FDA’s Pre-Cert Pilot Ends. Will there be a Sequel?November 1, 2022
FDA began the Software Precertification (Pre-Cert) pilot program in 2017 to evaluate an alternative approach to regulation of software as a medical device (SaMD) over the total product lifecycle (TPLC). We have followed it over the last five years, with blog posts here, here, here, here, here, here, and here. The Pre-Cert pilot program, as described in the Working Model, included an Excellence Appraisal (pre‑certification) to demonstrate a culture of quality and organizational excellence, a review determination to evaluate whether a pre-market review would be required for the particular SaMD, a streamlined premarket review, when required, and evaluation of real-world performance to verify the product’s continued safety, effectiveness, and performance. As part of the pilot program, FDA included a Test Plan.
As of September 2022, FDA has now announced the conclusion of the pilot program. It also has released a report summarizing the findings from the pilot program. The Software Precertification (Pre-Cert) Pilot Program: Tailored Total Product Lifecycle Approaches and Key Findings (“Key Findings”). The Key Findings are summarized below, along with our thoughts on the lessons learned in the pilot and how some of the pilot’s findings may be useful to industry while we wait to see if there will be a sequel.
FDA found that further development would be needed before they would be comfortable identifying low-risk devices where the Excellence Appraisal alone could be relied upon without further premarket review. The basis of this conclusion relates to the large range of device indications and technology and nuances of testing to support them. It seems that FDA may have validated that, even with an Excellence Appraisal above and beyond standard compliance with good manufacturing practices, there is a need for pre-market review for Class 2 devices. FDA notes that “a future approach could build on features of the current regulatory system, where a Quality Management System and other general and, in some cases, special controls provide a reasonable assurance of safety and effectiveness for certain low to moderate risk devices.” Key Findings at 12.
In their evaluation of the Streamlined Review process, FDA identified an issue that industry has complained about for years – that there is variability in how information is used in making regulatory decisions. FDA proposes that “modern mechanisms that support clear and consistent communication of device information could facilitate efficient device review activities” and notes that they “consistently observed the need for structured data . . . .” Id. One struggle that many companies face is the need to generate data in compliance with their quality system and then reformat or restructure it for review in a premarket application. We hope that any regulatory tools proposed by the Agency to support premarket review are developed in conjunction with FDA’s post-market reviewers and inspectors so that changes are helpful across the TPLC and do not introduce new burdens.
As discussed in our previous posts, we were concerned that the Agency may not be able to adequately validate the program. FDA did find that their ability to fully test the program was limited. Challenges to a successful test program included there being only nine companies in the pilot program and the need to limit formal implementation to the De Novo classifications. It was noted that authorizing a device through a De Novo would create pilot specific special controls that other companies would need to follow, even if not participating in the pilot program.
Despite the limitations in testing, some of the findings were noteworthy. As part of the testing, FDA retrospectively reviewed applications to determine if the Excellence Appraisal could be used in lieu of certain other premarket software documents. They found that either a “concise statement of precertification without the expectation of further premarket review of the Excellence Appraisal information, or a detailed report of the Excellence Appraisal process and results to be reviewed in-depth in the context of the device subject to review” would be useful, but a summary that fell in between these two approaches was least valuable. Id. at 8. While the former approach would likely be welcome by sponsors, the need to write a detailed report of the Excellence Appraisal as it applies to the device may be equally burdensome to the software document it is intended to replace.
Key Performance Indicators
The pilot resulted in many findings related to excellence principles and key performance indicators (KPIs) that may be useful to manufacturers of SaMD, as well as other device types, both with and without software. Within the current legislative framework, the efficiency of the premarket review process will be best when device information is clearly presented and answers key questions related to review of the device’s safety and effectiveness.
Implementing some or all of the KPI objectives presented in Appendix List A of the Key Findings document may, therefore, improve efficiency of the regulatory process regardless of whether there is a new regulatory framework for software devices. You may also note that many of these key findings can be considered best practices for implementation of a quality management system in compliance with the Quality System Regulation, 21 C.F.R. Part 820. The Descriptive KPI Objectives include:
- Processes engage the right people, at the right times, to the right degree: “We use knowledgeable, qualified, and multidisciplinary teams throughout the TPLC.”
- Development process results in well-characterized software: “Our software behaves as expected.”
- Deployment and monitoring process confirms well-characterized software in context of use: “Our software behaves as expected in the real-world.”
- Patching process ensures timely resolution of issues across the entire installed base: “We can fix our software when it doesn’t behave as expected.”
- Update process ensures modifications meet user needs identified through real-world use: “We can identify and implement improvements to the expected behavior of our software.”
Key Findings at 18.
Finally, FDA concluded that there were challenges with implementing the Pre-Cert approach under the current statutory authorities. While there is a need for new statutory authority, the Key Findings report does not provide any details as to what type of regulatory framework would be useful to achieve the goals of the Pre-Cert program. FDA provides only high-level information on the need for a “flexible, risk-based approach to regulation” that “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.” Id. at 13.
While it is not clear if there will be a sequel to the Pre-Cert pilot program, we look forward to seeing what new ideas FDA’s Digital Health Center of Excellence will bring to reducing regulatory burdens to allow software devices to be developed and iterated efficiently, while ensuring safety and effectiveness.