E-DRUG: Need for a protocol repository
--------------------------------------
Dear WB and others,
Have found this service very useful re. some HIV work I was doing. On the more general question of availability and transparency of evidence (re. the recent Vioxx recall, etc) I'd like to point out that trial registries are not enough. The disparity between the analyses reported and what may or may not have been the protocol pre-specified analyses is the root of many of the problems with both CLASS and VIGOR, and there is no reliable way of
achieving clarity on these issues without the availability of the protocol itself, in its entirety, including the critical statistical analysis plan which is often only finalized just before data-unlocking.
These issues lead us to argue in a 2002 Lancet editorial (Lassere M, Johnson K. Power of the Protocol. Vol 360:1620-2) for a protocol repository which would be designed to protect intellectual property concerns yet provide for eventual full disclosure to the journal writers and editors, and then to the public generally.
Kent Johnson, Marissa Lassere
Newcastle University
Australia
Kent.Johnson@newcastle.edu.au