been available since December 2022, so they are relatively new.\n\nFDA is always more cautious with new technologies and regulatory approaches, so expect the admin overhead to be higher until more guidance has been published.\n\nIn September 2022 FDA published their final report for their Pre-Cert Pilot program. In this report they state (emphasis mine):\n\n**The faster cycles of innovation and the speed of change for medical device software would benefit from a new regulatory approach…**\n\nUltimately, the approach to regulating novel, swiftly-evolving medical device software must foster, not inhibit, innovation, while continuing to provide reasonable assurance of safety and effectiveness. These aspects are not mutually exclusive. A flexible, risk-based approach to regulation could allow FDA to tailor regulatory requirements more efficiently for devices based on the latest science, the benefits and risks posed by devices, their real-world performance, and their contribution to promoting health equity. **It could leverage the capabilities of evolving medical device software so that health care providers, patients, and users can benefit from advancement and innovation, and so that risk for such devices can be reduced through swift software and cybersecurity updates throughout the TPLC, when needed.** **New legislative authority establishing such an approach could be supplemental to, and not replace, the established regulatory pathways.**\n\nFDA felt that the older regulatory approach with 510(k)s was not fast enough to allow for software innovation. Congress gave FDA authority to administer PCCPs in December 2022.\n\nFDA reviews a lot of marketing submissions. They would like to spend their time focused on the high-risk, new technologies that require their detailed oversight. This is the same reason they set up the third-party review program.\n\nThe QVD product code is (as best we are aware) the first product code created specifically for PCCP plans. It falls under the 892.2055 regulation and has the following description:\n\nA radiological machine learning quantitative imaging software with predetermined change control plan is a software-only device which employs machine learning algorithms on radiological images to provide quantitative imaging outputs. The device includes functions to support outputs such as view selection, segmentation and landmarking. The design specifications include planned modifications that may be made to the device consistent with an established predetermined change control plan.\n\nAs is often the case, a lot of the juicy details are found in the De Novo submission that established the product code. Here are the special controls required from the De Novo. I suspect new product codes for other types of PCCP plans will contain similar controls.\n\nIn addition, this is a prescription device and must comply with 21 CFR 801.109.\n\nNatural Cycles is Software as a Medical Device (SaMD), intended for women 18 years and older, to monitor their fertility. Natural Cycles can be used for preventing a pregnancy (contraception) or planning a pregnancy (conception). It was founded by particle physicist Elina Berglund (who was on the team that discovered the Higgs Boson).\n\nIn 2017, Natural Cycles submitted a De Novo classification request and created the [PYT product code](https://fda.innolitics.com/submissions/OB/subpart-f—obstetrical-and-gynecological-therapeutic-devices/PYT) for “Device, Fertility Diagnostic, Contraceptive, Software Application”.\n\nIn 2020, they submitted a 510(k) to add support for the Ōura smart ring. As part of the submission they did a clinical study with 40 women and compared the effectiveness of their algorithm with the original oral thermometer with the Ōura ring.\n\nIn 2023, they submitted another 510(k) for using the Apple Watch as a temperature source.\n\nIn 2024, they submitted a 510(k) for their device (K241006) which did not include any changes to their software, but only included the proposed PCCP. The PCCP is meant to obtain authorization from FDA to add more wearables providing temperature as an input to the Natural Cycles fertility algorithm in the future without requiring additional 510(k)s.