An Unsupervised Smart App–Optimized HIV Self-Testing Program in Montreal, Canada: Cross-Sectional St

'); //]]>‎Figure 1. Screenshots of the HIVSmart! app. View this figure

Participants were deemed eligible to join the study if they were 18 years or older, self-identified as an MSM, of unknown HIV status, and comfortable using smartphones or tablets. Participants receiving pre- or postexposure prophylaxis (PrEP or PEP) antiretrovirals for prevention (ARVs) were also invited to participate so as to not exclude these high-risk groups. Participants were excluded if they self-reported a previously confirmed positive HIV diagnosis.

A flow chart outlining the study flow is represented in . Once deemed eligible to participate, the study procedure was explained to participants. Informed consent (written) was obtained from all participants before beginning the study. Participants could withdraw consent at any point throughout the study. At this time, participants were informed about the strategy that involved deidentified data collection on a confidential, compliant secure server.

Study Procedure

As per Health Canada’s product monographs and Investigational Testing Authorization recommendations, the usual clinic procedures were followed in the study. Following participants’ informed consent and agreement to participate, participants were explained about data storage and confidentiality. Following this, a rapid test (INSTI HIV-1/HIV-2 Antibody Test; bioLytical Laboratories) was performed by the research nurse. However, the rapid test result was not revealed to participants at this time. Blood was drawn for laboratory-based confirmatory testing, following the clinic’s protocol (Centre hospitalier de l’Université de Montréal): p24 antigen testing and anti-HIV-1/2 testing for all samples, confirmed by Western Blot if abnormal. Samples underwent RNA testing in case of suspected acute HIV infection. Lab testing conformed to the national testing algorithms.

Participants were then brought to a private office in the clinic and provided the OraQuick Test kit along with a tablet installed with HIVSmart!. The process of self-testing went thus: participants were left alone (unsupervised) to navigate the app on the tablet, stage their personal risk for HIV, perform the self-test as per instructions on HIVSmart!, and conduct the test unsupervised, mimicking a home environment. All participants followed instructions on pretest counseling, staging, conducting results, and storing their results on screen. Participants were asked to wait for their test result using a timer built into the app. They were encouraged to use the phone number provided in the app to call for counseling or assistance at any time and respond to a research questionnaire within the app. Upon completion of the self-testing process, participants were offered the choice of calling the provided phone number to receive posttest counseling or receiving face-to-face posttest counseling with a research nurse. Phone-based counseling was rapidly followed by face-to-face counseling with the research nurse. Upon meeting with the research nurse, who interpreted and recorded the self-test results in real-time separately, the INSTI rapid tests results were provided to participants and they were offered posttest counseling based on their INSTI results. Participants with positive HIV results were seen by a physician immediately for follow-up care. Participants with negative self-tests were encouraged to return for retesting in 3 months. As per the Institutional Review Board recommendations, all participants were offered a modest compensation for their time (Can $20).

'); //]]>

Participants were well educated, with 84.7% (371/438) educated beyond high school; 79.5% (348/438) were employed, and 52.5% (230/438) had been tested in the past 6 months. Participants self-identified themselves as male (98.6%, 432/438) and homosexual (91.3%, 400/438); 96.1% (421/438) of participants were sexually active, with 25.1% (110/438) stating 6-10 partners in the past 6 months; 21.9% (96/438) stating ≥11; and 70.1% (307/438) engaged in condom-less sex. In total, 11.5% (52/451) participants were currently taking PrEP and 1.1% (5/451) were taking PEP.

Acceptability was high at 98.5% (451/458); 7 participants refused to participate or withdrew themselves from the study.

Feasibility, as documented by the completion rate of the HIVSmart! self-testing strategy, was computed by taking an average of the 3 steps of the self-testing strategy—(1) self-test conduct; (2) self-test interpretation; and (3) linkages to care.

1. Self-test conduct: 97.1% (438/451) of the participants conducted the self-test successfully.
2. Self-test interpretation: 92.9% (419/451) of participants interpreted their self-test successfully.
3. Linkages to care: 100% (451/451) of the participants sought linkages to care successfully.

To compute feasibility, we took an average proportion of the 3 steps highlighted above, resulting in the overall feasibility of 96.7%.

