The article describes the most critical part of the preparation for a De Novo Classification Request, the De Novo pre IDE meeting.

There are two critical differences between a De Novo classification request and a 510k submission. First, 510k clearance is based upon a substantial equivalence comparison of a device and a predicate device that is already marketed in the USA, while a De Novo classification is based upon a benefit-risk analysis of a device’s clinical benefits compared with the risk of harm to users and patients. Second, 510k clearance usually does not require clinical data to demonstrate safety and efficacy, while a De Novo classification request usually does require clinical data to demonstrate safety and efficacy. Therefore, it makes sense that the two most common challenges for innovative medical device companies are: 1) learning how to write a benefit-risk analysis, and 2) clinical study design. Success with both of these tasks can be significantly improved by requesting a De Novo pre IDE meeting with the FDA.

Benefit-risk analysis questions to ask during a De Novo pre IDE meeting

Most device companies are only familiar with substantial equivalence comparisons–not a benefit-risk analysis. The statement “the benefits outweigh the risks” is not a benefit-risk analysis. Medical device regulations have been changing toward an emphasis on benefit-risk anlysis. For example, the European MDD mentions benefit-risk analysis eight times, while Regulation (EU) 2017/745 mentions benefit-risk analysis 69 times. Despite the obvious increased emphasis on benefit-risk analysis in the new EU Regulations, ISO 14971:2019 only requires a benefit-risk analysis for unacceptable risks. The international standard also does not clearly explain how to perform a benefit-risk analysis. The best explanation for how to perform a benefit-risk analysis was provided in the FDA guidance, but now the ISO/TR 24971:2020 guidance includes detailed guidance in Clause 7.4.

In addition to reading ISO/TR 24971 and the FDA guidance, you will need to systematically identify all of the current alternative methods of treatment, diagnosis, or monitoring for your intended use. Therefore, you should ask in a pre-submission meeting if there are any additional devices or treatments that the FDA feels should be considered. You should review each of the alternative treatments for clinical studies that may help you in the design of your clinical study. You should carefully review the available clinical data for alternative treatments to help you quantify the risks and benefits associated with those treatments too. Finally, you should consider whether one or more of these alternative treatments might be a suitable control for your clinical study. Ideally, your clinical study design will show that the benefits of your device are greater, and the risks are less, but either may be enough for approval of your classification request. If you think the risks of your device are significantly less than alternative treatments, then ask the FDA about using this factor as an endpoint in your study design.

Clinical Study Design Considerations

Ideally, there is already a well-accepted clinical model for assessing efficacy for your desired indications. This means multiple, published, peer-reviewed journal articles. You might have a better method for evaluating subjects, but don’t propose that method instead of a “gold standard.” If you feel strongly that your method is more appropriate, propose both methods of evaluation. You also need multiple evaluators who can be objective. Randomization, blinding, and monitoring of clinical studies is critical to ensure an unbiased evaluation of clinical results. In general, it is difficult to design an unbiased post-market clinical follow-up (PMCF) study. A common deficiency identified by the FDA is that post-market study performed outside the United States (OUS) has selection bias and covariate imbalance.

You also need to design your study with realistic expectations. Murphy’s law is always active. That means, “things will go wrong in any given situation if you give them a chance.” Therefore, you must avoid optimism and devise methods for detecting errors quickly. This is why electronic data capture systems and eSource is preferred for data collection instead of the manual collection of data on paper case study forms. Not only does it reduce errors in data collection, but it also facilitates remote monitoring of clinical sites. This includes asking questions that are open-ended or quantitative–instead of Yes/No questions or qualitative evaluations that encourage subjectivity. You can always anticipate every mistake that will be made, and open-ended questions often capture essential data that would otherwise be lost. Asking the quantitative questions also will provide you with additional data you can analyze, which may reveal unexpected relationships or help you to explain unexpected results. To help facilitate the development of these questions, try asking yourself how you could detect an error for each data point you are collecting. Then add a detection mechanism to your data collection plan wherever and whenever you can.

Goals of De Novo pre IDE Meeting

A pre-IDE meeting is not typically your first pre-submission meeting with the FDA. Usually, your first pre-submission meeting is to verify that the FDA agrees that the regulatory pathway is a De Novo classification request rather than a 510k submission. Hopefully, you also were able to review your overall testing plan with the FDA during your first pre-submission meeting. You may have even reviewed a clinical synopsis with the FDA during your initial pre-submission meeting. During the pre-IDE meeting, your goal is to finalize your clinical study protocol. That doesn’t mean that the FDA should agree 100% with your draft protocol. You want positive and negative feedback on all aspects of your protocol before the IDE submission. During the IDE review, changes will be made.

