Objective: The main objective of this study was to determine the effectiveness of the clinic-based HPV self-sampling approach at a rural HIV clinic in Uganda.
Methods: A simple single-blind randomized controlled trial was used to estimate the efficacy of a clinic-based (intervention) compared to a home-based (control) HPV self-sampling approach among 150 HIV-infected women aged 25-49 years at the rural HIV clinic. The Health Promotion Model (HPM) guided the broad conceptualization. The participants were randomized to either clinic or home-based HPV self-sampling using a ratio of 1:1. The assignment was concealed and handed over to the midwife. The outcome was a continuation rate for HPV self-sampling at 6 months follow-up among participants tested for HPV. Survival analysis was used to determine the effectiveness of the clinic-based HPV self-sampling. The continuation rate was determined using the intention to treat analysis. The predictors of time to continue with HPV self-sampling in both arms, recovery, and Cox proportional hazards regression were used. The multivariable model was built using the forward modeling approach while controlling for confounding variables. The best model was selected after comparing the fitness of the final models (global test under chi-squared distribution). A Cox–Snell residual plot was fitted to ensure the final model did not violate the proportional hazards assumption. Adjusted Hazard ratios and their 95% confidence interval were reported.
Results: The results show that of the HIV-infected women in the clinic-based arm, 92% (69/75) had attained at least a primary level education, 78.7% (59/75) were self-employed, and 74.4% (56/75) had a history of STIs compared to those in the home-based HPV self-sampling arm. The overall continuation rate was generally low at 30.7% (46/150) among HIV-infected women receiving HPV self-sampling in both arms. The continuation rate of HPV self-sampling was higher at 41.3% (31/75) among HIV-infected women randomized to receive clinic-based HPV self-sampling compared to the 20% (15/75) of those in the home-based arm. The average time for timely continuation of HPV self-sampling of 8.2 days while those in the home-based arm had 22 days. The factors that influenced the continuation of HPV self-sampling were residing between 6-10km (HR:0.11; CI: 0.014-0.84), a history of taking alcohol (HR: 7.74; CI: 1.06-56.54), age of a sexual debut above 18 years (HR:1.48; CI:1.08-2.03) and educated about HPV by the health worker (HR: 17:22; CI: 1.106-268.42).
Conclusion: The overall continuation rate for HPV self-sampling is low. The clinic-based HPV self-sampling is more effective compared to the home-based HPV self-sampling approach. The factors influencing the increased continuation of HPV self-sampling are having a history of taking alcohol, age at sexual debut of >18 years, and receiving HPV education from health workers. Therefore, it is important to improve HPV self-sampling services at the HIV clinics and boost the capacity of health workers to educate HIV-infected women on HPV as well as provide integrated HPV self-sampling within the HIV clinics.