I found a presentation that was put together by BioTechLogic to address this issue ( http://c.ymcdn.com/sites/www.casss.org/resource/resmgr/2016_CMCS_TreDenickTracy.pdf). It seems like it is a very involved process and I've just listed the major steps below:
1. Quality System Gap Assessment
2. CAPA
3. Form Cross Functional Team
4. Develop “Device Family” Bracketing Strategy
5. Update Policies and Create SOPs
6. Prepare Design and Development Plan and Define High Level Milestones
7. Create Design History File and Index
8. Create User-Needs Requirements Document
9. Compile and Conduct Risk Analyses
10. Prepare Design Input and Output Documents
11. Compile Historical Design Control Verifications
12. Verify Proper Design Transfer
13. Review Change Controls for Design Changes
14. Conduct Design Review/Verification Meetings per 820.30(e)
15. Conduct Risk-Based Remediation(s)
16. Prepare Design Verification Traceability Matrix
17. Final Plan and Close DHF
With respect to recreating the documentation, companies never work without any documentation. Whatever limited paperwork they may have should serve as basis for creating the DHF. "A compilation of documents/records necessary to demonstrate that the design was verified in accordance with the approved plan and specified requirements."

Fady,
This is a very interesting and also important subject matter. To answer your question, even products that were designed and built pre 21 CFR, has to have some sort of design and technical documentation somewhere, and especially so if these products were patented. They won't be as completes as the DHF, however, we now have a starting point. Once all these documents are identified, the next step would be to compare these docs and against the design control requirements and find the gaps. Once the gaps has been identified, the real work would begin in closing these gap and bringing all parts of control designs up to date. This will be done by the information and timeline derived from the existing documentation. The difficulty of the process will depends on the level of details and documentation that the company has retained for the specific product.

Thanks
Adrian

Fady,

That is an interesting point to bring up. With combination products, if I understand correctly, the company must identify first the necessary center for the primary use of the product. Once it is designated into CDER, CBER, or CDRH then that would specify how the approval process would be for the product. Usually the pre-amendment device (product release before the act of 1976) tends to be the secondary product in these cases that still hold the substantial equivalence. If the device is class III, for instance, it would require a PMA or notice of completion of product development protocol (PDP). But if PMA is not required, then a 510K containing a Class III certification and summary is necessary. This provides details of the search done by company about the device type. Also, it provides verification of the safety and effectiveness for the devices and would be signed by the submitter.

http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/HowtoMarketYourDevice/PremarketSubmissions/PremarketNotification510k/ucm134578.htm
http://www.schiffandcompany.com/backupFiles/fusra-6ed-201-215.pdf

Good Discussion so far. A legacy device is one that incorporated a drug before combination product regulations arrived. Many medical device manufacturers have mistakenly assumed that since legacy products were initially reviewed by Centers for Devices and Radiological Health (CDRH), but for combination products that is not true. Developing combination products requires expertise from engineers and pharmaceutical scientists to develop combination products. It is time consuming at the same time. Some combination products examples are,

Drug Delivery
• Implantable drug pumps
• Transdermal Patches
• Metered dose Inhalers
• Antimicrobial urinary catheters
• Eye drops for glaucoma

Orthopedic and Dental
• Antibiotic Bone cements
• Woundcare
• Antimicrobial dressings
• Steroid-containing Creams
• Antimicrobial impregnated Sutures

Vascular
• Drug coated balloon catheters
• Coated vascular prostheses

Source: https://www.medicalalley.org/media/105661/canopy-medical-thought-leadership.pdf

-Hetal

Personally, I think that this topic should be considered with the famous motto "If it ain't broke, don't fix it" . Obviously, these documents and the protocol to make all these devices safe and effective are necessary. However, if these legacy products have been used in hospitals, doctors offices, and over the counter for years already, and there hasn't been a reason for retraction or recall, it's not worth it to open up these cold cases. If it's working, let it work. Precious time should not be wasted on things that don't require the time. Especially with the healthcare system as out of whack as it is, paperwork time is precious because of the amount there is.

-John L.

I agree with you, John. As others have mentioned, companies that have developed legacy products must have some level of documentation on the design and manufacturing processes at minimum. I think that it may be worthwhile to compare their level of documentation with a DHF to identify gaps, but I think if the gaps are small and the history of the product usage has been favorable over a relatively long period of time, it may not be worthwhile to initiate more paperwork (as it would simply just cost the company resource and therefore money, when it already has actually demonstrated the quality of the product).

These products, although no longer under development, as you mentioned, are still being manufactured. If this is the case, then there exist documentation in order to create the product. I do not think it will be too difficult to use these documentation in order to create a DHF. At the end of the day, a DHF is a document that entails a list of requirements needed in order to build the device. Therefore, it is not difficult to figure out how the device is made if they are still being manufactured. They are just not developed and no changes are being made to them. Also, if no documentation exist on how to build the device (which is unlikely), reverse engineering the device would allow us to gain enough knowledge to write a DHF.

Legacy devices can be existing devices, modified hardware, and modified software. Legacy medical devices should conform to IEC 62366, a standard that specifies usability requirements for the development of medical devices. Having a summary of known use errors is a good start. A full use-related risk analysis should be performed. Additionally, primary operating functions should be documented. Application specification should also be documented. Performing all these steps can surely help in creating a DHF.

I also didn't think it would be difficult. There had to be some type of lab notebook or documentation process for the development part and for testing just pull the materials from manufacturing and go from there. I think the problem would be if it fails safety testing how can the company come back and fix it.

As some of you mentioned to retro spectically find as many document as possible and move forward from them but here is something which I would suggest as a plan to go for especially for legacy products to get approval.
So here is the flow. Correct me if wrong.

1. Get all documents available.
2. Perform Gap analysis.
3. Identify what data needed for fixing gaps.
4. Collect information as much as available to fill the gaps
5. Create design doc as per "Current" procedures and templates.
6. Perform reviews and approvals.
7. Note changes happened in the design
8. Documents changes as per procedures and version control
9. Perform reviews and approvals.

This plan covers most part of the documentation but things to looks out would be make a checklist of current document needed. Sometimes you wont needed a review and approval for some document like DMR, as the product is already being manufactured for a long time. so make a list of such document which doesn't require approval so that you don't waste time.