Regulatory Open Forum

 View Only
  • 1.  IVD performance study for 510(k)

    Posted 16-Jan-2017 13:37
    Does the performance study to evaluate our IVD's analytical and clinical performance characteristics for our 510(k) submission have to include the same of study sites and number of subjects as our chosen predicate?

    ------------------------------
    Noemi Olivo
    Milford CT
    United States
    ------------------------------


  • 2.  RE: IVD performance study for 510(k)

    Posted 18-Jan-2017 11:53

    The answer, as is often the case, is "it depends." There area  number of considerations:

    - Were they complying with a guidance document or recognized standard?
    - Was the study powered to demonstrate a certain endpoint?
    - Has your team come up with a different, scientifically valid, number that you think is appropriate?
    - When was your predicate cleared? Has the Agency's thinking in the area changed over time?
    - Do you have the same scientific technology, operating principles, etc.?

    The safest route, of course, is to submit a pre-sub.



    ------------------------------
    Paul Swift, RAC
    Director, Global RA
    Fort Worth, TX
    United States
    ------------------------------



  • 3.  RE: IVD performance study for 510(k)

    Posted 19-Jan-2017 06:20
    Thank you Paul; great advice.  I'm currently working on the Pre-Sub with hopes to submit by end of this month. 

    Here's another question, if you don't mind:  I am working on the final drafts of the proposed study protocol for the Pre-Sub and wondering if the FDA reviewers will be looking for a "complete" statistical plan - I have included a section with what amounts to an outline of the requirements, but since we don't have a statistician, yet, it's not a concise plan.  My company is a small start-up with limited resources and does not want to contract a statistician until the "final" performance studies protocol is being written for the 510(k) performance studies.  Is this not a good plan?

    ------------------------------
    Noemi Olivo
    Milford CT
    United States
    ------------------------------



  • 4.  RE: IVD performance study for 510(k)

    Posted 27-Jan-2017 12:58
    I don't think designing a clinical study without the involvement of a qualified biostatistician is ever a good plan.

    ------------------------------
    Julie Omohundro, ex-RAC (US, GS), still an MBA
    Principal Consultant
    Class Three, LLC
    Durham, North Carolina, USA
    919-544-3366 (T)
    434-964-1614 (C)
    julie@class3devices.com
    ------------------------------



  • 5.  RE: IVD performance study for 510(k)

    Posted 18-Jan-2017 13:16
    There is no requirement that the study be identical; however, you can use that study as an indicator of the FDA's current thinking as you design your own appropriate study. The considerations in Paul's response are all very good and I agree with his recommendation of a pre-submission to get concurrence before starting any large IVD study.

    ------------------------------
    Michael Nilo
    Network Regulatory Partners
    Nilo Medical Consulting Group
    Portland OR
    United States
    ------------------------------



  • 6.  RE: IVD performance study for 510(k)

    Posted 19-Jan-2017 06:22
    Thank you Michael for your good advice.  I posted a response to Paul if you will please take a look at that; I'd like your opinion too if you will please.

    ------------------------------
    Noemi Olivo
    Milford CT
    United States
    ------------------------------



  • 7.  RE: IVD performance study for 510(k)

    Posted 20-Jan-2017 19:30
    A set of planned "fence posts" for a study is fine during a pre-submission. Since you have the opportunity to specify questions in a pre-submission, I would suggest asking the FDA to comment on your outline with regards to the statistical aspects as well. 

    ------------------------------
    Michael Nilo
    Network Regulatory Partners
    Nilo Medical Consulting Group
    Portland OR
    United States
    ------------------------------



  • 8.  RE: IVD performance study for 510(k)

    Posted 22-Jan-2017 11:51
    Will do - thank you Michael!

    Noemí

    ------------------------------
    Noemi Olivo
    Milford CT
    United States
    ------------------------------



  • 9.  RE: IVD performance study for 510(k)

    Posted 27-Jan-2017 13:06

    I agree on the "concurrence," meaning you've developed a scientifically and statistically valid study design and analysis plan and now you want FDA's concurrence that your design and analysis plan are appropriate to evaluate SE to the predicate. 

    I don't agree with trying to use a pre-sub to get FDA to tell you how to design a scientifically and statistically valid study  FDA's pre-sub guidance seem to be an effort to screen out such requests, based on the information they encourage you to include with your pre-sub request, but I don't know what happens reliably in practice.



    ------------------------------
    Julie Omohundro, ex-RAC (US, GS), still an MBA
    Principal Consultant
    Class Three, LLC
    Durham, North Carolina, USA
    919-544-3366 (T)
    434-964-1614 (C)
    julie@class3devices.com
    ------------------------------



  • 10.  RE: IVD performance study for 510(k)

    Posted 27-Jan-2017 16:41
    Thank you Julie for adding your thoughts.  I agree with your comment on the need for a biostatistician...working hard to get that across.

    Noemi

    ------------------------------
    Noemi Olivo
    Milford CT
    United States
    ------------------------------