Rapid prescreening is as effective at reducing screening error as postscreening with the FocalPoint automated screening device.

Both manual (rapid prescreening) and automated methods have been devised to reduce screening error compared with routine review of 10% of negative gynecologic cases. To date these two methods have not been compared. A total of 5,139 liquid-based (SurePath Pap Test, BD Diagnostics) cases were subjected to 100% rapid prescreening and routine manual screening. All cases diagnosed as negative on routine screening were placed on the FocalPoint (BD Diagnostics, Franklin Lakes, NJ) automated screening device. All cases that were negative on routine screening and abnormal on rapid prescreening or in the top 15% of cases by FocalPoint were rescreened and a final diagnosis obtained. Cases were blinded during rapid prescreening and routine manual screening. Abnormal was defined as any diagnosis of atypical squamous cells (ASC) or worse. A total of 427 (8%) of cases were abnormal on routine screening. The sensitivity of rapid prescreening was 44.6%. Rapid prescreening identified an additional 14 abnormal cases (13 ASC and 1 LSIL) and FocalPoint identified nine cases (eight ASC, one LSIL) that were not detected by routine screening. Three of these cases were detected by both methods. The sensitivity of routine screening was 93.1%. Rapid prescreening increased the overall sensitivity significantly (96.0%, P = 0.04); FocalPoint increased the sensitivity but this change was not significant (95.0%, P = 0.21). Estimated screening time was 30 seconds for rapid prescreening and 6 minutes for routine screening with the result that rapid prescreening required 2,570 minutes while review of the FocalPoint Slides required 2,694 minutes. The sensitivity of routine screening in the second half of the study (95.3%) was higher than that in the first half of the study (91.6%) but the difference was not significant (P = 0.11). Rapid prescreening is as effective as directed review using FocalPoint at reducing screening errors and requires no additional screening time. Continued use of these methods may improve the sensitivity of routine screening