Analysis of false-negative and underreported smears in the Florence District screening program for cervical carcinoma

Aims and background: To review false-negative or underreported (reactive changes, squamous or glandular atypia) smears performed in women developing histologically proven CIN2 or more severe lesions within 24 months and evaluate error causes. The study setting was the Florence District cervical cancer population-based screening: about 60,000 women age 25-60 years screened per year. Methods: 118 false-negative or underreported cases were identified at screening files-cancer Registry matching, and the original smears were reviewed by six independent readers to judge smear adequacy and error type. Results: Sampling errors (reported as inadequate, negative or less severe than CIN1 at review) accounted for 74% and screening!interpretation errors (reported as CIN1 or more severe at review) accounted for 26% of studied cases. Screening/interpretation errors were more likely ascribed to misinterpretation and underreporting than to misperception of cellular abnormalities. Conclusions: Quality control should above all address the problem of sampling adequacy. Due to the rarity of misperceived abnormalities (true screening errors), manual or automated rescreening of negative smears would not be an effective procedure for quality control.