OBJECTIVE We prospectively studied the diagnostic utility of our Bethesda system-based cervical cytology screening program with colposcopy and biopsy as the criterion standard. METHODS We prospectively collected and studied the correlation of cytologic, colposcopic, and histologic data in women referred for colposcopic examination because of "nonnormal" cytologic results or other risk factors. RESULTS We found that 771 of 5585 initial colposcopic visits yielded high-grade (cervical intraepithelial neoplasia II or worse) biopsies (13.8% prevalence); 13 showed invasive cancer (0.23% prevalence). Only 132 of 771 cases of high-grade dysplasia (17%) and 5 of 13 cases of invasive cancer (38%) followed Papanicolaou smears suggesting high-grade intraepithelial lesions or cancer, with 77% being discovered after "minor" Papanicolaou smear abnormalities. High-grade disease or cancer was confirmed in 1 of 2 high-grade or cancer Papanicolaou referrals and in 1 of 11 referrals with atypical squamous cells of undetermined significance. CONCLUSIONS Papanicolaou smears, especially those that are low grade, should not be equated with histologic sampling in association with poor cytohistopathologic correlation. Most high-grade dysplasias and cancers occur in women with either minor Papanicolaou smear abnormalities or visible lower genital tract lesions or both. Colposcopy for women with any "nonnormal" screening result is feasible.