Accuracy of colposcopy in the Swedish screening program.

Area of Science:

• Gynecologic oncology
• Cervical cancer screening
• Diagnostic accuracy studies

Background:

• Colposcopy's diagnostic accuracy varies significantly in real-world settings.
• The impact of colposcopist experience on assessment accuracy remains unclear, with conflicting study results.
• Evaluating colposcopy performance within national screening programs is crucial for understanding its effectiveness.

Purpose of the Study:

• To assess the accuracy of colposcopic examinations within the Swedish screening program.
• To investigate the variability in colposcopists' diagnostic assessments.
• To determine if colposcopist experience influences accuracy in routine clinical practice.

Main Methods:

• A cross-sectional register study analyzing 82,289 colposcopic assessments linked with histopathological biopsies.
• Data collected from women aged 18+ in Sweden between 1999 and September 2020.
• Accuracy calculated by agreement with biopsies, analyzing outcomes: Normal vs. Atypical, Normal vs. Low-Grade vs. High-Grade Atypical, and Non-High-Grade vs. High-Grade Atypical. Time-trend and experience-based analyses were performed.

Main Results:

• Overall accuracy for distinguishing Normal from Atypical findings was 63%, with overrating four times more common than underrating.
• Accuracy improved when distinguishing High-Grade from Non-High-Grade lesions (76%).
• Identifiable colposcopists showed an overall accuracy of 67%, with significant performance variations but no correlation with experience.

Conclusions:

• Colposcopy demonstrates low accuracy in differentiating Normal from Atypical cervical lesions, even in referral settings.
• Increased colposcopist experience alone does not enhance diagnostic accuracy.
• Substantial performance disparities among colposcopists highlight the need for further investigation into factors influencing accuracy.