The most important aspects of getting right before the IDE submission are the fundamentals. Most of our De Novo clients feel that a control group is not possible, because they think that test subjects will know when a sham device is used. However, trying to avoid a control group is nearly impossible. The most important factors for why a control group is needed are:

• you need to minimize differences between experimental and control subjects, but you can’t do that if you are relying on data from other clinical studies
• you also need to ensure that your evaluation methods are identical, which is nearly impossible when performed by different people, at different facilities, using slightly different protocols

Another area of weakness in most draft clinical protocols is the method of evaluation. Specifically:

• Who is doing evaluations?
• Which endpoints are important?
• When are your endpoints?
• What are your acceptance criteria?

The last area to consider in a pre-IDE meeting is your statistical plan. You need a statistical plan, but the statistical analysis seldom appears to be the reason for the rejection of clinical data. The reason is that changes can be made to your statistical analysis of data after the study is completed, but you can’t change the data once the study is over. The FDA is now accepting adaptive designs that allow the company to analyze data during the study to recalculate the ultimate sample size needed based upon actual data rather than initial assumptions.

What are the basic milestones in an FDA pre IDE meeting?

Once you have prepared your De Novo pre IDE meeting request and the FDA PreSTAR template indicates that the submission is complete, then you are ready to submit the PreSTAR to the FDA. There are 11 steps in the process for an FDA pre-submission meeting request (see flow chart above). The following is a brief summary of each milestone the process:

1. Upload your completed FDA PreSTAR to the FDA Customer Collaboration Portal (CCP)
2. Within seconds of uploading your file, you will receive an automated email acknowledging the uploading of the file to the CCP, but your real acknowledgement is a letter your receive by email the following day that has the Q-Sub number assigned to your request (i.e., QYYXXXX).
3. A preliminary review of the PreSTAR used to be performed and a checklist was filled out, but with the use of the PreSTAR template the FDA is now only conducting a technical review. Once the technical review is completed, a lead reviewer is assigned and you will receive an email notifying you of who your reviewer is. This process usually is completed in 15 days.
4. Once the FDA lead reviewer is assigned, the lead reviewer will ask subject matter experts to review and respond to each of the questions in the De Novo pre IDE meeting request. Pre-submissions are limited to a maximum of four topics, and therefore, there should only be a need for a four subject matter experts. The lead reviewer may be one of the subject matter experts. Typically the assistant director of the review panel (i.e., medical specialty) will also participate, and sometimes the director will also participate. This review process will occur from day 15 until day 70 of the pre-submission process. The lead reviewer will also contact you by email to coordinate a date and time for scheduling the pre-submission teleconference.
5. Approximately five days before the scheduled De Novo pre IDE meeting teleconference, the FDA lead reviewer will provide a email response to your questions. The file name of the document sent by the lead reviewer will be in the following format: “QYYXXXX.Notification.EMFB.”
6. Your team will need to review the FDA responses to each question and decide whether you want to ask any clarification questions. If you don’t have any clarification questions, you can cancel the FDA teleconference. A slide deck is not required for the teleconference, but if you decide to create a slide deck the FDA would like to receive it by email ~48 hours before the teleconference.
7. For the De Novo pre IDE meeting teleconference, the FDA lead reviewer provides login information for a Microsoft Teams or Zoom teleconference in advance of the meeting. Attendees can login approximately 5 minutes before the start. The meeting begins with introductions, and then the company will present their slides and ask clarification questions. At the end of each slide, it is a good practice to ask if anyone from the company has additional questions, and then you should ask if the FDA have anything they would like to ask. We recommend alternating speakers for presentation of slides. This gives multiple people practice presenting to the FDA, it provides some variety of speakers, it increases engagement during meetings, and it allows people that are not speaking time to catch-up on their notes. The FDA will not permit recordings during the meeting. At the end of the meeting, you will want to leave approximately 5 minutes for summarizing any action items for your company or the FDA.
8. After the teleconference most companies will conduct a debrief meeting without the FDA. Notes from each person will be shared with the person designated for creating draft meeting minutes. The minutes are intended to be only a summary of what was discussed–not a transcription. You have 15 calendar days to create the draft.
9. Once the company agrees on a final draft of the meeting minutes you will prepare an FDA eCopy and upload it to the FDA CCP. If you created a slide deck, the slide deck should be included with the meeting minutes as a second document in the FDA eCopy. Lead reviewers will also sometimes request an MS Word version of the minutes be emailed directly to them to facilitate editing the minutes.
10. Within seconds of uploading your FDA eCopy of the minutes, you will receive an automated email acknowledging the uploading of the FDA eCopy to the CCP. The document number assigned for meeting minutes is in the following format: QYYXXXX.A001.
11. The FDA will take 30 days to review your draft meeting minutes. The minutes will be redlined by the FDA with what the FDA intended to say regarding your questions so it may differ from exactly what was said. You will receive an email with a letter attached. The filename of the letter will be: QYYXXXX.A001.Meeting Minutes.REVS. You can dispute the minutes if you disagree with the FDA’s redlines.

How to you use meeting minutes in your final De Novo application?

In your FDA eSTAR, you will need to attach a copy of the meeting minutes from any pre-submissions you had with the FDA. You will also need to provide a response memo indicating how you addressed any concerns or questions the FDA raised during those pre-submissions. Both documents will be attached to the eSTAR. We recommend using a tabular format and alternating between blue and black font to clearly separate the minutes from your responses.

Do you need more information about De Novo applications?

We created a cornerstone webpage that summarizes our content about De Novo applications. All of our webinars about De Novo applications and our De Novo templates can be purchased as part of our 510k Course Series. You can also learn a lot about clinical study design by purchasing Medical Device Academy’s Clinical Procedure (SYS-009). Finally, try searching the De Novo Reclassification Summaries for examples of how other companies designed their clinical studies to demonstrate safety and efficacy for a De Novo application.

The new FDA goal is to reduce the De Novo review timeline to 150 days for 70% of De Novo submissions, but how long does it take now?

What is an FDA De Novo submission?

An FDA De Novo submission is an application submitted to the FDA for creating a new device product classification. There are three classifications of devices by the FDA: Class 1, Class 2, and Class 3. Class 1 devices are the lowest-risk devices, and they only require general controls. Class 2 devices are moderate-risk devices that require “Special Controls,” and Class 3 are high-risk devices that require Pre-Market Approval (i.e., PMA). De Novo applications can only be submitted for Class 1 and Class 2 devices, and most of the De Novo submissions require clinical data to demonstrate that the clinical benefits of the new device classification outweigh the risks of the device to patients and users. It’s the need for clinical data that is partly responsible for the longer De Novo review timeline.

What is the De Novo review timeline?

Initially, the FDA required that Class 2 devices must be first submitted as a 510k submission. If the device did not meet the criteria for a 510k, then the company could re-submit a De Novo Classification Request to the FDA. On July 9, 2012, the regulations were revised to allow companies to submit De Novo Classification Requests directly. This makes sense because some devices have novel indications for use, and submission of a 510k would be a complete waste of time in money. For example, the first SARS-COV-2 test had to be submitted as a De Novo by Biofire to obtain permanent approval for the test instead of emergency use authorization (EUA). This change in 2012 dramatically reduced the De Novo review timeline.

On October 4, 2021, the FDA published a final rule for De Novo Classification Requests. This new regulation identified the De Novo review timeline as 120 calendar days. Even though 120 days is 30 days longer than the FDA review clock for a 510k, the actual timeline to review De Novo submissions was much longer.

Every five years, when Congress reauthorizes user fee funding of the FDA, new MDUFA goals are established. The draft MDUFA performance goals (which impact FDA funding) were published recently. The specific performance goal to review De Novo submissions is:

FDA will issue a MDUFA decision within 150 FDA Days for 70% of De Novo requests.

There are two problems with this goal. First, the term “FDA Days” is based on calendar days minus the number of days the submission was placed on hold, and we don’t have any visibility into the number of days submissions are placed on hold. In the past, submissions could be placed on hold multiple times during the Refusal to Accept (RTA) screening process, and the “FDA Days” is reset to zero days each time the company receives an RTA hold letter. In addition, even after the submission is finally accepted, the FDA places the submission on hold when they request additional information (i.e., AI Hold). RTA and AI Hold periods can last up to 180 days, and during the Covid-19 pandemic, companies were allowed to extend this up to 360 days.