\n\nThis seems like a great use case for a PCCP. The company had done two studies already and understood how to evaluate adding new temperature-measurement devices. Furthermore, it was not possible to add support for new devices using a letter-to-file since the new devices do introduce new questions of safety and effectiveness. Furthermore, it is clear they were able to define very precise acceptance criteria and KPIs for use when evaluating new hardware devices like the Apple Watch and Ōura smart ring:\n\nHere is a list of KPIs that are used to evaluate the new hardware devices:\n\nKey Performance Indicators (KPIs):\n\nSpecifications\n\n| KPI | Specification |\n|---|---|\n| Std | CI upper bound ≤ 0.234 |\n| Ratio | CI lower bound ≥ 2.04 |\n| Ovulation detection | CI lower bound ≥ 85.5% |\n| Ovulation resolution | CI lower bound ≥ 76.1% |\n| PPA | CI lower bound ≥ 96.5% |\n| Fraction of green days | No more than 2 additional red days per cycle compared to the oral thermometer |\n\n23andMe does genetic testing for health and ancestry purposes. They have an interesting FDA history that reaches back over a decade. More recently, they provide an example of using a PCCP for a non AI/ML purpose.\n\nIn 2013, 23andMe received a [warning letter](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4330248/) from the US Food and Drug Administration (FDA) to discontinue marketing its health-related genetic tests in the United States.\n\nIn October 2015, after two years of work with FDA, they were granted a De Novo ([DEN140044](https://www.accessdata.fda.gov/cdrh_docs/pdf14/den140044.pdf)) for “the detection of the BLMAsh variant in the BLM gene from saliva collected using an FDA cleared collection device (Oragene DX model OGD-500.001).”\n\nIn April, 2017, they were granted authorization for a genetic health risk report for 10 diseases and conditions.\n\nIn March, 2018, 23andMe received the first-ever FDA authorization for a direct-to-consumer genetic test for cancer risk. The authorization allows 23andMe to provide customers, without a prescription, information **on three genetic variants** found on the BRCA1 and BRCA2 genes known to be associated with higher risk for breast, ovarian and prostate cancer.\n\nIn August, 2023, they were cleared for **41 more BRCA1 and BRCA2 genetic variants and they established a PCCP for reporting additional variants and associated cancer risk information to the report**.\n\nDetails about their PCCP plan may be found both in their 510(k) letter and also in their labeling:\n\nThe PCCP plan included specific test methods for clinical and analytical validation to\n\nestablish substantial equivalence relative to DEN170046, and include sample size determination, analysis methods, and acceptance criteria. The Sponsor will perform testing of the additional BRCA1/2 variants according to the specified protocols, and if the validation data meet the specified acceptance criteria, they may add those variants to the BRCA1/2 report without additional premarket review.\n\nThe PCCP is limited to the addition of single nucleotide variants and small insertions and deletions (≤20 bp) in the BRCA1 and BRCA2 genes. The plan describes the specific clinical validation criteria that must be met to demonstrate that the new BRCA1/2 variants are high-risk, highly penetrant BRCA1/2 variants (i.e., those that are demonstrated to be linked to hereditary breast and ovarian cancer (HBOC) syndrome). Specific analytical validation protocols and acceptance criteria are also detailed in the plan to ensure that the device maintains the following performance characteristics for each new BRCA1/2 variant:\n\nAdditionally, software verification and validation activities are detailed in the PCCP and all must be completed successfully to modify the report to add the new BRCA1/2 variants. The plan specifies change control for genotype calling definitions and labeling updates to ensure the device remains as safe and effective as the predicate device.\n\nAlthough details aren’t included, it is clear that the nature of the planned changes and the controls for the changes must be specified in detail.\n\n| Date | Changes |\n|---|---|\n| 2024-08-07 | Initial Version with just the basic questions. |", "url": "https://wpnews.pro/news/pccps-best-practices-faqs-and-examples", "canonical_source": "https://innolitics.com/articles/pccps-best-practices-faqs-and-examples/", "published_at": "2026-07-17 05:00:00+00:00", "updated_at": "2026-07-23 22:29:19.096349+00:00", "lang": "en", "topics": ["ai-policy", "ai-products"], "entities": ["FDA", "Innolitics", "Predetermined Change Control Plan", "PCCP", "510(k)", "De Novo", "PMA"], "alternates": {"html": "https://wpnews.pro/news/pccps-best-practices-faqs-and-examples", "markdown": "https://wpnews.pro/news/pccps-best-practices-faqs-and-examples.md", "text": "https://wpnews.pro/news/pccps-best-practices-faqs-and-examples.txt", "jsonld": "https://wpnews.pro/news/pccps-best-practices-faqs-and-examples.jsonld"}}