RT Journal Article
SR Electronic
T1 P5.042 Quality Assurance in Visual Inspection of the Cervix - the South African Experience
JF Sexually Transmitted Infections
JO Sex Transm Infect
FD BMJ Publishing Group Ltd
SP A347
OP A347
DO 10.1136/sextrans-2013-051184.1086
VO 89
IS Suppl 1
A1 C Firnhaber
A1 L Mao
A1 D A Lewis
A1 B Goeieman
A1 A Swarts
A1 M Faesen
A1 S Levin
A1 N Rakhombe
A1 S Williams
A1 J S Smith
YR 2013
UL http://sti.bmj.com/content/89/Suppl_1/A347.3.abstract
AB Background Invasive cervical cancer is the most common cancer is Sub-Saharan Africa partly due to lack of effective screening programmes and the HIV epidemic. The baseline analysis of a South Africa study comparing three cervical cancer screening methods, i.e. cytology, visual inspection with acetic acid (VIA) and HPV testing in HIV-infected women, has been completed and quality assurance of VIA interpretation was evaluated. Methods HIV-infected women, aged 18–65, were recruited from Johannesburg, South Africa and VIA performed with 5% acetic acid. Nurses received a two-week training on the VIA procedure. VIA interpretation by the nurse and recording of a digital image was performed in real time. The study physician reviewed weekly the digital imaging and recorded their own interpretation of the VIA appearance. Women with a positive VIA and 25% of women with a negative VIA had confirmatory colposcopic biopsy. Sensitivity and specificity for VIA interpretation for cervical intraepithelial neoplasia (CIN)-2 + for both the nurse and the doctor were compared retrospectively. Results 1202 HIV infected women with a median age of 37 years and CD4 of 394 cells/mm3 participated. 45% of the women had a positive VIA result. Sensitivity and specificity for VIA to determine CIN 2+, as performed by the nurse and doctor, was 65.4% (59.7–71.1)/68.5% (65.3–71.7) and 75.5% (70.5–80.4)/68.1(65.0–71.3) respectively. There was no statistical difference in the VIA readings comparing the first 600 VIA interpretation to the final 593 VIA readings [nurse (p = 0.613) or doctors (p = 0.624)]. Evaluating the first 300 VIA to the final 300 results was also similar (p = 0.505 for nurse; p = 0.802 for doctors). Conclusion Throughout the study there was no statistical difference in the sensitivity in interpreting VIA. These results show that after two weeks of training, nurses were consistent in their VIA readings and could accurately predict 65.4% of CIN 2+.