Please select Print from the file menu or right mouse click and choose print from the menu to print this page.

[349] Effectiveness of Cervical Cancer Screening Using Correlation Analysis

SS Raab, DM Grzybicki, RJ Zarbo, JE Janosky, C Jensen, FA Meier, SJ Geyer, KR Geisinger. University of Pittsburgh, Pittsburgh, PA; Henry Ford Hospital, Detroit, MI; University of Iowa, Iowa City, IA; Western Pennsylvania Hospital, Pittsburgh, PA; Wake Forest University, Winston-Salem, NC

Background: Although the limitations of cervical cancer screening are well known and the estimated sensitivity of a single Pap test is approximately 60%, the impact of cervical cancer screening correlation discrepancies is less understood.
Design: Four hospital labs measured the frequency of Pap test-histologic correlation discrepancy over a 6-year period. We calculated the discrepancy proportion using a denominator of total Pap tests. The labs standardized the correlation process in the mid-point of the study. Standardization was based on a 6-month look back between the histologic and Pap test specimens and discrepancies were defined as greater than a 2-step difference in the Pap test and histologic diagnoses. We adjudicated the cause of discrepancy as sampling, if tumor cells were not seen on either specimen, or interpretation, if diagnostic cells were seen on a slide on review. Outcomes were obtained by chart review and classified into the categories of no harm, near miss and harm. The category of harm was classified into minimal (e.g., delay in diagnosis), mild (e.g., repeat non-invasive test), moderate and severe.
Results: A total of 6,260 cytologic-histologic discrepancies were found 1,657,528 Pap tests (1 in every 265) across the 4 institutions. A missed squamous cancer occurred in 1 of every 166,000 Pap tests, although this discrepancy was adjudicated as secondary to interpretation error in only 1 of every 552,000 Pap tests. Approximately 40% of discrepancies resulted in harm, mainly classified as minimal or mild harm (delays in diagnosis and repeat testing). Institutional discrepancy frequency varied from 0.14% to 0.62% of all Pap tests and reflected differences in correlation processes and/or biases in reporting. Histologic misinterpretation was the cause of 12% of all discrepancies.
Conclusions: Longitudinal, multi-institutional data show that cervical cancer screening using Pap testing is remarkably effective in preventing cervical cancer, as only 1 in 550,000 women had a missed squamous cancer. However, the problem with the current cervical cancer screening system is the large number of low-impact errors from over and under diagnosis resulting in over treatment.
Category: Cytopathology

Tuesday, March 27, 2007

Poster # 74, Tuesday Morning