Humans cannot perform tasks correctly each time. Unaided, 98 percent correct performance is excellent. Nevertheless, when the consequences of error are severe, multiple system redesign steps may be needed to reduce the rate to an acceptable level. Achieving zero error may be impossible. Transfusion medicine provides a good example of the roadblocks to perfection.
Hospitals have taken vigorous steps to reduce the risk of major hemolytic transfusion reactions resulting from an incorrect crossmatch. In a recent paper (Ansari and Szallasi), the authors look at their success at eliminating discrepancies between the patient identified on the tube’s label and the patient whose blood is in the tube (wrong blood in tube — WBIT).
In large-scale studies, the detected WBIT rate tends to be about 0.04 percent, but the real WBIT rate is even higher. Detection of the error only occurs when some red flag arises, such as a mismatch with an earlier recorded blood type, realization by the phlebotomist that he made a patient recognition error or post-facto analysis of an unexpected transfusion reaction.
The authors noted that their rate was similar to the typical rate, and WBITs persisted even when policies were changed to have two witnesses to patient ID and bedside tube labeling. They then added a second control, the two-tube/check-type method. This requires two different persons to draw blood from the same patient at two different times for patients who had no previously-recorded blood type. This relatively simple step provides a “defense in depth,” adding a second, independent type check, markedly reducing the risk of WBIT.
With the two-tube/check-type method, a WBIT event will be caught when the ABO or Rh blood types are discrepant. However, WBIT may still occur when the two samples come from patients with the same major blood type (e.g., A+.) Then the crossmatch, which also screens for rarer incompatibilities, may still be in error. Indeed, the authors detected several potential errors by other means that were not detected by the two-tube/check-type method.
Further rate reductions may require more automation and innovations in patient and blood sample identification. To quote Richard Friedberg of Baystate Health System: “Absent a bar-code on the patient’s forehead, the rest are all compromises.”
Ansari S, Szallasi A. “Wrong blood in tube”: solutions for a persistent problem. Vox sanguinis. 2011;100(3):298-302.