Table 1.
Study characteristics by technology used
| Author | Study type | Participants | Study results |
|---|---|---|---|
| Social media | |||
| Patel et al., 2020 [26] | Randomized comparative trial to study feasibility, acceptability, and impact of peer-delivered messaging for HIV testing and consistent condom use. Two groups: (1) approach (desirable outcome to be achieved) group and (2) avoidance (undesirable outcome to be avoided) group. | 18+-year-old MSM | Increased self-reported HIV testing from baseline to 6-months follow-up (31.5 to 43.8%, P=0.04). No significant difference in condom use between baseline and follow up. HIV testing and condom use did not differ between two different intervention groups. |
| Rhodes et al., 2016 [22] | Randomized community trial to increase HIV testing through posts on social media. Two groups: (1) two intervention communities received information about HIV testing information and (2) two control communities did not receive any information. | 18+-year-old MSM | Post-test HIV testing for intervention group was 63.7% compared to 42.0% for comparison group (AOR 2.9, 95% CI 1.8–4.7). |
| Washington et al., 2017 [24] | RCT to examine feasibility of video intervention on Facebook to increase HIV testing. Two groups: (1) intervention group viewed five 1-minute videos weekly and (2) control group viewed standard HIV-related text information only. | 18–30-year-old Black MSM | OR for 6-week post intervention HIV testing 7.0 (95% CI, 1.72, 28.33, p=0.006). |
| Young et al., 2015 [25] | RCT to examine the efficacy of peer models who engaged participants on Facebook to increase request for HIV test. Two groups: (1) intervention groups with peer models and (2) control groups without peer models. | 18+-year-old MSM | HIV testing for intervention group was 17.0% compared to 7.0% for control group (AOR 2.61, 95% CI 1.55–4.38). |
| Study websites | |||
| Bauermeister et al., 2019 [29] | RCT to determine the acceptability and efficacy of an online intervention to change sexual risk behaviors and HIV testing. Two groups: (1) intervention group received distinct and affective content designed to build HIV risk reduction skills and positive sexual health behaviors and (2) control group provided with general HIV prevention content. | 18–24-year-old gay, bisexual, and MSM | 45% of participants reported testing for HIV at 3-month follow-up, with no differences observed between the two groups (46.7% vs. 40.0%; X [2] (1) = 0.49, p = 0.49). |
| Boni et al., 2019 [27] | Intervention study to examine the feasibility of an internet-based HIV self-test kit approach to increase HIV testing. | 18+-year-old MSM | The study website garnered 67,225 page views with 17,786 unique visitors during the study period. A total of 2526 HIV self-test kits were delivered and 21% of participants reported results online or by mail. |
| MacGowan et al., 2020 [28] | RCT to evaluate the effect of providing HIV self-test kits on testing frequency, diagnoses of infection, and sexual risk behaviors. All participants received information through study website. Two groups: (1) intervention group which received a self-test kit in the mail and (2) control group which did not receive a kit. | 18+-year-old MSM | HIV testing for intervention group at least 3 or more times during the trial was 76.6% compared to 22.0% for control group (P<0.01). |
| Live chat | |||
| Stephenson et al., 2020 [31] | RCT to test an online HIV prevention intervention using telehealth and home-based HIV self-test kit. Two groups: (1) intervention group received an HIV self-test kit and an appointment with a remote counselor who supervised administration of the test and (2) control group that was mailed one HIV self-testing kit. | 15–24-year-old transgender youth | 65 participants randomized into the intervention group did not participate in the intervention. Of the remaining intervention participants, HIV testing during the trial was 96% compared to 91% for control group. At 3-months follow-up, STI testing for intervention was 72.5% compared to 42.9% for the control group (p = 0.001), but declined by 6-months follow-up for both groups (intervention arm 19.2%, control arm 15.3%). |
| Wang et al., 2018 [30] | RCT to examine the efficacy of an online counseling session during a home-based HIV test. Two groups: (1) intervention group received the same video as control plus an additional video, an HIV self-test kit, a video chat counseling session and (2) control group received a 3-minute online HIV prevention video and encouragement to tested. | 18+-year-old MSM | HIV testing of any test type for intervention group was 89.8% compared to 50.7% for control group at 6-month follow-up (RR 1.77, p<0.001). |
| Short message service/text message | |||
| Kelvin, George, Kinyanjui, et al., 2019 [35] | RCT to test the efficacy of a text message reminder for a self-test HIV kit. Three groups: (1) intervention group received three text messages about the new self-test HIV kit, (2) enhanced standard of care group received the standard text message reminder for a finger-prick test administered at the clinic three times, and (3) standard of care group received the standard text message once. | 18+-year-old male truckers (driver or assistant) | HIV testing during the study period for the intervention group was 3.5% compared to 1.3% for the enhanced standard of care group (OR 2.7, 95% CI 1.3–5.5). No difference was found in testing rates for enhanced standard of care group and standard of care group groups. |
| Kelvin, George, Mwai, et al., 2019 [36] | RCT to test the efficacy of a text message reminder for a self-test HIV kit. Three groups: (1) intervention group received three text messages about the new self-test HIV kit, (2) enhanced standard of care group received the standard text message reminder for a finger-prick test administered at the clinic three times, and (3) standard of care group received the standard text message once. | 18+-year-old female sex workers | HIV testing during the study period for the intervention group was 10.8% compared to 6.1% for the enhanced standard of care group (OR 1.9, 95% CI 1.3–2.7). No difference was found in testing rates for enhanced standard of care and standard of care groups. |
