This “FDA GPSV Traceability Expectations” post is only available to Premium subscribers. See our Subscribe page for information on subscriptions.
Going way back to the late 1990’s, FDA had an expectation that safe and effective software would require a well thought out development lifecycle that includes many activities designed to ensure the correctness and robustness of all software that was part of the medical device. A key guidance document was created known as the General Principles of Software Validation, or GPSV for short. One method that one will find in the GPSV is the use of traceability analyses for several verification activities. FDA GPSV traceability expectations include at least twelve traceability analyses. This could vary depending upon the risk profile of the medical device and other regulatory factors. Theses trace analyses are illustrated in the diagram below:
FDA GPSV Traceability Expectations Diagram
As you can see, this does not perfectly align with IEC 62304 so one should carefully consider and plan how traceability analysis will be used in the SDLC process for products marketed in the US to better fulfill FDA GPSV traceability expectations.
Traceability analysis can be a powerful tool to help gain confidence in the quality of a software program. It is just one verification method that one should use. The V-model diagram illustrates how verification activities essentially trace back to the “flowdown” portion of requirements and specifications. See our post on the V-model here: Enhanced V Model Diagram
