Using Batch Validation
The Batch Validation job, accessed from the Conduct menu by selecting Data Validation and then Batch Validation Session, checks data in a single study for all types of discrepancies. Most organizations set up Batch Validation to run at regular intervals, usually nightly.
Note:
Only alpha-numeric printable characters are supported by this feature. Non-printable characters, such as ASCII controls or graphical characters are not supported.
Which Data is Processed During Batch Validation
Batch Validation will run, including all phases of the "Batch Validation Process" outlined below for RDCM or RDCI's that are locked (if the study they are in is not Frozen). If you do attempt to run Batch Validation for a Frozen Study, these phases of the Batch Validation Process do not run for the Study.
Note:
The user that runs Batch Validation is the user shown in the audit trail of any associated updates made during this process. A user only requires privileges to run Batch Validation for a given Study for the updates outlined above to be completed; Privileged Update is not required
Batch validation always runs incrementally; that is, it runs on selected data based on changes that have occurred since the last Batch Validation, as follows:
-
If a patient has any new or changed data, Batch Validation runs all Procedures over all data for that patient.
-
If the definition of a Procedure has been changed, Batch Validation runs that Procedure on all data for all patients, even if patient data has not changed.
Note:
If a derivation Procedure definition has been changed, and other validation or derivation Procedures are dependent on the changed Procedure (they use the value it derives), the system detects the dependencies and also runs the dependent Procedure(s) on all patients.
-
If the definition of a question has been changed, Batch Validation runs a validation check on all responses to that question.
Discrepancy Creation and Obsolescence
As the Batch Validation job processes data, it opens and "obsoletes" discrepancies and counts those that remain current, and reports all in its output file (viewable from the Batch Validation submission window by selecting Job Status and then View Output) as follows:
-
New Discrepancies. The system creates new discrepancies as appropriate on new or changed data or as a result of changes to Procedure or question definitions.
-
Obsolete Discrepancies. The system gives a status of Obsolete to previously created discrepancies that have been manually closed, either by correcting data or by giving them a status equivalent to Irresolvable.
-
Remain Current. Although the system runs the same Procedures and other validation checks on the same data (because it processes all of a patient's data whenever any of the patient's data is new or changed), the system does not replace existing discrepancies with new ones for the same error in the same data.
In the Batch Validation output file you can see information on the number of discrepancies created during that run, the number that remain current, and the number that became obsolete. However, these numbers reflect only one run of Batch Validation, so the number of discrepancies remaining current includes only those records that were processed during that run because data or definitions had changed. If you want an accurate figure for the total number of existing discrepancies, go to the Maintain Discrepancy Database window and query for the information you need. The number of records retrieved by the query appears at the bottom of the window.
To see the total number of current discrepancies generated in the study by a single Procedure, execute the Procedure explicitly; from the Conduct menu, select Data Validation, then select Execute Single Procedure, and look at the .out file. See Executing a Single Procedure in the Oracle Clinical Creating a Study manual.
See the following topics: