Reflecting on the First Year of CDRH’s 510(k) Refuse to Accept Policy

By Allyson B. Mullen & Jeffrey N. Gibbs –

The end of 2013 marked the first anniversary of CDRH’s Refuse to Accept Policy for 510(k)s (the “RTA”).  FDA, Guidance for Industry and FDA Staff, Refuse to Accept Policy for 510(k)s (December 2012).  Therefore, we thought it would be a good time to reflect on how well the RTA has been working (or not working) from our perspective.  The stated intent of the RTA is modest: “to assess whether a submission is administratively complete.”  RTA at 2.  Like others in the device regulatory space, we had our reservations about the RTA when it was issued.  You may remember from our prior post, we were concerned that reviewers may need to do a substantive review of the submission in order to complete the RTA checklist (thus, making the RTA more than an administrative check of completeness).  Based on the RTA comments we have seen (and heard), it seems that our substantive-review concerns understated the issues created by this policy. 

As of September 2013, CDRH reported that approximately 60% of all newly filed 510(k)s were refused under the RTA.  CDRH Meeting FDASIA Goals, Confirms Compliance Re-Org, DEVICES & DIAGNOSTICS LETTER (FDA News, Falls Church, V.A.), Sept. 30, 2013, at 1.  According to Dr. Jeffrey Shuren, Director of CDRH, the majority of the refusals were caused by the failure to provide shelf life and biocompatibility information.  Id.  However, the RTA comments we are familiar with have not been that simple (or well founded).  To illustrate, we wanted to share, in no particular order, some of our top (or worst) RTA comments from 2013.

• In response to a 510(k) for a condom, in which the applicant supplied the required biocompatibility data, the applicant was cited in the RTA checklist for failure to specifically state whether the condom was patient contacting, and if so, for more or less than 24 hours.  It seems obvious to us that a condom, solely by virtue of its well known intended use, is patient contacting, and (we would hope) for less than 24 hours.  Even if the reviewer did not know this, the answer to the comment was, in fact, clearly stated in the submission itself.
• In response to a 510(k) in which the applicant cited its compliance to a Special Controls Guidance Document for labeling of the particular device type, the Center responded with an RTA comment asking for an explanation of how following the guidance satisfied the requirements of 21 C.F.R. § 807.87.  The guidance contained labeling requirements.  Labeling is a required element under 21 C.F.R. § 807.87.  It would seem self explanatory that by conforming the proposed device labeling to the requirements of the Special Controls Guidance Document that the applicant was complying with the labeling requirement in 21 C.F.R. § 807.87(e).  But, either it was not so obvious to the reviewer, or everything needs an explanation when it comes to the RTA.
• One set of RTA comments cited the applicant for the failure to provide “a table format with short summary/discussion clearly labeled ‘SE comparison discussion’” in the 510(k) summary.  We remember the good old days when a 510(k) summary could just contain a few sentences comparing the proposed and predicate devices in terms of intended use and technological characteristics.  Now, you had better not forget to replicate your full substantial equivalence table.  Also, this is just one example of the 510(k) summary comments that companies have been receiving during the RTA review.  In our view, 510(k) summary comments are inappropriate at the RTA stage since the 510(k) summary is a synopsis of the basis for a determination of substantial equivalence, and at the RTA stage no determination as to substantial equivalence is being made.  Moreover, summary tables tend to change during the 510(k) review process.  Furthermore, parsing the details of the 510(k) summary seems unrelated to the ability of the FDA to review the 510(k) itself – the stated purpose of the RTA.
• In response to a 510(k) for a device that is exempt from the requirement of adequate directions for use under the Act because the device’s use is commonly understood by the ordinary person, the Center noted in the RTA that although “the submission stated that the device is exempt from section 502(f)(1) of the Act (common instructions for use),” “the labeling still did NOT include a indications for use (IFU) that is IDENTICAL to the IFU; which is a requirement of the RTA-checklist” (emphasis added).  If there is a requirement for device labeling to contain an IFU statement when the device is exempt from the requirement of adequate directions for use, we would expect the Center would have referenced the CFR rather than the RTA as support for this requirement.  It is striking that the RTA guidance is now creating a “requirement,” rather than simply ensuring that 510(k)s “meet[] a minimum threshold of acceptability” to be accepted for substantive review.  
• As we have previously blogged about, in response to a software 510(k), the Center indicated that the submission could not be accepted because the applicant had not demonstrated that the proposed software complied with the Draft Guidance for the Content of Premarket Submissions for Management of Cybersecurity in Medical Devices (underlined for emphasis).  In effect, an administrative checklist is being combined with a draft, non-binding document to create substantive requirements before FDA will proceed with fulfilling its statutory obligation to review 510(k)s.
• In response to a 510(k) for a vinyl glove, the Center noted the following deficiencies in the RTA, “information regarding whether the device contains software is not provided,” and “information regarding whether the device requires EMC and Electrical Safety evaluation is not provided.”  Yes, this 510(k) was for the type of vinyl glove worn by the doctor when doing a patient exam.  Gloves do not contain software; that should be self-evident.  The RTA clearly stated that the information was not applicable, “however, the submission did not state [this].”  Once again, common sense seems to be subordinated to a checklist. 
• Lastly, one reviewer asked for the six-month shelf life rationale for a device during the RTA review.  It is a substantive question, not a procedural one.  Also, there is no reason for the question since the RTA simply asks if “proposed shelf life/expiration date [is] stated.”  The application met the “requirement.”  Asking for a rationale went beyond the checklist.  The shelf life was stated and the reviewer should have moved on.

We have also heard of the Center citing typos and misprints in RTAs.  For example, in response to one 510(k), FDA cited the applicant for having two page 75s (rather than a page 75 and a page 76).  In another instance, an applicant received an RTA comment stating that although the applicant cited the correct classification regulation for the proposed device, the wrong classification name was provided.  It turned out that the applicant used the device name from the product code rather than the name from the classification regulation.  In the comment, FDA told the applicant the correct classification name (would it not have been easier for FDA to simply note that themselves as they reviewed?).

Last September, Director Shuren said that the intent of the RTA was to return “low quality submissions” thereby “reducing inefficiencies caused by the need to review poorly or incomplete applications.”  CDRH Meeting FDASIA Goals, Confirms Compliance Re-Org at 1.  Based on the examples we have seen, we question whether the 60% of rejected submissions are really “low quality,” or, if the Center is being overly picky (and in some instances failing to use common sense), and consequently, has created a process that actually further delays the overall review time for 510(k)s rather than improving it.  In a recent discussion with Karen Becker of Precision for Medicine, she told us that on average it is taking her clients approximately four weeks to respond to the RTA comments.  This length is due to a number of factors, for example sometimes they have discussed the comments with the reviewers to ensure that all questions are being addressed completely so that there is only one RTA, rather than multiple rounds.  

Since we know we are only seeing a fraction of the RTA comments, we are inviting you, our readers, to share your RTA experiences and comments.  We are opening up our comments section to allow you to include, in the comments below, examples of problematic RTAs. 

One final note: CDRH’s final Pre-Submission Guidance also contains a checklist (blog post is forthcoming).  Thus, Pre-sub’s can now be rejected for administrative reasons.  We certainly hope that CDRH does not convert the pre-sub checklist into yet another potential trap for companies.