Lowtimer's analysis of the leaflet is very good - this is one of the worst documents (or at least the worst paragraph of a document) I've ever seen come out of the CAA. I'm not saying it's wrong, without seeing the underlying numbers I don't know (although I know where my suspicions would lie).
I would concur. The underlying numbers are available in CAP667. There were 13 accidents of the total of 166 that involved loss of control in IMC. "More than 3/4 had no IR or IMC rating (one had an IR)". I would suggest that the most likely distribution would be:
1 IR
2 IMC rating
10 no IR or IMC rating
It's not possible to determine from CAP667 which accidents involved the IMC rated pilots and, even if it were, the reports are not available online.
But the claimed inference in SSL23 is, in scientific terminology, utter bollocks.