I was the IO for a Class A where the mechanism of failure was completely mechanical. During the 4 months of the investigation (including several "pauses" for technical analysis), there was never a HF person assigned to the SIB, nor was one ever volunteered. When I sent the draft Tab T to AFSEC to review about a week out from the outbrief to COMACC, I got back "but you need an HFAC in there" and the ACC HFACS chief was hastily summoned to assist. Since this person had not spent the last 4 months on the SIB like the rest of us, it was ridiculous to expect any useful input at that point.
There is an HFAC for Acquisition/Design that we ultimately ended up throwing in there to appease the AFSEC QC review. The MOFE process then removed this as a causal factor 90 days later due to lack of evidence to support the claim. I'm so glad we modified our Tab T for that end result.
In between convenings of the Class A SIB, I was the SIO of a Class B. The causal findings revolved around a mechanic and a QA supervisor at an overhaul facility failing to ensure a bolt inside a component was properly torqued. The time between the overhaul and the mishap was 4 years, so it's not like I could get any useful info from them; he mechanic no longer worked for the company anyway and I had no way to track him down. AFSEC wanted me to invent HFACS to assign to these two individuals with no shred of evidence (they QC rejected my final message the day my Class A SIB reconvened for the final push to the finish line). Luckily my NAF (convening authority for the Class B) went to bat for me and we were able to get the final supplemental pushed through.
AFSEC also told me I had too many acronyms in my one-liner (insert eye roll here). Well if you only give me 40 fucking characters, I have to abbreviate somewhere!
I think the QC process at AFSEC is broken. It should occur BEFORE the SIB deconvenes. What's the purpose of having an AFSEC rep (telephonic or in person) if their review and input to the SIB/SIO isn't good enough to pass muster?
If every final message needs an HFAC, which is definitely the vibe I have received in the last several years from AFSEC, then the IOs need to be better trained on HFACS, and the convening authorities need to provide an HF consultant to every. single. investigation. The problem right now is that IOs are being forced to simply invent a link to an HFAC when one might not be present, thus skewing the data over time.