That article's definition of "lab accident" seems narrow and legalistic to me. In either case, the virus spent 1950-1977 in a lab freezer. It ended up in the wild, with ~700k people dead. The only question is whether it escaped in an infected researcher (or in infectious lab waste, or in whatever else you'd consider a proper lab accident) vs. in that failed vaccine candidate.

Those details do inform some details of the correct policy response. For example, they determine the relative importance of better PPE at the bench vs. better QA before allowing the vaccine to leave the lab. They don't change the overall question of whether scientific research has ever caused a pandemic, though. That causality is what matters, not whether the sign on the door said "lab" vs. "experimental vaccine nurse".

For example, if the pandemic originated from a WIV researcher who became infected in the field (during their many expeditions to remote bat caves that no other humans would routinely enter), was that a "lab leak"? Literally no, since they weren't in the lab. The causality would still be the same, though--if not for that scientific research, that virus would likely have never left the cave.

To avoid such confusion, it's probably better to say something like "unnatural origin", or "origin arising from scientific research". A much bigger mouthful than "lab leak", though.

I'd add that the article does not state that there have never been cases of accidental releases of pathogens from laboratories, only that such accidents had likely not led to a 'global epidemic' as of the date the article was written (2015).

The article's abstract opens with the statement 'The 1977-1978 influenza epidemic was probably not a natural event'.