Regulatory Open Forum

 View Only
  • 1.  Early signal detection process

    Posted 07-Jul-2020 10:40

    In a recent post Ginger Glaser reported concern. She said, "having now been through two different situations of FDA using their 'early signal detection process' in which the entire signal was 'there seem to be a lot of MDRs' …

    While I read the guidance document about early signal detection, I didn't know that FDA had actually implemented it and sent out information. In particular, Ginger said, "given that HOSPITALS think that FDA has actual facts when they send out notices", which raises a concern for me.

    MDR signals should be a rate calculated using numerator (number of MDRs) and a denominator (number of devices or procedures that could result in an MDR). FDA does not have, as far as I know, the denominator information.

    Can somebody point me to information about early signal detection in practice. In particular, does FDA include the communication on its website for the public to see?

    Dan O'Leary CQA, CQE
    Swanzey NH
    United States

  • 2.  RE: Early signal detection process

    Posted 07-Jul-2020 17:45



    Hi Dan,


    This page gives an overview that's pretty consistent with my experience:



    FDA seems to call it "Signal Management" and it's not at all a transparent process.  I'm not aware that anything is public, unless a safety notice is issued.  It's also not very transparent to the manufacturer, as they are often unaware of whether FDA's looking at a device class or just their device.  Worst is that the manufacturers are not usually informed of any resolution--it just ends up going silent (which is admittedly better than a safety notice).


    In the several times I've observed it happen, the process starts with a list of questions for the manufacturer--including sales by year.  So that gives them a denominator of sorts, though it's often not clear that all devices sold were used in patient care.  Other requests include any related CAPAs or investigations information. 


    Hope that's helpful.


    Best regards,






  • 3.  RE: Early signal detection process

    Posted 08-Jul-2020 03:23
    Hello Dan,

    I would agree with Ted after spending a little time on this as part of some post market surveillance activities.  There is not any "public" information related to this and as far as I know FDA's gets some denominator information from the company itself on a one-to-one basis.  Though I have heard they also use other information like across all manufacturers, the TPLC information, and inspection information to also help create some signal detection.  Personally, I think they just have internally some "thresholds" they have established and use those when they need to make public notices.

    Richard Vincins RAC
    Vice President Global Regulatory Affairs

  • 4.  RE: Early signal detection process

    Posted 08-Jul-2020 10:27
    As for transparency, we have tried to get the materials relating to a "signal" situation that applied to us via FOI. Still waiting over 18 months, because basically FDA "has no process for sorting out material from other vendors that may be in the file."

    As Ted said, FDA does ask for denominator information, but ovbiously the initial "trend" does not use it. It has also been my experience that in conversations they only speak of numbers, not rates. As best I can tell, many of them don't really understand the difference between a malfunction MDR and an injury MDR either, as it often got translated "but there have been XXX injuries" - um, no there were not.


    Ginger Glaser RAC
    Chief Technology Officer

  • 5.  RE: Early signal detection process

    Posted 10-Jul-2020 16:56
    Good evening,

    let me recommend an excellent book, Cobert's Manual of Drug Safety and Pharmacovigilance

    It is useful even for device vigilance. 


    Peter Mikó M.D
    ArtPharm Ltd.