Regarding Test Interpretations

Incomplete test conduct occurred in the initial set-up stages of the study when participants were unable to submit their results through the app because of Wi-Fi connectivity issues. These were later resolved by resolving incompatibility issues of the app and the clinic’s Wi-Fi server.

Regarding test interpretations, a few participants mistakenly interpreted their negative result (the control line) as their positive result, despite instructions. An invalid result was also recorded that was truly negative.

Regarding test results, 3 participants tested positive for HIV (0.7% seropositivity) with both the self-test and rapid test, which were rapidly confirmed by laboratory results, and the participants were linked to a physician within the same day (linkage: 3/3, 100%) and returned for a follow-up appointment. All negative rapid and self-test results (448/451, 99.3%) were confirmed negative through laboratory testing, and all were linked to counseling (448/448, 100%). Lab testing conformed to the national testing algorithms.

Regarding linkages, all participants (451/451, 100%) were linked to in-person counseling following the self-testing procedure. All participants used the phone line and later met the research nurse. The average turnaround time to linkage to counseling ranged from 2 to 6 hours.

Regarding the preference for counseling, participants were in favor of counseling over the phone, followed by face-to-face counseling in a clinic. Some favored counseling over the internet (chat or website; 132/421, 31.4%) or in a pharmacy (121/421, 28.7%), both followed by face-to-face counseling in clinic and counseling in clinics only (132/421, 31.4%). Participants were generally not in favor of no face-to-face counseling (28/421, 6.7%) or no counseling at all (3/421, 0.7%).

Regarding cost preferences, half of the participants (206/421, 48.8%) selected Can $10-20 and 27.4% (115/421) selected <Can $10. In terms of the usefulness of the app, 98.8% (417/422) of the participants found the app helpful in guiding them through the self-testing process. Finally, 94.3% (395/419) of the participants said that they would recommend this self-testing strategy to their partner.

'); //]]>

Discussion

Principal Findings

In this Canadian Institutes of Health Research funded innovative Canadian study, we investigated the feasibility of implementing an app-optimized unsupervised HIV self-testing strategy in a clinical setting.

Our unsupervised self-testing strategy was found to be feasible to operationalize (419/451, 96.6%), was well accepted (451/458, 98.5%), and preferred by participants. HIV was detected, and all participants were linked to care within a working day. All of our self-testers were linked to counseling or directed to a physician within hours, an essential service to offer with self-testing. Participants found the app-based approach to be a useful (417/422, 98.8%) in completing the HIV self-testing procedure, and a majority (395/419, 94.3%) wanted to recommend it to their friend.

This app-optimized, self-testing strategy was aimed to plug gaps in the self-testing process that are associated with an accurate detection, self-test interpretation, and rapid initiation of linkages to counseling and care. We were limited by the lack of approved HIV self-tests in Canada, which restricted our evaluation to a clinic environment [-] instead of homes. To stimulate a home environment, we set up kiosks in clinics where participants could test unsupervised, yet a nurse was always available to offer counseling and support []. Regarding costs, we found that a majority of participants believed that an acceptable price to pay for an HIV self-test in Canada was between Can $10 and $20. This finding is in line with 3 studies from New York City, where participants felt that an affordable and accessible price point was US $15-$25 (compared with the prevailing US $40) [-].

This is the first Canadian study to report data on the use of an app-based strategy. Many digital innovations (ie, Web-based programs, kiosk-based tablets, and short message service [SMS] text messaging services) are available, yet a complete, portable, and patient-friendly app-based solution for self-testing from engagement to linkage to counseling and care is novel [].

In 5 studies that have evaluated some digital innovations, we observed a few limitations in their offer of services that impacted the process. A Dutch study evaluated a Web-based strategy, where participants could purchase self-tests and access self-test instructions and counseling, but they did not provide data or information on linkages []. Of 4 US studies, 2 reported positive findings on the feasibility of use of a kiosk tablet for HIV self-testing in emergency rooms but reported poor engagement (50%) of participants and limited data on detection and linkages [,]. Another US study evaluating the use of SMS text messaging self-test results only by participants found that it was preferred by participants []. The fourth study provided written and Web-based video instructions to participants to choose a picture that resembled their self-test result and reported comparable results to ours, with 100% of positive OraQuick results and 98% of negative results being interpreted correctly []. None of these studies evaluated linkages to care.

