Is there a difference in sample adequacy when vaginal HPV DNA samples are self-collected at home, a health post, or a primary care health center in rural Ethiopia? Implications for community cervical cancer screening.

Self-collection of HPV DNA samples is recommended as a cervical cancer screening method in areas with high barriers to clinical examination, such as Ethiopia. Self-collected sample adequacy in clinical settings is high compared to clinician collection, but less is known about self-collected sample adequacy in community settings. We evaluated sample adequacy differences when samples were taken at women’s homes, a local health post, or a primary care health center in rural Ethiopia. Women either self-collected vaginal samples for HPV DNA at home (Arm 1, n = 100), at a health post (Arm 2, n = 100), or at a health center (Arm 3, n = 200). Women received identical sample collection kits and illustrated instruction pamphlets; all samples were treated the same way once collected. HPV DNA testing was performed using Ampfire Multiplex High-Risk HPV (Atila, Mountain View, California). Thirty-two (8%) of the 399 samples were inadequate (negative β-globin gene). Sample inadequacy frequency was highest in Arm 1-Home (15%) compared to Arm 2-Health Post (7%) and Arm 3-Health Center (5%) (p < 0.05)). Arm 1 had significantly higher odds of inadequacy than Arm 3 (aOR: 2.8, 95% CI 1.2–6.8, p = 0.02) when adjusted for age, education, and marital status. There was no difference in adequacy between Arms 2 and 3. HPV prevalence was lowest in Arm 1 (6%) and significantly higher in Arms 2 and 3 (20.4% and 15.9%, respectively, p = 0.02). More women in Arm 2 reported positive views of self-sampling, and fewer reported embarrassment when using the self-test compared to Arm 1. Collecting self-samples at home yielded more inadequate samples than collecting at a health center; however, self-samples collected at the health post were not significantly different from those collected at the health center. In rural areas, self-sampling at locations proximal to women’s residences but providing privacy may increase screening campaigns’ success. While more work is needed to confirm this finding, home sampling may miss women who should be referred for diagnostic testing.