The second problem with the MDUFA goal is that we only have visibility into the outcome of De Novo submissions that were granted. More than 60 De Novo submissions are submitted each year, but the number of De Novo Classification Requests granted ranged between 21 and 30 over the past three years. Therefore, the 50%+ of De Novo applications denied could skew the % of submissions that meet the MDUFA goal for the De Novo review timeline.

What is the FDA track record in reviewing a De Novo?

Every CEO I speak with asks the same question: “How long does the FDA review take?” In preparation for a webinar I taught about De Novo Classification Requests in 2019, I researched the latest De Novo review timelines. I expected the review timelines to be close to 150 calendar days because the FDA decision goal was 150 FDA days. The 150-day goal was set in 2018 when Congress approved MDUFA IV. The 2019 data held two surprises:

1. only 21 De Novo requests were granted in 2019, and
2. the average review timeline was 307 calendar days (i.e., the range was 108 days to 619 days).

FDA days are not the same as calendar days. Only 23.8% of De Novo submissions were reviewed within 150 calendar days. The FDA doesn’t calculate the number of FDA days as calendar days, but there is no way to know how much time each De Novo spent on hold publicly. Upon seeing the announcement of a new decision goal for MDUFA V on October 5, 2022, I decided to revisit my previous analysis.

*Only 9+ months of data for 2022, because data was collected on October 17, 2022.

We can blame the Covid-19 pandemic for the slower De Novo review timeline during the past few years, but you would expect a longer average duration in 2020 if that was the root cause of the FDA’s failure to achieve the MDUFA IV target of 150 calendar days. You would also expect 2021 to have the longest review timelines. Instead, the review timelines are the slowest for 2022. The number of De Novo submissions remains small, and therefore it is hard to be conclusive regarding the root cause of the failure to reach the 150-day decision goal. In addition, the percentage of De Novo applications granted within 150 calendar days was lowest in 2021, as you would expect if the reason for delays is primarily due to the Covid-19 pandemic.

Is there any good news?

The FDA is allowing the new eSTAR templates to be used for De Novo Classification Requests. These new electronic submission templates standardize the format of all 510k and De Novo submissions for FDA reviewers. The eSTAR also forces companies to answer all questions in the FDA reviewer’s checklist to ensure the submission is complete and accurate before the new submission is submitted to the FDA.

The new eSTAR templates were first used in 2021, and our firm has observed shorter overall review timelines and fewer deficiencies identified by FDA reviewers when they submit an “Additional Information Hold” (AI Hold) to companies.

How can the FDA improve De Novo timelines?

The FDA, industry, and Congress seem to be taking the same approach pursued five years ago to improve the review timeline for De Novo submission. MDUFA V authorized additional user fees for De Novo submissions (i.e., 17.8% increase), and the FDA will be authorized to hire additional employees each year during MDUFA V if the performance goals are met. However, there are three other options that the FDA and industry should have seriously considered during the FDA-industry negotiations.

The first option that should have been considered is to allow third-party reviewers to review the elements of a De Novo that are identical to a 510k submission:

1. sterilization validation
2. shelf-life testing
3. biocompatibility testing
4. software validation
5. electrical safety testing
6. EMC testing
7. wireless testing
8. interoperability testing
9. benchtop performance testing
10. animal performance testing
11. human factors engineering

The above approach would require blended pricing where the FDA charges a smaller user fee than a Standard De Novo user fee, and the third-party reviewer charges a smaller fee than a 510k. The combined cost would be higher than the FDA Review of a De Novo, but this would reduce the number of hours the FDA needs to complete their review of a De Novo, and it would allow for pricing that is much lower than the De Novo standard user fee for qualified small businesses.

A second approach would be to pilot a modular review approach. A modular review would be similar to modular reviews for PMA submissions. In a modular review, the FDA can review most submission sections and provide feedback before the human clinical performance data is available. This would not help the few De Novo submissions that do not include human clinical performance data, but this would have a profound positive impact on most De Novo projects. First, the FDA would be able to complete the review of all sections in the submission except the human clinical performance data without delaying the final De Novo decision. Second, a successful review of non-clinical data by the FDA would give investors more confidence to fund pivotal clinical studies required to complete the De Novo submission.

A third approach would be for the FDA to force manufacturers to submit testing plans and protocols as pre-submissions to the FDA. This approach would give the FDA more familiarity with each device and the testing plan before reviewing the data. This approach would also reduce the hours FDA reviewers spend reviewing data that doesn’t meet the requirements and writing deficiencies. This approach would also give investors more confidence to fund De Novo projects for all V&V testing.