Regulatory Open Forum

 View Only
Expand all | Collapse all

FDA Traceability Question

  • 1.  FDA Traceability Question

    Posted 22-Apr-2019 15:28
    Hi everyone, 

    For clarification, what is the difference in traceability requirements between hardware and software? As per FDA premarket submission  requirements for software, there is a traceability requirement linking the SRS to SDS, V&V and risk control measures.  How can we establish traceability in design control of hardware?  Is the design input matrix in addition to risk analysis table sufficient?  What if the device has both hardware and software components? How are hardware and software traceability established? What kind of matrix would like that look like? 

    Thanks. 


    ------------------------------
    Karen Zhou
    ------------------------------


  • 2.  RE: FDA Traceability Question

    Posted 23-Apr-2019 11:23
    Karen, sorry you haven't gotten any responses, at least in the Forum.

    Could I ask you a couple of potentially odd questions?  What do you see as the purpose of traceability in design control of hardware?  And in establishing hardware and software traceability?

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



  • 3.  RE: FDA Traceability Question

    Posted 24-Apr-2019 11:44
    Karen,

    There is no traceability requirement in 21 CFR 820.  It is however, a best practice to show that you meet other requirements on Design Verification (21 CFR 820.30 (f) and Design Validation 21 CFR 820.30 (g)).  it allows you to provide objective evidence that all of the User Requirements leading to Design Inputs (21 CFR 820.30 (c) which in turn lead to Design Outputs in 21 CFR 820.30 (d) have been addressed.  As devices become more complex (and you have indicated you have both hardware and software components in your devices), it becomes difficult to keep track of which Design Inputs have been addressed by Design Outputs, and in turn, which Design Inputs have been Verified (as in many cases a single Design Input can lead to more than one Design Output). Since the Design Validation process is intended to show that all User Needs have been fulfilled, you need to be able to provide that trace back so you can ensure that the objective evidence is available to provide the objective evidence that the Design Validation is complete. 

    In software, trace tools have become a best practice as well, in fact some indicate that is where they began, to provide the trace for the software lifecycle requirements, such as those of IEC 62304 or other lifecycle process standard.  Tools, such as IBM's DOORS and the Requisite Pro tool originated in software and have migrated to complete systems.  Since you have software and hardware in your device you probably need to consider it as a system to assure all User Needs and Design Inputs are met.  Those would include Regulatory Requirements and requirements from voluntary standards that you claim compliance.

    I can point to a requirement for traceability that is in the risk management standard ISO 14971:2007 (and EN ISO 14971:2012) in Clause 3.5.  It will also be part of the new edition of the standard to be released late this year.  That requirement states you "shall provide traceability for each identified hazard to the risk analysis, the risk evaluation, the implementation and verification of risk control measures, the assessment of the acceptability of any residual risk(s).  The bolded part of the requirement is part of the Design Verification and Design Validation activities of Design Controls.  So in this case a traceability requirement, while not part of a regulation, per se, is a part of a standard normally used to meet a regulatory requirement.  

    Hope this helps.

    ------------------------------
    Edwin Bills MEd, CQA, RAC, BSc, CQE, ASQ
    Principal Consultant
    Overland Park KS
    United States
    elb@edwinbillsconsultant.com
    ------------------------------



  • 4.  RE: FDA Traceability Question

    Posted 24-Apr-2019 13:51
    There is a traceability requirement in 820.65 which I think is closely linked to the identification requirement in 820.60.  I have a suspicion that both of these should be linked to design traceability and are not, but I'm still working on that.

    Questions:

    When you say (2nd sentence) "it allows you to provide objective evidence..." what is "it."?

    How do you interpret 820.30(g), "Design validation shall ensure that devices conform to defined user needs," so that it is not a requirement for traceability?

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



  • 5.  RE: FDA Traceability Question

    This message was posted by a user wishing to remain anonymous
    Posted 24-Apr-2019 15:21
    This message was posted by a user wishing to remain anonymous

    There is a traceability requirement in 820.65 which I think is closely linked to the identification requirement in 820.60


    This may be in reference to traceability of the product, for example, in the case of a customer complaint. Trace the lot number to the the components to do the complaint investigation.


  • 6.  RE: FDA Traceability Question

    Posted 24-Apr-2019 15:27
    Edited by Julie Omohundro 24-Apr-2019 15:32
    Yes. I think I'm saying that I suspect there should be a link between the design and the product, and therefore between traceability of the two.

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



  • 7.  RE: FDA Traceability Question

    Posted 24-Apr-2019 17:57
    That is correct, there is a traceability requirement in 21 CFR 820.65, but it does not relate to design, as implicated in the original question.  820.65 relates to a device that has been produced and is "intended for surgical implant into the body or support or sustain life and whose failure to perform when properly used in accordance with instructions for use".  The requirement is for procedures to facilitate corrective action.

    There is no written requirement for design traceability.  The Design Controls do require Design Verification showing that Design Outputs meet Design Inputs, and Design Validation that demonstrates the device, including software, as designed, meets user needs.  A traceability tool is a best practice to demonstrate these requirements can be met, but is not a requirement.

    ------------------------------
    Edwin Bills MEd, CQA, RAC, BSc, CQE, ASQ
    Principal Consultant
    Overland Park KS
    United States
    elb@edwinbillsconsultant.com
    ------------------------------



  • 8.  RE: FDA Traceability Question

    Posted 24-Apr-2019 10:51
    The theory of design traceability between software and hardware is the same. The inputs flow into your outputs (specifications) which then need to be validated as meeting the requirements. You're documenting all the design requirements (user, business, cost) as the inputs. All the inputs should have an importance level associated with them from nice-to-have to critical. From there, you should have a design outputs document that addresses if and how all inputs were manifested in the outputs. My suggestion would be to keep the software and hardware traceability separate as the details are usually quite different. I wouldn't think the matrix would look too different from a software one but probably fewer columns. Have you looked ISO 13485:2016 for how they talk about D&D? The documents that contain the information listed there would be the traceability if you didn't have a standalone trace matrix.

    ------------------------------
    Maddi Myers
    Regulatory and Quality Project Manager
    Edina MN
    United States
    ------------------------------



  • 9.  RE: FDA Traceability Question

    Posted 24-Apr-2019 13:40
    Maddi, nice summary.  I have one question.   When you say "design traceability between software and hardware," this seems to conflict with your suggestion to keep them separate.  Could you clarify?

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



  • 10.  RE: FDA Traceability Question

    Posted 24-Apr-2019 15:37
    Hi Julie,

    Thanks for the question. The key word in that sentence is "theory". At a conceptual level, they are the same when considering inputs, outputs, validation and verification. However, once you get into actually creating the documentation, there are a lot of differences in jargon and logisitics which would make combining the documentation difficult. Also, I like Edwin's suggestion about using traceability tools for software. I've worked on software projects where they were not used and it makes it very difficult to organize a valuable traceability matrix. 


    ------------------------------
    Maddi Myers
    Regulatory and Quality Project Manager
    Edina MN
    United States
    ------------------------------



  • 11.  RE: FDA Traceability Question

    Posted 24-Apr-2019 21:52
    Thanks, Maddi.  Software everything is always a mental stretch for me.

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



  • 12.  RE: FDA Traceability Question

    Posted 25-Apr-2019 08:25
    Thanks for the great responses everyone.

    ------------------------------
    Karen Zhou
    ------------------------------