Sorry to ask a question that may be off-topic. But we’re working on a study project where we need to collect patient data. It’s not a question of encrypting the database, but rather the tables, i.e. directly identifying columns. So far, I haven’t found a solution on any of the redcap communities. I’m not necessarily looking for a solution, but for solutions and feedback, if possible via R or Python APIs.
Best practice is to use the
REDCap (Research Data Capture) API for fetching data securely via an API key that the
REDCap server provides to you. A workflow for this in R. is here when you click on the
REDCap tab. I will be updating that to reflect the latest changes in the R
REDCap API functions.
Next, consider this workflow for encrypting your analysis datasets, relying on several nice R package including the wonderful
qs package for efficient binary R object file storage.
Taken together, these workflows ensure that unencrypted personal identifiers exist only briefly on temporary files that are not in your project directory.
Thank you @f2harrell for your reply. But we have a study where we store data for years. So we need to encrypt the data inside Redcap, not outside then import. The idea is to find a process that encrypts the identifying data in Redcap as soon as it is entered. I’m looking at the links you’ve provided, but I’m not sure they’ll answer my concerns.
That’s really a misunderstanding of
REDCap and all of its built-in security IMHO.