Question 15 from the first paper of 2004 and Question 11 from the first paper of 2003 asked the candidates about prone ventilation. Unfortunately, these questions preceded the wildly awesome PROSEVA trial. Contemporary studies which those college answers were based on were largely negative owing to weird patient selection. The best resource to revise this topic is the relevant LITFL CCC page. Of academic literature, a good article to read about the mechanisms of improved oxygenation from prone positioning was "Pragmatics of Prone Ventilation" by Messerole et al. (2002).
LITFL have an excellent literature summary section on their prone ventilation page, which has been pillaged for references. In brief, three contemporary studies (Gattinoni, Beuret and Guerin) were available to the trainees at the time of writing Question 15 from the first paper of 2004 and Question 11 from the first paper of 2003, in addition to early pioneering work by champions of "extreme positioning".
Piehl et al (1976) were the first to play with prone position: their study had 5 patients in it.
Gattinoni et al (2001): multicentre RCT - 304 patients with ARDS
Beuret (2002): single centre RCT - 54 patients with coma (not ARDS)
Guerin (2004): multicentre RCT - 791 patients with acute respiratory failure
On the basis of the available evidence, the college answer to the 2004 question mentioned the 2001 Gattinoni paper, lamenting the negative data. The college answer was fairly optimistic; they quoted an opinion piece by Alain Broccard from 2003, who also felt that there were "good reasons not to regard the recent negative prone positioning study as indicating that the prone position is of no interest". Broccard pointed out that it took five RCTs to finally arrive at the correct conclusion that low tidal volume ventilation improved survival in ARDS. This was prescient. Many trials followed.
Sud et al (2010) collected the data into a meta-analysis (n=1,867).
PROSEVA (2013); multicentre RCT - 466 patients with severe ARDS