Quality assurance of gram-stained direct smears.

Clinical laboratories do not always prepare and examine direct smears. In part, this may be explained by lack of control over the precision and accuracy of the procedure, which has diminished confidence in direct smears among both laboratory scientists and clinicians. Objective criteria were established for enumeration of cells and placement of bacteria in identification categories. Two methods for monitoring the accuracy and reproducibility of their application were studied: (1) examination of previously examined and reported direct smears of clinical specimens by independent observers with subsequent review of discrepancies with decisions by personnel; (2) preparation of suspensions of cells and bacteria that yielded identical smears for subsequent examination as unknowns. Method 1 showed agreement in bacterial identification category in 65%, quality score 74%, enumeration of neutrophils 79%, squamous cells 76%, and erythrocytes 92%. Method 2 showed that conformance of technologists' observations to expected results varied, for neutrophils conformance was 72-78%, squamous cells, 68-78%, bacterial identification category, 100%, and bacterial enumeration, 45-96%.