Jenifer Otieno ^1* Lisa Were ¹ Moriasi Nyanchoka ² Easter Olwanda ² Mercy Namuma Mulaku ^1,3 Xiao Hui Sem ⁴ Mikashmi Kohli ⁴ Jessica Markby ⁴ Angela Muriuki ⁴ Eleanor Ochodo ^1,5

• ¹ Centre for Global Health Research, Kenya Medical Research Institute (KEMRI), Kisumu, Kenya
• ² KEMRI Wellcome Trust Research Programme, Kilifi, Kenya
• ³ Department of Pharmacology, Clinical Pharmacy, and Pharmacy Practice, College of Health Sciences, University of Nairobi, Nairobi, Kenya
• ⁴ Foundation for Innovative New Diagnostics, Geneva, Geneva, Switzerland
• ⁵ Centre for Evidence-based Health Care, Stellenbosch University, Cape Town, South Africa

Introduction: HPV self-sampling is a relatively new, cost-effective and widely accepted method, however, uptake in LMICs remains limited. We aimed to map out the evidence and identify gaps in accuracy, acceptability, cost, equity and uptake of self-sampling vs. provider-sampling in LMICs. Methods: We searched: MEDLINE, EMBASE, CINAHL, SCOPUS, Web of Science, and Global Index Medicus, from 1946 to July 2023. Inclusion criteria entailed studies focusing on self-sampling alone or compared to provider-sampling for HPV testing and reporting on at least one outcome of interest (accuracy, acceptability, cost, equity, or uptake). Two authors independently screened titles, abstracts, and full texts, resolving disagreements through discussion. Data was extracted by one reviewer independently, with quality checks by senior authors, and results were synthesised narratively. Results: Our search yielded 3,739 records, with 124 studies conducted on 164,165 women aged 15-88 years between 2000 and 2023 included. Most studies were from the African region (n=61, 49.2%). Designs included cross-sectional (n=90, 81.1%), randomised (n=5, 4.5%), modelling (n=4, 3.6%), micro-costing (n=2, 1.8%), and non-randomised crossover (n=1, 0.9%) studies. Outcomes included; acceptability (n=79, 63.7%), accuracy (n=51, 41.1%), cost (n=7, 5.6%), and uptake (n=7, 5.6%). Most studies reported that participants preferred self-sampling, with only a few studies (n=7, 8.9%) studies favouring provider-sampling. The sensitivity and specificity of self-sampling ranged from 37.5% - 96.8% and 41.6% - 100.0%, respectively. One study directly compared the sensitivity and specificity of dry self-collected vs. wet provider-collected sample transportation. Laboratory costs were similar, but overall costs were lower for self-sampling. Uptake was higher for self-sampling in five of the seven studies. Most studies (n=106) mentioned equity factors like age (n=69, 65.1%), education (n=68, 64.2%) and place of residence (n=59, 55.6%) but no analysis of their impact was provided. Conclusion: HPV self-sampling is acceptable and cost-effective but, evidence of its accuracy shows varying sensitivity and specificity. Evidence on the accuracy of dry self-collected vs. wet provider-collected sample transportation is limited. Research evaluating HPV self-sampling’s accuracy, including comparisons of transportation modes, uptake, the impact of equity factors in LMICs and comparisons with high-income countries is essential to inform cervical cancer screening uptake.