The issue is whether accepting "historically negative" units without retyping is a reasonable risk or not. The underlying question is whether the unit is from the donor the system says it is from. No collection-distribution system is perfect. Every time a unit is manipulated in a processing station, it is usually manipulated at the same time as multiple other units. Some processing systems attach a filter after the unit has passed all testing to save the cost of the filter on units that are rejected for one reason or another. This is a point at which units can be mislabeled. Confounding the chances of not catching this error is the habit of processing like ABO units at the same time. Donor A's ABO was A neg prior to filtration, the unit labeled as Donor A tested as A neg after filtration, so it must be Donor A's blood in the bag labeled as Donor A. Anyone who believes this assumption to be true has never worked as QA in a blood processing facility, which I have. There is always the chance that the blood in the bag can't be truly linked back to the donor. Thus, the "historically antigen negative" label is a crap shoot. That said, the issue will always be whether it was reasonable for you to accept the label as true. I can tell you that a jury would not find it reasonable in a survivor's lawsuit. I did a complete flipflop on my attitude toward risk and liability after two events in my life. The first event was when I had to defend my practices in a wrongful death lawsuit. The second was obtaining my law degree. The rule of thumb is that if you could have prevented an accident, then you will be held liable for the accident if you had a duty of any sort to prevent it. Providing safe blood is one of the highest duties you can imagine. Skimp because of cost, and you will pay the price if something goes wrong. It doesn't matter that you were right 99.9999% of the time. BC, J.D.