I promise this is the last time...

[Mabel Adams]

One more thing that happened with our barcode scanners for bedside ID: they would have connectivity problems to the wireless network so sometimes the labels wouldn't print and the phlebs would work around the lack of label. Later, when someone else was using the same scanning device, the backed up labels would print with a string of labels for a different patient. Aaaagh! As I said, make sure they are working as intended.

If we could use one absolutely unique identifier (say MRN), get it assigned to the right patient's body, never change it, never remove it from that patient's body, never make mistakes with it when using it to identify them or the blood, then we would be fine. Problem is we humans can't be that accurate so we have systems to have multiple identifiers and recheck them at every step--which, as John says, only complicates things. The complexity often obscures the identity issue as much as clarifies it. We need to identify and quantify the sources of error and the associated outcomes. Then we need to decide how often it is okay to hang the wrong blood. I know, never--but in the world of probability, chance may never be zero. I think I would be okay with once in the US per every 500 or 1000 years. That would make the chance of ABO mismatch and fatalities some percentage smaller. Right now, per the FDAs latest release we had 4 fatalities from ABO mismatch in the US in 2009--2 phleb errors, one hanging blood error, one lab error. Maybe once we knew how the different variables contributed to the errors, we could find the most effective way to improve the process. OK, I am really "thinking aloud" and defining the problem for the mathematicians here so help me figure out if I have left out anything.

Or we just need to implant ID chips like we do for our pets.

[BSIPHERD]

Our absolute on this is that the two identifiers on the armband must match the specimen must match the Blood Tag on the unit. If the name is truncated on the armband, it must be truncated exactly the same on the specimen and Blood Tag. The specimen must be labeled directly from the armband. (And armbands always have the last name,first name so we don't accept specimens that are labeled first name last name because they couldn't have been copied directly from the armband.)

If the patient's name is Jones, Robert but the armband says Jones, Bobby, then the specimen and the Blood Tag must also say Jones, Bobby.

And we are also very strict on the rules being followed for Blood Bank specimens.

I read a study once quite a while ago that pointed out that even specimens with minor errors (letters reversed, wrong vowel, etc) had a much greater risk of being Wrong Blood in Tube than correctly labeled specimens.

Belva in Lincoln

[tbostock]

I use this one ALL the time:

Lumadue JA, Boyd JS, Ness PM. Adherence to a strict specimen-labeling policy decreases the incidence of erroneous blood grouping of blood bank specimens. Transfusion 1997;37:1169–1172.

This study found that specimens with an obvious labeling error were 40 times more likely to contain WBIT (wrong blood in tube).

I read a study once quite a while ago that pointed out that even specimens with minor errors (letters reversed, wrong vowel, etc) had a much greater risk of being Wrong Blood in Tube than correctly labeled specimens.

Belva in Lincoln