In 2012, the OraQuick ADVANCE Rapid HIV-1/2 Antibody Test (OraSure Technologies Inc, Bethlehem, USA) was the first HIV self-test (HIVST) approved for sale in the United States as an over-the-counter HIVST rapid diagnostic test (RDT) for individuals with no prior HIV testing experience.¹ Since then, over 2.5 million HIVST kits have been sold globally and more than 4 million have been distributed through donor funded programmes.² The World Health Organization (WHO) strongly recommends that HIVST be utilised as a way to complement existing HIV services³ as self-testing may reduce barriers associated with traditional facility-based testing, like travel, wait times and privacy concerns.^4,5

Based on this growing body of evidence, South Africa became one of over 40 countries to have incorporated HIV self-testing into their national HIV policies,^6,7 with self-testing introduced as a way to help close the gap between the 84.9% of adults living with HIV who know their HIV status and the 90% target of the UNAIDS 90-90-90 initiative.^5,8,9,10 The introduction of HIVST programmes will improve access to further HIV diagnostic services, prompting an increase in testing uptake and frequency, which could lead to earlier diagnosis.¹¹

There are, however, several concerns related to HIVST, as there is no formal pipeline for users to self-report their results or be linked to care following the self-test. These HIVST kits are not diagnostic, but rather considered tests for triage, and all positive results should prompt the user to seek confirmatory testing by a trained healthcare professional.¹² Furthermore, the independence of HIVST presents considerable challenges surrounding the monitoring and evaluation (M&E) of HIVST programmes, which are required by public health stakeholders to understand the uptake and effectiveness.¹³

Strong mobile phone penetration in low- and middle-income countries (LMIC)^14,15 has led to the development of a variety of mobile health (mHealth) interventions to complement HIVST. These include telephone hotlines, short message service interventions, internet-based platforms and mobile applications (apps).^{16,17,18,19,20} A Brazilian study conducted in 2019 showed that an internet-based intervention targeting men who have sex with men led to 21.4% of online participants self-reporting, whilst an interactive voice response telephone line in South Africa was found to link 9.8% of participants to care.²¹ Whilst these platforms have shown varied success, the introduction of mHealth interventions for linkage to care and M&E are in line with the South African National Department of Health mHealth Strategy (2015), and should be explored further.²²

Despite data concerns in LMICs,²³ recent trends are towards the development of downloadable apps due to their agility and scalability.²⁴ The app interface also provides developers with a malleable platform that can be tailored to individual users, allowing them to curate a collection of HIVST information, resources and guidance for testers, whilst also capturing the HIVST result data.^19,20 Recently, HIVSmart^TM, a Canadian app, was developed to guide users through the testing process, link them to care, and store the HIVST result data. Preliminary evaluations in key Canadian populations, as well as healthcare workers in South Africa have shown the app to be feasible and acceptable; however, neither HIVSmart^TM, nor any other app, has been developed or tested for the general population in LMICs.^9,20,25

South Africa has shown previous acceptance of HIV-related mHealth interventions with SmartLink, an app that improved linkage to care for clinic-based HIV testing in participants under 30 years of age.²⁶ Another successful mHealth intervention, MomConnect, has been used by over 2 million pregnant South African women with information regarding their pregnancy, whilst also creating a national pregnancy registry.^27,28

The Aspect^TM HIVST app was developed to help strengthen and complement HIVST programmes by supporting self-testers through testing, facilitating linkage to care and digitising the reporting of HIVST results through an operational dashboard for M&E. The specific objective of this pilot study was to evaluate the acceptability and feasibility of the Aspect^TM HIVST app for individuals from the inner city of Johannesburg, in order to advise further scale-up. We present the findings from this pilot.

This pilot study is the first investigation of an mHealth app to enhance monitoring and evaluation of HIVSTs for individuals from the inner city of Johannesburg, and the findings from this pilot have established that participants showed high acceptability of the intervention, whilst also identifying challenges that can be targeted for improvement as the platform scales up. The high acceptability was similar to that of the HIVSmart^TM app and a Brazilian internet-based intervention; however, these studies only evaluated the feasibility of using the app to link patients to care or increase testing uptake, respectively.^9,20,21,37

The Aspect^TM HIVST app, instead, aimed to guide participants through the testing process, then upload the results to a central server for M&E, and this additional layer of complexity has introduced more opportunities for user error. The majority of errors, however, were not as a result of the app functionality, but rather test usability and the IFU that guided the self-testing process. Errors stemming from the IFUs have been well documented in a number of HIVST studies, including ones from South Africa.^38,39,40 Suggestions like clarifying the instructions, incorporating video or voice notes, and offering additional languages should all be taken into consideration, especially as more HIVSTs, each with specific IFUs, become available to the market. Some of these suggestions have already been implemented by other platforms, as the HIVSmart^TM app is already available in both of Canada’s national languages, and provides supplemental video content.²⁰

