RAP - The Main Game, page-1098

Each test case will be anonymised and tagged with a unique identifier. The Resapp predictions will go into one data store. The traditional test results will come back some time later to be separately collated and stored by a different team that does not have access to the Resapp predictions.

Early in the trial a sample of Resapp predictions may be matched against test results to identify methodological problems that need to be addressed while the trial is being conducted.

But for the most part the correlation of Resapp predictions to test results will not be done until after the trial is complete in order to preserve the blindness of the trial.

The results have to be collated by two separate teams and any differences reconciled to eliminate human error in matching results. Then they have to test the accuracy of the Resapp predictions using the appropriate statistical methods. And finally the submission will have to be prepared and submitted. There is likely to be a trial report published in a medical journal as well. Somewhere between FDA submission and medical report preparation is where we'll get an announcement of the results.

It occurs to me that some of the time estimates I'm seeing here might be a little opimistic...