Having just read the paper in question, it's not obvious to me that we have clean data. I don't see that the GF was downloaded from a computer, but rather that it was calculated somehow from a) analysis of some computer download, or b) computed from reported dive characteristics. I just don't know. It says that 39,000 dives were digitally recorded, so I presume there's some sort of computer download.
But let's assume that the GF data is valid. Here's what they say:
To which my response is, "Wait! What!?"
A "deserved" hit is one whose GF is over 1 and all the others are undeserved ??? Not in my world.
If I look over the dives I do (including deco), the vast majority end with a surfacing GF<50, which I presume is the same as their GF 0.5. And every deco dive ends with a GF <70. If I'm wrong about their terminology, somebody correct me.
But if that's true, their whole database of DCS cases is skewed toward higher surfacing GF's than my dives, and all this population reflects is the distribution of GF's among bent divers. "Only 8%" has GF's over 1 because damned few dives end that way. If it's 8% of your sample DCS population when it's 1% of dives, I'd say that's a pretty definitive association. But we don't have a denominator.
And the plurality of DCS cases occurring in the GF70-90 range tells me that they're the majority of dives that are pushing my envelope. To me, the "undeserved" hits (which are probably just missing some key additional data) are the DCS cases occurring at GF<70 (14.6%).
Those are the ones I'd want to look at.
I am not overly impressed with this study, though the conclusions make some sense.
Am I not understanding how they are measuring GF?