Self-collected versus medic-collected sampling for human papillomavirus testing among women in Lagos, Nigeria: a comparative study

Objective: To evaluate the feasibility and performance of self-collected vaginal swab samples for HPV screening among women in Lagos, Nigeria. Methods: A cross-sectional study was implemented from March to August 2020 among sexually active women. Study participants provided same-day paired vaginal swab samples. Medic-sampling and poster-directed self-sampling methods were used to collect the two samples per participant. A real-time PCR assay detected HPV 16, HPV 18, other-high-risk (OHR) HPV, and the human beta-globin gene. The self-collected samples' sensitivity, specificity, and accuracy were determined against the medic-collected samples using the MedCalc Online Diagnostic Calculator. Results: Of the 213 women aged 16 similar to 63-year-old recruited, 187 (88%) participants had concordant results, while 26 (12%) participants had discordant results. Among the 187 concordant results, 35 (19%) were HPV positive, 150 (80%) participants were HPV negative, and two (1%) were invalid. 18 (69%) out of the 26 discordant samples were invalid. The self-collected sample was invalid for 14 (54%) participants. Two (8%) medic-collected samples were invalid. Compared to the medic-collected sample, the self-collected sample was 89.80% (95% CI: 77.77 similar to 96.60%) sensitive and 98.21% (95% CI: 94.87 similar to 99.63%) specific, with an accuracy of 96.31% (95% CI: 92.87 similar to 98.40%). The mean age for HPV positive and negative participants were 39 and 40, respectively, with an ANOVA p-value of 0.3932. The stratification of HPV infection by the age group was not statistically significant (P > 0.05). Conclusions: With high accuracy of 96%, self-collected sampling is adequate when tested with real-time PCR and may increase the uptake of HPV testing. Though more self-collected samples were invalid than medic-collected samples, most likely due to poor collection, they could be identified for repeat testing. Future implementation can avoid this error with improved guidance and awareness.