DHF, Risk Management, Use Engineering

One of the participant who attended the 6-Hour Virtual Seminar on The DHF, DMR, DHR, EU MDR Technical Documentation Similarities, Differences and The Future asked:

I would like to ask what I need to do for legacy medical devices (FDA Class 2).

My company has 510(K) clearance back in 2000. Since most of the requirements happened post 2000, may I know what I should for the legacy medical device related to:

• DHF (should I remediate it?) – some of the info may not be available (i.e., design review/meeting minutes/decision, formal approval (no proper documents control before), other validation records)
• ANS:  Where the DHF was complete in 2000 . it does not need remediation.  Areas of incompleteness can be added by researching old documentation, interviews, lab books, etc.  and added (not backdated) to the DHF with explanation. Known missing data can also be stated and a document / memo to file added (actually or as an addendum).  Subsequent changes are addressed in the DHF if your company keeps it open / controlled, but as I mentioned in the webinar, I don't recommend that.  I recommend changes controlled by 1) a new DHF if extensive, 2) an addendum to the old DHF if extensive, or 3) use the CGMP Change Order system, 820.40(b).  In all cases, a change, single or cumulative, must be evaluated / documented, as to the need to file a new 510(k). Remember to view the DHF through both regulatory and IP (intellectual property) "eyes".

• Risk Management (RM) – DHF has been closed and now tracked under DMR – do I need to go back to update RM (per latest standard) during design stage which has been closed? Or update incremental to the latest standards? Or it’s OK to meet RM requirements at the time of design stage & no further work required (perhaps only periodical review post-market)?
• ANS:  Although RM should be done as part of the Design (Design Control, 820.30, ISO 13485  7.3) process. since RM drives all device decisions throughout its lifecycle, the RM File must be a living / controlled document, updated as new applicable information becomes available (through CAPA, V&V, industry data, annual quality review, etc.). That's why I recommend in the webinar that the RM File and Use Eng'g File (if any) have a non-controlled copy in the DHF (or a pointer to it/ them), of the version used during the design phase, prior to Design Transfer , and the actual RM (UE) Files be active and controlled (change controlled). The new version of ISO 14971 adds the need to add systemic RM considerations to the QMS.  Any change in emphasis re: Device (not QMS) RM based on the new 14971 rev could be addressed during one of those reviews / file updates. 

• IEC62366 – As it comes after 2000 which was not done before during 510(K) approval, do I still need to do it if no major changes to medical devices which have been shipped to market for ~20 years? I come across User Interface of Unknown Provenance (UOUP), what’s the minimum efforts that I need to take?
• ANS:  You as a company need to decide based on novelty of your device and any user interface concerns that are still applicable 20 years later.  Human factors was a concern with the FDA in 2000, when they starting publishing documents on it.  If your product / family has minimal field problems due to design / interface issues, I personally haven't seen regulatory agencies raise an issue about it.
  
• ANS:  If the use interface falls under UOUP, you should consider all 9 stages of 62366-1, and revisit those that don't appear to be addressed, and/or pose a high risk to the end user / patient; and document this evaluation - basically a Gap analysis. Some devices are so obvious as to use (or are subject to med school, et al, training) that a UE analysis may not be justified, e.g., standard needles. Address in your applicable SOPs, and by a written rationale / letter to file.

--  jel@jelincoln.com