Excellent analysis. My experience over several decades of 510(k) submissions, is that the process and documentation demands have become much more rigorous, with more data required. The 20 hour average cited in your article, if based on actual FDA averages must be considered in light of the fact that the reviewer starts and stops the "clock" when requesting additional information from the submitter. Not exactly like a lawyer bills a client. And that average must also be considered in light of the fact that the majority of 510(k)s are basically "me too" products, i.e., the whole 510(k) process is predicated upon the requirement to identify a "predicate" device, and then prove "substantial equivalence" to it. The principle being that the new device poses basically the same risks (and benefits) as the predicate. and is cleared on that basis. Yes some 510(k) products push the envelope, but then the Agency usually requires an IDE (Investigational Device Exemption) submission leading to clinical trials prior to the submission of the 510(k). I very much agree that today's requirements are much more extensive than those existing in 1982. The e-copy requirement is not only for filing purposes, but to allow the review to involve others in the Agency when questions of safety and efficacy arise. Requirements such as Design Control, started in 1996-7, device risk management per ISO 14971, and human factors engineering / usability engineering per IEC 62366-1:2015, and -2:2016, and similar, have also been incrementally added as appropriate. And a growing list of software / hardware requirements. New guidance documents, which state that they are voluntary, but I have found to be required if applicable to a submission, e.g., the documentation required for devices containing software, cybersecurity, et al. The sheer volume of documentation required in a submission between 1982 and 2017, has grown exponentially as a result, and my experience is that it is all given careful study based on the follow-up questions I receive from the reviewer(s).