Self-sampling for human papillomavirus DNA detection: a preliminary study of compliance and feasibility in BOLIVIA

Background: Cervical cancer incidence and mortality rates in Bolivia are among the highest in Latin America. This investigation aims to evaluate the possibility of using simple devices, e.g. a cotton swab and a glass slide, for self-sampling in order to detect human papillomavirus (HPV) DNA by PCR in cervico-vaginal cells. Methods: In the first phase of our study we evaluated the use of a glass slide as a transport medium for cervical cells. A physician took paired-cervical samples from 235 women. One sample was transported in Easyfix (R) solution and the other sample was smeared over a glass slide. Both were further analyzed and compared for human DNA recovery and HPV detection. A kappa value was determined to evaluate the agreement between the HPV DNA detection rates. In the second phase of the study, 222 women from the urban, peri-urban and rural regions of Cochabamba were requested to perform self-sampling using the following devices: a cotton swab combined with a glass slide, and a vaginal tampon. Women gave their opinion about the self-sampling technique. Finally, the agreement for high risk-HPV detection between self-and physician-collected samples was performed in 201 samples in order to evaluate the self-sampling technique. Results: Firstly, the comparison between Easyfix (R) solution and the glass slide to transport clinical samples gave a good agreement for HPV DNA detection (kappa = 0.71, 95% CI 0.60-0.81). Secondly, self-sampling, especially with cotton swab combined with glass slide, would generally be preferred over clinician sampling for a screening program based on HPV detection. Finally, we showed a good agreement between self-and physician collected samples for high risk-HPV detection (kappa = 0.71, 95% CI 0.55-0.88). Conclusions: Simple devices such as a cotton swab and a glass slide can be used to perform self-sampling and HPV DNA detection. Furthermore, most Bolivian women preferred self-sampling over clinician-sampling for cervical cancer screening.