Understanding Oracle Clinical

Discrepancies and Derivations

Clinical data needs to be cleaned and manipulated, via discrepancy checks and derivations, before it is ready for analysis. Deriving values and checking for errors are tasks that can be done in SAS, OC, or by a PL/SQL program or listing. However, if you can check for errors sooner, you should have a resolution sooner. Thus it is generally better to do checks in OC than in SAS when the effort involved in programming is equivalent. Similarly, within OC it is better to have a check fire online than to wait for a Data Clarification Form to be faxed to the site.

Basic Discrepancy Management

In general, the sooner a discrepancy is raised, the easier and quicker it is to fix. The advantages of having RDC raise an error at the time of data entry start with time not spent pulling out a subject's medical record, since it's right there on the desk, and continue from there. The time-saving advantages of RDC relate almost entirely to discrepancy management not data entry; site personnel cannot be expected to enter data as quickly as trained data entry staff.

Thus it is extremely important to have your discrepancies programmed when an RDC study starts so that you will not lose this time-saving advantage. However, I have noticed that if only two or three very complicated checks are keeping a study from starting, there won't be any significant losses from adding them at a later time. Just make sure that all your standard checks as well as all the simple and obvious checks ...