Meta-analyses and megatrials: neither is the infallible, universal standard. Issue 2 (23rd April 2004)