There were a number of discrepancies between HCW-recorded and app-captured data on participant demographic information. There were also some difficulties in the uploading of the test strip photo via the app. A simple summary page, similar to that seen on a banking app, before completing a transaction, could provide the user with an opportunity to review their information before submitting it through the app. This additional checkpoint should help prevent any data entry errors. One variable, however, previous HIV testing history, had 34 (12.7%) discordant entries between what the HCW recorded and what the app captured; all 34 entries reported never having HIV tested to the HCW, but were captured in the app as having previously tested. It is possible that privacy of the app has revealed an interviewer bias, where some participants may not have felt comfortable sharing sensitive information with the HCW, but felt free to do so through the app. Previous mHealth studies have also found that self-administered tools may decrease interview bias;⁴¹ however, further evaluation of this app and its users would be required before stating that the app is responsible for removing or decreasing this interviewer bias.

Some participants also had difficulty understanding how to take a picture of the test strip. When test images were reviewed on the Aspect^TM dashboard, the images were quite variable in terms of quality. The purpose of this functionality was to allow a third party to manually review test images and flag potential discordant results for follow-up. However, similarly to other studies,^20,42 we had high concordance between participant and HCW interpretation of the self-test and, thus, this step may not even be necessary if lay persons are able to interpret results as accurately as trained HCWs. In low bandwidth environments, the requirement to upload images may also incur additional data charges and may not be cost effective.

With the number of countries adopting HIVST policies being on the rise, the M&E of these programmes poses a unique set of challenges¹² and measurement of uptake and effectiveness becomes difficult. The Aspect^TM HIVST app facilitated the capture of HIVST data directly to an operational dashboard, namely Aspect^TM. This dashboard was developed by SystemOne and is currently being used to report tuberculosis and HIV viral load results from over 3000 diagnostic instruments across 43 countries.²⁹ For this study, the dashboard displayed very basic summary HIV statistics, a list of individual test results and also supported the downloading of automated reports. This could allow a programme manager to remotely monitor indicators such as uptake, demographics of the testing population, HIV positivity rates, invalid rates and improve reporting against key performance indicators. The functionality of the dashboard also allows for the pushing of automated SMS notifications directly to the tester based on their HIV result, which could be used to promote confirmatory testing and help link them to care.⁴³ This is especially important for HIVST, as one of the problems with home testing is that people receiving a positive diagnosis are suddenly faced with a serious diagnosis and no immediate access to information, counselling or treatment resources.¹¹ The feasibility of these dashboard features should be considered for future research.

Data concerns are also an important issue in South Africa, with previous mHealth studies highlighting data costs and phone memory as a barrier to entry.^26,44 Future app development should focus on keeping storage requirements minimal to ensure that the app is available for as many individuals as possible. Furthermore, the necessity to upload images may also incur additional data charges and may not be affordable for all users.

The study had several limitations. Convenience sampling from one sub-district from inner-city Johannesburg was used to recruit participants limiting the generalisability of the findings, and the compensation of participants may have accounted for the very high participation rate. Furthermore, the majority of participants were under 35 years old, which may have made it easier for them to navigate a mobile app as they may be more tech-savvy than older age groups. The Aspect^TM HIVST app was only available in English. It was also only tested on a Samsung phone, and it may not reflect the usability of the app on other phones owned by the general population, especially across different operating systems and memory capacities. The discordance between HCW-recorded and app-captured demographics may reflect an interviewer bias, whilst the process of testing in front of a HCW may have increased the number of forced errors due to the pressures of being observed. Performing the HIVST with the app in a clinic, with a HCW present, may also present bias, as the app is intended to be used independently of a clinic setting. Another limitation of the pilot process was that the HCWs did not record the participants’ interpretation on paper and, thus, results discordance could not be verified, as was done for the other variables.

Although recent studies have introduced validated data collection tools for mHealth usability,⁴⁵ at the time of this study, there were also no validated data collection tools to measure the acceptability and feasibility of mHealth apps for HIVST, hence the study-specific questions may not be used to reproduce these results in similar settings. Similarly, the use of only one HIVST kit and its accompanying IFU means that these results cannot be generalised across all HIVSTs, especially since many of the errors were related to the interpretation of the IFU.

With millions of HIVST kits distributed worldwide without adequate tracking, the need for M&E of these kits is ever increasing. On an individual level, this may lead to better linkage to care and follow-up with patients and, on a national level, tracking can identify areas of need to optimise kit distribution, marketing and supplementary information. Despite some challenges with IFU interpretation and data capture via the app, this pilot study has shown that the Aspect^TM HIVST app is an acceptable way to upload mobile HIVST results and demographic information to a central database.