The HIVSmart! app is an integrated innovation. It offers a personalized experience of self-testing from access to linkage, which is a step up from a website-, tablet-, or SMS text messaging-only service and is housed in a secure Health Insurance Portability and Accountability Act-compliant cloud-based platform. Integrated innovations [], like HIVSmart!, are a new trend in the digital innovations space, as reported in a recent systematic review []. In 2011-2013, we evaluated a Web-based HIVSmart! strategy successfully in South African health care professionals []. We are currently testing the strategy at scale in South African townships in a project funded by both the Governments of South Africa and Canada []. A prototype of this strategy was evaluated in students way back in 2009 [].

With the increasing availability of PrEP in Canada, and the desire to self-test frequently expressed by those on PrEP, we included a subsample of participants on PrEP in this study. However, neither our main outcomes nor HIV status differed by the PrEP status. Our study was underpowered to detect subgroup differences (small number of PrEP participants). Studies suggest that PrEP increases the window period for seroconversion, taking longer to get a positive test result [,]. Yet, many of our self-test results were consistent with the rapid test results. Concordant with a study in PrEP users from Kenya, participants were in favor of self-testing [].

Limitation

A limitation of this study included convenience sampling that raises a concern of selection bias.

Conclusions

In Canada, future research with HIV self-tests from provinces with high rates of undocumented HIV infection (Saskatchewan) and marginalized populations with undiagnosed HIV infection is warranted. Future research that incorporates digital strategies to plug service delivery gaps important for HIV self-testing will make it easier to offer and document self-testing.

Rapid approvals by Food and Drug Administration, Conformité Européenne, and World Health Organization Prequalification of self-tests, both oral and blood-based HIV self-tests, will help expand options to self-test in Canada. It will also increase the visibility of HIV self-tests in pharmacies, clinics, and outreach settings and democratize the process of HIV self-testing. Finally, the adoption of proven digital solutions will help improve engagement and expedite rapid linkages to care. Doing so will help address the last mile problem of detecting undiagnosed HIV infection in marginalized Canadians, thereby accelerating progress toward Joint United Nations Programme on HIV/AIDS 90-90-90 targets in Canada.

We conclude that the HIVSmart! app-optimized strategy is feasible, accepted, and preferred by an educated, urban MSM population of Montreal. With the app, participants were able to perform, interpret, store results, and rapidly link to care. The HIVSmart!-optimized, self-testing strategy could be adapted and contextualized to many at-risk populations within Canada and worldwide, thereby maximizing its public health impact.

Acknowledgments

The authors are grateful to all the staff at Clinique Médicale L’Actuel Montreal. We would also like to acknowledge Dr Marc Steben, Dr Bertrand Lebouché, and Dr Jean Pierre Routy for supporting the study. This work was funded by an operating grant from the Canadian Institutes of Health Research (grant #HHP-137872) and the Fonds de recherche du Québec - Santé Research-Scholar Junior 2 and Senior, awarded to NPP.

Authors' Contributions

NPP was involved in the concept, execution, write-up, critique, and overall responsibility of data for the project; MS was involved in execution, data analyses, write-up, and critique; LD was involved in execution, data collection, write-up, and critique; AG was involved in execution, data collection, write-up, and critique; KB was involved in execution, data collection, and write-up; AFV was involved in execution, data collection, write-up, and critique; LJ was involved in data analyses, write-up, and critique; and RT was involved in execution, data collection, write-up, critique, and responsibility for the project implementation.

Conflicts of Interest

None declared.

Abbreviations

Edited by G Eysenbach; submitted 20.03.18; peer-reviewed by M Steben, S Badman, C Figueroa; comments to author 05.08.18; revised version received 23.08.18; accepted 14.09.18; published 27.11.18

Copyright

©Nitika Pant Pai, Megan Smallwood, Laurence Desjardins, Alexandre Goyette, Krisztian G Birkas, Anne-Fanny Vassal, Lawrence Joseph, Réjean Thomas. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 27.11.2018.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.

 Sections

Journal of Medical Internet Research ISSN 1438-8871

Copyright © 2018 JMIR Publications

NEW: Help Desk Now Available