Regulatory Open Forum

 View Only
  • 1.  software versions-DoC/Letter to file

    This message was posted by a user wishing to remain anonymous
    Posted 25-Mar-2024 08:37
    This message was posted by a user wishing to remain anonymous

    Hello RAPS Community,

    If software versioning scheme is as follow: X.Y.Z where, X= major release, Y=low frequency updates (minor), Z= high frequency updates e.g. bugfixes; does a manufacturer of class I MD software should prepare the DoC for each software version e.g. 3 (X), 3.1 (X.Y) 3.1.1 (X.Y.Z)?

    Same question regarding Letter to File: should letter to file be created for each above version independently?



  • 2.  RE: software versions-DoC/Letter to file

    This message was posted by a user wishing to remain anonymous
    Posted 01-Apr-2024 09:17
    This message was posted by a user wishing to remain anonymous

    Letter To File - yes, definitely for every change.

    DoC - not for bug fixes. That's sort of like changing a fastener in a piece of hardware - no new DoC. If the change is big enough to be considered a new version or model, with a new UDI, perhaps a new catalog number - then a new DoC would seem warranted. If the change resulted in NB review and updated CE certificate, then again a new DoC.

    You might consider writing rules into a work instruction, so it's not left to individual judgments, which might vary.




  • 3.  RE: software versions-DoC/Letter to file

    Posted 01-Apr-2024 09:30

    Anon,

    This type of question has been asked a few times on the forum, which I would refer to those threads for additional information.  If the device is a Class I software only product with no Notified Body intervention, then really the information contained on the Declaration of Conformity (DofC) can be defined by the organisation based on the needs of the customer.  So the question about whether X, or X and Y, or X and Y and Z are needed should be based on what the needs are for identification to the users.  If the software application falls into the Measuring sub-category where Notified Body intervention is required, this might be a discussion or agreement with them as well what needs to be on the DofC.  This also in combination with your user needs.  In general practice, most organisations have an X.Y designation on DofC though this can depend on often low frequency updates are made.  In my awareness, not many put the X.Y.Z because that would be difficult to track and maintain the multiple versions of DofCs over time.

    The Letter to File or addressing each of the changes really depends on the software type and what type of information or clinical decision is being made.  All changes should be recorded on a change control process, so even the 'Z' changes would still be documented.  What I have seen is most companies are doing the "regulatory notification" determination at the 'Y' level with proper rationale and justification for these changes.  In most cases, the 'X' level changes usually requires a review or even new submission by a regulatory agency - but again these would also all be documented through the change control process.



    ------------------------------
    Richard Vincins ASQ-CQA, MTOPRA, RAC
    Principal Strategy Consultant
    NAMSA
    ------------------------------



  • 4.  RE: software versions-DoC/Letter to file

    Posted 02-Apr-2024 02:39

    Our NB took a very strict stance and requires us to have full SW identifier in the DoC. Their basis was that MDR requires "unambiguous reference allowing identification and traceability". Therefore, we write all three digits in the DoC from now on. 



    ------------------------------
    Hannes Hyvonen
    RA Manager at Icare Finland
    Icare Finland
    Vantaa
    Finland
    ------------------------------



  • 5.  RE: software versions-DoC/Letter to file

    Posted 02-Apr-2024 09:37

    Hi Hannes,

    thanks for sharing that information.

    Out of curiousity: Did the NB address how they interpret the next sentence of Annex IV (4) (where I think the "unambiguous" quote comes from) that states: "Except for the product or trade name, the information allowing identification and traceability may be provided by the Basic UDI-DI referred to in point 3."? Because referencing the Basic UDI-DI as a suitable unambiguous reference does not necessarily reflect the need to include a SW identifier down to the last digit.

    Best regards, Christoph



    ------------------------------
    Christoph Kiesselbach
    Schrack & Partner
    Reutlingen
    Germany
    ------------------------------



  • 6.  RE: software versions-DoC/Letter to file

    Posted 02-Apr-2024 09:55

    I think they did not consider Basic UDI-DI adequate to distinquish between different software versions, as different versions can be placed on the market at the same time. Let's say 1.7.6 and 1.7.7 are on the market at the same time, the have the same basic UDI-DI, so how you differentiate between the two.

    Something along that lines was their reasoning, and we decided not to challenge it as it's pretty easy to write a DoC.



    ------------------------------
    Hannes Hyvonen
    RA Manager at Icare Finland
    Icare Finland
    Vantaa
    Finland
    ------------------------------



  • 7.  RE: software versions-DoC/Letter to file

    Posted 02-Apr-2024 09:58

    Plus in any case there's need to include UDI:s (UDI-DI + UDI-PI) as part of technical documentation per MDR article 27(7) and MDCG 2021-19.

    We use SW version as our UDI-PI.



    ------------------------------
    Hannes Hyvonen
    RA Manager at Icare Finland
    Icare Finland
    Vantaa
    Finland
    ------------------------------



  • 8.  RE: software versions-DoC/Letter to file

    Posted 02-Apr-2024 10:18

    Thanks for the quick reply. If you can live with it there is no reason to challenge the interpretation of the NB, it is just another topic where I think the MDR is not as unamibiguous as I would like it to be and I am always interested in different interpretations.

    Same with the requirement to keep an up-to-date list of UDIs in the technical documentation per MDR article 27(7): Even though MDCG 2021-19 (correctly) adds that UDI means UDI-DI+UDI-PI I would usually interpret this as meaning the respective UDI-DIs plus the rule how to generate the UDI-PI, not a day-to-day update on your actual production identifiers (that may be feasible for software, but not so much for other devices with acutal manufacturing processes).



    ------------------------------
    Christoph Kiesselbach
    Schrack & Partner
    Reutlingen
    Germany
    ------------------------------