"Opening a can of data-sharing worms" (General, not specific to any specific medical research)

[Dolphin]

"Opening a can of data-sharing worms"

http://blogs.scientificamerican.com/absolutely-maybe/2013/09/10/opening-a-can-of-data-sharing-worms/

A blog post on data sharing, from a conference for medical editors.

 

[Dolphin]

It talks about the difficulty of getting data from trials, with only 7% of it accessible from 1991.

This, it could be argued, is possibly one reason why it is important that data be given out with Freedom of Information Act requests; if it's not released now, it may never be released.

 

[Dolphin]

“We’ve been too industry-focused: academics are resisting this [sharing their data] too”

I do get frustrated when the discussions only focus on industry.

 

[Dolphin]

Dickersin was concerned that researchers have to “become detectives” to find out treatment effects and resolve discrepancies between different sources of data: “There needs to be more scrutiny by regulators.”

Unfortunately with non-pharmacological interventions like CBT and GET, there is little if any regulation.

 

[Esther12]

It talks about the difficulty of getting data from trials, with only 7% of it accessible from 1991.

This, it could be argued, is possibly one reason why it is important that data be given out with Freedom of Information Act requests; if it's not released now, it may never be released.

Ta D.

At the very least, data should be backed up at shared centres in a standard format. Crazy to have this stuff slipping through our fingers.

 

[Dolphin]

Ta D.

At the very least, data should be backed up at shared centres in a standard format. Crazy to have this stuff slipping through our fingers.

Yes, although would still need to be maintained I imagine with software changing over time.

 

[user9876]

Yes, although would still need to be maintained I imagine with software changing over time.

There is a whole archive industry which aims at keeping data secure, searchable and available (with growth driven by the e-discovery rules in the US). If universities cannot run systems with professional data security and data retention policies then they shouldn't be allowed to run trials. I say that from the perspective of both funding and the ethics committee approving a trial. Patients agree to take part in a trial to help move treatments on - to not properly store data to back up results is breaking the trust of patients.

I remember there was a series of letters where someone points out an issue with stats in one of Wessely's papers and his response was oh the data has been lost due to a computer theft. This is just not an acceptable answer for a professional research organization where data should be kept secure, computers physically secured and offsite backups kept. I have talked to a few people working at university IT departments who basically have a huge problem because academics want to do their own thing and don't use the services offered.

The data they have is not large and can be archived in comma separated lists which is about as basic a format as is possible. I think scanned copies of all patient survey data should also be kept (as pdfs) which is a common archive format and I believe that there are processes in place to preserve the format if Adobe ever go bankrupt.

Drug companies are very aware of the data archive problems and have appropriate systems in place.