Quantification of errors in laboratory reports. A quality improvement study of the College of American Pathologists' Q-Probes program.

Over a 3-month period, 61,496 errors were detected in clinical laboratory reports by 631 participants in the College of American Pathologists' 1990 Q-Probes program. Each error detected was defined as an opportunity for improvement. Almost 4% of the detected errors were attributed to nonlaboratory personnel and approximately 4% (A errors) had a major impact on patient care. Rates of B (serious errors, but unlikely to affect patient care) and C errors (minor clerical errors) were approximately equal. When expressed in terms of measures of laboratory work loads, four of six measures of mean errors were lowest in blood banking, intermediate in chemistry and microbiology, and highest in hematology. Thirteen percent of participants did not have an error detection system in place. We conclude that many errors go undetected, and we recommend that an effective system for error detection in patient reports should be employed in all laboratories.