I read in another post regarding Bernard Lagat that someone speaks about doping because "1:46 is to slow for running 3:26".
There are many reasons for PB in different event that seem not to have any relation with the best PB.
For example, the PB of Lagat in 800m was at the beginning of the season (also with 13:19 in 5k), when his real value could be about 3:34 in 1500, and he never ran 800 at the top of his shape.
I want to remember some "strange" personal best in big champions of the past.
SYDNEY MAREE had a PB of 1:48.11 (82), when was able running 3:32.12and 13:32.40 (and already there were no relations among 1:48 and 3:32), then in 85 ran 3:29.77 and 13:01.15. Also his PB in 10000 (28:21.46 in '80) is ridicolous if compared with 5000.
HENRY MARSH, the same year that he ran 8:09.17 on steeple (1985), ran 3000 in 7:56.2 : of course, no relation. Also his PB in 5000 (13:45.2 in '84) is something wrong if you think about 8:09 in steeple.
STEVE CRAM ran 49.1 in 400m in 1982. In the same year he ran 1:44.45 and 3:33.66. Doyou think possible running 1:44 if you value is 49 in 400 ?
STEPHEN CHERONO has yet a PB in 3000 of 7:46. Last year, winning against El Guerrouj in 12:48.81 in Ostrava, the split of both the athlete was 7:37, 9.0 better than his PB.
HICHAM EL GUERROUJ himself has an official PB in 800m of 1:47.8, and, before running 5000 last year, had a PB in 5000 of 13:46.79 in 1992 (only 18, before running 1500).
How can you consider these data ?
So, what I want to say, is that the most part of times PB in secondary events are not significant, and is not possible to have precise indications from them.
Expecially if you use these data for speaking about drug.