(This article was originally published at Statistical Modeling, Causal Inference, and Social Science, and syndicated at StatsBlogs.)
James Coyne writes:
For those of you who have not heard of the struggle for release of the data from the publicly funded PACE trial of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome, you can access my [Coyne’s] initial call for release of the portion of the data from the trial published in PLOS One.
Despite the investigators’ having promised the data would be available as a condition of publishing in PLOS One, 18 months after making my request, I still have not received the data and the PACE investigators are continuing to falsely advertise to readers of their PLOS One article that they have complied with the data-sharing policy.
Here’s what the Pace team wrote:
We support the use of PACE trial data for additional, ethically approved research, with justified scientific objectives, and a pre-specified analysis plan. We prefer to collaborate directly with other researchers. On occasion, we may provide data without direct collaboration, if mutually agreed. . . .
Applicants should state the purpose of their request, their objectives; qualifications and suitability to do the study, data required, precise analytic plans, and plans for outputs. See published protocol for details of data collected.
Dayum. Them’s pretty strong conditions, considering that, as Coyne points out, the original Pace trial didn’t follow a pre-specified analysis plan itself.
Data will be provided with personal identifiers removed. Applicants must agree not to use the data to identify individual patients, unless this is a pre-specified purpose for record linkage.
Individual researchers who will see the data must sign an agreement to protect the confidentiality of the data and keep it secure.
This seems fair enough, but I don’t see that it has anything to do with pre-analysis plans, qualifications and suitability, or direct collaboration. The issue is that there are questions with the published analyses (see, for example, here), and to resolve these questions it would help to work with the original data.
Putting roadblocks in the way of data sharing, that seems a bit vexatious to me.
P.S. There’s nothing more fun than a story with good guys and bad guys, and it’s easy enough to put the Pace struggle into that framework. But, having talked with people on both sides, I feel like lots of people are trying their best here, and that some of these problems are caused by the outmoded statistical attitude that the role of a study is to prove that a treatment “works” or “does not work.” Variation is all. I don’t know how much can be learned by reanalysis of these particular data, but it does seem like it would be a good idea for the data to be shared as broadly as possible.
The post Whassup, Pace investigators? You’re still hiding your data. C’mon dudes, loosen up. We’re getting chronic fatigue waiting for you already! appeared first on Statistical Modeling, Causal Inference, and Social Science.
Please comment on the article here: Statistical Modeling, Causal Inference, and Social Science