Please please do!
If you could get them to confront the problem of unblinded trials with subjective primary outcome measures and make a policy decision not to include any such trials in Cochrane reviews that would be amazing.
The trouble I see with this is that this would mean they could not include any of the trials of CBT and other therapies, since by the nature of the treatment they are unblinded, and by the nature of the conditions usually treated (anxiety and depression), there are no objective things to measure.
If you could at least get them to exclude trials of psychological treatments for conditions such as ME and the other physical conditions dumped in the MUS dustbin where there are objective outcome measures possible, but none have been used, that would be a huge step forward.
This would mean, for example, reanalysing the effectiveness of the PACE trial using only the 6 minute walk test (taking into account that lots didn't do it), step test and employment data, and excluding all subjective measures. It might also force the likes of Esther Crawley to start using actometers. Still doesn't get around all the other problems with PACE like whether the participants actually had ME and whether they complied with treatments though. I don't really think PACE can be rescued and provide any meaningful data, apart from the fact that despite their best efforts to get the result they wanted, they inadvertently managed to show their treatments make no long term difference.
You point out that part of the problem lies in peer review. The problem is, of course, that the BPS cabal all work together to approve each others trials and review each others papers. And funding bodies are in their pockets too. In a sense Cochrane is the summit of a stinking pile of bad science right up from the foundation of bad trial design. Until that is sorted out, nothing will change.
Sorry, you know all this already. Just letting off steam!
I believe there should be something more formalised regarding the design and running of clinical trials, and maybe a professional specialism in this regard. By formalised I do
not mean rigidly set in stone, because I fully appreciate that if you try to clone the management and methodologies of trials, you might seriously hinder the inventiveness needed. But if a specialism existed, then appreciating that very need to allow free thinking and flexibility would be part of the skill set. (I have your comments in mind
@Jonathan Edwards from a few days back regarding this). Whilst at the same time fully understanding the fundamentals, the rules that should never be broken, as well as the rules that maybe don't apply in certain circumstances.
To me it feels that maybe every trial should include such a role, perhaps mandatory, and that someone couldn't just self-elect themselves into that role but have qualifications of some kind for it. Such a role would be highly skilled, and would in many ways cut across medical specialism, albeit probably need a sound understanding of them.
It's a bit analogous to safety engineers and safety engineering. When this discipline first came into being, many other engineers no doubt thought it very over the top. But you really would not want to be flying in modern aircraft, or driving modern cars, or having x-rays done, or using all manner of other modern equipment, without safety engineers (who will be degree-qualified in that discipline) being part of the design teams for such things. Safety engineering cuts across other disciplines, and a safety engineer has a lot of authority within a design team, and if the safety engineering aspects are not up to scratch then it's a no-go. There is of course considerable discussion and debate within the whole team regarding safety, guided and managed by the safety engineer - it's not a draconian exercise, but collaborative. Wherever there is significant potential risk to human life or the environment, these folks apply their skills, science and tools to identify the risks and manage them, again with considerable involvement and collaboration of the rest of the team. I can't help thinking something similar applies to clinical trials, where human life is a key facet.
If all trials were run by teams with the competence and will to design and run them well, then such roles would never be needed. But the same could be said for safety engineers and teams designing airliners. There will always be trials lacking the competence and/or will to run them well, and as sure as eggs are eggs, some will go dreadfully wrong - Sod's law.
I have to say, I found it really shocking once I realised this sort of "safety control" does not exist for clinical trials, where human safety is so paramount, not just during the trial, but from the consequences of their reported outcomes. Especially when I then realised something like PACE so tangibly proves the need for such control. I find it mind boggling and horrifying really.