O'Dowd-Crawley early intervention study

Exactly. And writing up the circumstances surrounding the non-publication based on the public record and statements also does not depend on their cooperation or permission.

 

But we don't have their full write-up of the protocol or methods either. Can we compel them to release whatever draft report they did submit to journals?

 

Protocols still aren't pre-registered on record precisely to avoid this scenario?

Prior edits, sure, but the registered protocol submitted for public funding? If this were serious research anyway...

 

If this were "serious" research, they'd only publish the protocol once the main study had finished!

 

This is an interesting suggestion. In days gone by I would not have doubted it for a minute. But things are changing and one could argue that nobody owns data if it is in the public domain and is of public interest in terms of medical implications.

I think this would be a very interesting test case. Clearly the Wilshire et al paper has pushed the boundaries a bit already. I suspect that there are editors around prepared to help push a bit further.

The study should be published because there are inferences to draw and it involved patients. If the investigators cannot persuade an editor to publish then others should be entitled to try.

 

So we have the full protocol. What would be "private" then? If it's already published? Sounds like everything that's needed besides the data (and someone wanting to do the actual work of deciphering weird cryptic non-specific nonsense and mathemagics).

 

Fantastic! And apt.

 

Would this be the first time that third parties have published a report on someone else's study that they tried to hide in their bottom drawer?

 

I can't find the appendix mentioned in the protocol though... It would be useful to know what was in the Information Booklet.

 

Surely they would assume that this small band of militant activists who do nasty things like FOI requests and say mean things to/about psychiatrists would be in close communication though.

 

There frankly isn't much usable data. This is low effort even by Chalder's standards. It seems they are just playing with spreadsheets and copy-pasting the same work over and over again, doing roughly the same analysis despite there being no value in it.

Most of the variables are useless, basic psychometric questionnaires and basically the same set as in every other Chalder research. There are only two partially relevant data points: SF-36 and hours missed from work. The latter is hard to interpret as it's compared to what? Are they all people who used to work 40h normal weeks before? Few actual changes at 6 months. It's the only objective measurement and somehow it manages to be mostly subjective and open to interpretation. All but 5 rated as having missed fewer than 10h in the past 7 days, with most actually rating 0. There is an outlier that "missed" 44h of work at baseline and "missed" 56h of work at 6 months. So I guess someone working 56h week having dropped to 12 then 0? Really not sure how to interpret that.

SF-36 is all over the place, from 5 to 100 (the 5 did not submit at 6 months). There are 3/44 at 100 and 8/44 >=85. It doesn't look to be a valid cohort to begin with, I don't know how anyone can both have a diagnosis of CFS and rate 95+ on SF-36. Because of missing data, somehow average of both columns (baseline and @ 6 months) add up to the exact same number (61.7045). Somehow removing entries that did not report at 6 months actually yields another exact tie (this time at 63.472). This is peak GIGO.

There is one data point that I can't understand. It's wpai_q1-2_: "number in paid work at baseline/6 months". It's either a 1 or a 2 but there is no explanation what either stands for. Number of what?

The only value I could see in publishing an analysis would be in just how incredibly bad and useless this kind of research is. This is an embarrassing waste of public funds and everyone involved should be ashamed of themselves, if they were capable of shame anyway.