Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
I wish to use raw data about clinical trials for the end year examination of my master students. These data could deal with any kind of molecule as long as the trials were complete (Phase 1 to 4).
Do you have an idea where to find such a free dataset on the web ?
You can have a look at the Clinical Trials Network, where data are available in CDISC format. You must agree to their terms and conditions, although the following point might be of concern for teaching purpose:
To retain control over the received data, and not to transfer any portion of the received data, with or without charge, to any other entity or individual
(Anyway, I think that you can just send an email to the contact support to check that their data can be used for teaching.)
The NIDKK Data Repository is specifically concerned with studies on kidney and liver disease, and diabete; however, you have to submit an application.
Otherwise, perhaps the ADNI project, which aims at characterizing change in cognitive functions and brain structures with age, with a particular emphasis on Alzheimer's disease and neuroimaging, might be interesting. This is not a clinical trial, but available data include: demographics, clinical and cognitive data, neuroimaging (MRI/PET) data. Details about protocols and data can be found under ADNI Scientist's Home, and data are available on ADCS website.
There doesn't seem to be anything related to RCTs on http://www.infochimps.com/. However, I remember having seen some clinical data used with the Weka software, as e.g. on this page: Data mining to predict patient outcome in a clinical trial of a lung cancer treatment.
It looks like NIMH has a number of slightly lobotomized data sets available, although I have no idea what the data use agreements entail. This will be a problem for almost every trial you come across I suspect - the need to either prevent their trial results from getting scooped, health information privacy concerns, or both.
Others I've found:
I'll edit this answer if the list grows. There's also always simulating a data set so it matches the results of a published trial.
You can find details of a trial downloadable in an XML format from http://clinicaltrials.gov/
Check out http://yoda.yale.edu/ and https://www.clinicalstudydatarequest.com/ both of which require lengthy applications.
