Managing Blinded Data
For additional information on data blinding, see "Security for Blinded Data" in the Oracle Life Sciences Data Hub Implementation Guide.
You may want to hide, or blind, treatment codes or other information that would reveal which patients were receiving which treatments.
Oracle LSH supports data blinding in:
-
the Oracle LSH Table instances you specify
-
all reports and other outputs generated using one or more blinded Table instances as a source
-
all Table instances downstream in the data flow from a blinded Table instance: If a Program instance that reads from a blinded or unblinded Table instance attempts to write data to a nonblinded target Table instance, the submission fails—unless the target Table instance is explicitly authorized to accept data from such a Program and a user with Blind Break privileges explicitly confirms that the Program can be executed.
All Table instances have a Blinding flag attribute that indicates whether or not they may contain data that is sensitive and must be blinded at some point in time.
If a Table instance's Blinding Flag is set to Yes, then Oracle LSH maintains two partitioned sets of rows for the Table instance: one set of rows of real data and one set of rows of dummy data. Programs that run on data in these Table instances operate on only one set of data at a time: either the real data or the dummy data.
Table instances also have a Blinding Status attribute. If a Table instance's Blinding flag is set to Yes, its Blinding Status can be either Blinded or Unblinded to indicate the current state of the data. If a Table instance's Blinding flag is set to No, its Blinding Status can be either Not Applicable (the default) or Authorized.
This section includes the following topics: