Using routine programmatic data to estimate the population-level impacts of HIV self-testing: The example of the ATLAS program in Cote d’IvoireUsing routine programmatic data to estimate the population-level impacts of HIV self-testingpreprintFotsoArlette SimoautJohnsonCherylautVautierAnthonyautKouaméKonan BlaiseautDiopPapa MoussaautSilholRomainautMaheu-GirouxMathieuautBoilyMarie-ClaudeautRouveauNicolasautDoumenc-AïdaraClémenceautBaggaleyRachelautEhuiEboiautLarmarangeJosephautTeamthe ATLASaut© 2022, Posted by Cold Spring Harbor Laboratory. This pre-print is available under a Creative Commons License (Attribution-NoDerivs 4.0 International), CC BY-ND 4.0, as described at http://creativecommons.org/licenses/by-nd/4.0/medRxiv2022-02-09monographic10.1101/2022.02.08.22270670https://www.medrxiv.org/content/10.1101/2022.02.08.22270670v1Background HIV self-testing (HIVST) is recommended by the World Health Organization as an additional HIV testing approach. Since 2019, it has been implemented in Côte d’Ivoire through the ATLAS project, including primary and secondary distribution channels. While the discreet and flexible nature of HIVST makes it appealing for users, it also makes the monitoring and estimation of the population-level programmatic impact of HIVST programs challenging. We used routinely collected data to estimate the effects of ATLAS’ HIVST distribution on access to testing, conventional testing (self-testing excluded), diagnoses, and antiretroviral treatment (ART) initiations in Côte d’Ivoire. Methods We used the ATLAS project’s programmatic data between the third quarter (Q) of 2019 (Q3 2019) and Q1 2021, in addition to routine HIV testing services program data obtained from the President’s Emergency Plan for AIDS Relief dashboard. We performed ecological time series regression using linear mixed models. Findings The results are presented for 1000 HIVST kits distributed through ATLAS. They show a negative but nonsignificant effect of the number of ATLAS HIVST on conventional testing uptake (−190 conventional tests [95% CI: −427 to 37, p=0·10]). We estimated that for 1000 additional HIVST distributed through ATLAS, +590 [95% CI: 357 to 821, p<0·001] additional individuals have accessed HIV testing, assuming an 80% HIVST utilization rate (UR) and +390 [95% CI: 161 to 625, p<0·001] assuming a 60% UR. The statistical relationship between the number of HIVST and HIV diagnoses was significant and positive (+8 diagnosis [95% CI: 0 to 15, p=0·044]). No effect was observed on ART initiation (−2 ART initiations [95% CI: −8 to 5, p=0·66]). Interpretations Social network-based HIVST distribution had a positive impact on access to HIV testing and diagnoses in Cote d’Ivoire. This approach offers a promising way for countries to assess the impact of HIVST programs. Funding Unitaid 2018-23-ATLAS Evidence before this study We searched PubMed between November 9 and 12, 2021, for studies assessing the impact of HIVST on HIV testing, ‘conventional’ testing, HIV diagnoses and ART initiation. We searched published data using the terms “HIV self-testing” and “HIV testing”; “HIV self-testing” and “traditional HIV testing” or “conventional testing”; “HIV self-testing” and “diagnosis” or “positive results”; and “HIV self-testing” and “ART initiation” or “Antiretroviral treatment”. Articles with abstracts were reviewed. No time or language restriction was applied. Most studies were individual-based randomized controlled trials involving data collection and some form of HIVST tracking; no studies were conducted at the population level, none were conducted in western Africa and most focused on subgroups of the population or key populations. While most studies found a positive effect of HIVST on HIV testing, evidence was mixed regarding the effect on conventional testing, diagnoses, and ART initiation.Added value of this study HIVST can empower individuals by allowing them to choose when, where and whether to test and with whom to share their results and can reach hidden populations who are not accessing existing services. Inherent to HIVST is that there is no automatic tracking of test results and linkages at the individual level. Without systematic and direct feedback to program implementers regarding the use and results of HIVST, it is difficult to estimate the impact of HIVST distribution at the population level. Such estimates are crucial for national AIDS programs. This paper proposed a way to overcome this challenge and used routinely collected programmatic data to indirectly estimate and assess the impacts of HIVST distribution in Côte d’Ivoire.Implications of all the available evidence Our results showed that HIVST increased the overall HIV testing uptake and diagnoses in Côte d’Ivoire without significantly reducing conventional HIV testing uptake. We demonstrated that routinely collected programmatic data could be used to estimate the effects of HIVST kit distribution outside a trial environment. The methodology used in this paper could be replicated and implemented in different settings and enable more countries to routinely evaluate HIVST programming at the population level.enPages: 2022.02.08.22270670medRxiv