| Njuguna et al., 2016 [33] | Quasi-experimental study to examine the effect of text messages in increasing HIV testing. Two groups: (1) intervention group received weekly HIV-related text messages and (2) control group did not receive text messages. | 18–24-year-old college women | HIV testing at the 6-month follow-up for the intervention group was 67% compared to 51% for the control group (hazard ratio 1.57, 95% CI 1.28–1.92, p<0.001). |
| Wettermann et al., 2019 [32] | Proof-of-concept study to examine the feasibility of a text message campaign to encourage patients to facilitate HIV testing discussion with their physician and increase HIV testing. Two groups: (1) group received HIV testing text message and (2) group received vaccine or calcium intake text message (control message). | 35–55-year-old patients in a healthcare system | 35% of the group who received an HIV text message requested a test and none of the patients who received the control message requested an HIV test. |
| Ybarra et al., 2017 [34] | RCT to test the effect of a text messaging-based HIV prevention program. Two groups: (1) intervention group that received HIV informational text messages and (2) control group that received healthy lifestyle text messages. | 14–18-year-old cisgender male who identified as gay, bisexual, and/or queer | HIV testing post intervention for the intervention group that was sexually active at baseline was 38.9% compared to 18.1% for the control group (AOR 3.39, p<0.001). At 90-day follow-up, testing for the intervention group was 55.1% compared to 28.2% for the control group (AOR 3.42, p<0.001). |
| Mobile/dating application | |||
| Balán et al., 2020 [37] | Multi-phase study to develop and test a smartphone-based app to increase HIV testing among users and their partners. | 18+-year-old MSM or transgender women who have sex with men | Participants found the app useful and engaging. Most watched the video tutorial before self-testing and scanned their test results with the app. |
| Horvath et al., 2020 [38] | RCT to test the feasibility, acceptability, and impact of a mobile app to increase HIV testing. Two groups: (1) intervention group had access to a mobile app with HIV-related information and (2) no treatment control group. | 18+-year-old MSM | Overall, the app was feasible and well accepted by participants. At 4-months, repeat HIV testing for the intervention group was 17% compared to 4% of the control group. However, at 8 months, repeat HIV testing for the intervention group was 50% compared with 40% of the control group. |
| Pai et al., 2018 [39] | Cohort to investigate if mobile app increased referrals, detected new infections, and expedited linkages to care. Three groups: (1) supervised self-testing, (2) unsupervised self-testing, (3) control with conventional testing. | Patients from primary care clinics | Referrals for self-testing was 17.4% compared to control group 2.6% (RR 6.69, 95% CI 4.47–10.01) and many new infections 86 vs 57 (OR 1.55, 95% CI 1.1–2.2). |
| Rosengren et al., 2016 [40] | Intervention study to examine the feasibility of banner ads placed in a dating app (Grindr) to distribute HIV self-test kits. Ads linked to a study website where participants were able to order a test kit. | 18+-year-old Black and Latino MSM | Banner ads placed on Grindr were effective in reaching high-risk populations. The study website received 4389 unique visitors and 333 test requests. Participants were open to providing personal information to obtain a free test kit, which participants found acceptable and easy to use. |
| Sullivan et al., 2017 [41] | Intervention study to examine the acceptability and usability of an Android-based HIV prevention app. | 18+-year-old MSM | Participants used the app for an average of 17.7 min over the study period. App users ordered condoms (63.3%) and HIV self-test kits (52.8%) through the app. During the study period, 77% of participants tested for HIV at least once and 56% tested multiple times. Participant thought the app was easy to use and 69% stated they would download the app again. |
| Wray et al., 2018 [42] | RCT to increase HIV testing, post-test follow-up care, and referral. Three groups: (1) eTEST group that was mailed a self-test kit in packaging attached to a device that alerted study investigators when the package was opened to initiate a follow-up call from a counselor, (2) standard group which received a self-test kit with no follow-up call, and (3) control group that was mailed informational letters about HIV testing lefts. | 18+-year-old MSM | HIV testing for eTEST and standard groups during the study period was 100% compared to 72% for the control group (p = 0.001). There were no significant differences in testing behavior between the two groups that received self-test kits. Repeat testing was also higher for participant in these two groups compared to control participants (81% eTEST, 77.2% standard vs. 40.9%, p<0.001). However, the eTEST group was significantly more likely to receive risk reduction counseling, prevention supplies, and PrEP referrals compared with the other two groups. |
| Zhu et al., 2019 [43] | RCT to test a mobile-based app designed to increase HIV self-testing. Two groups: (1) intervention group received two HIV self-test kits and access to study app that provided messages, videos, and referrals to health services related to HIV and (2) control group received two HIV self-test kits only. | 18+-year-old MSM |
HIV testing in the past 6 months at follow-up for intervention group was 92% compared to 68% for the control group (p< 0.001). Intervention group significantly increased HIV testing behaviors in the past 6 months compared to controls (ARR = 1.99, 95% CI 1.03–3.84, p < 0.05) and use of oral HIV self-test (ARR = 2.17, 95% CI 1.08–4.37, p < 0.05) |
| Informational videos | |||
| Aronson et al., 2016 [44] | Intervention study to examine the feasibility and acceptability of a tablet-based video designed to address barriers to HIV testing in the Emergency Department. | 18–24-year old Emergency Department patients | Of the 100 participants enrolled in the study, 98% completed the intervention (mean time <9 mins) and 30% accepted HIV tests offered by the tablets. Post-test showed significantly higher knowledge after than before the intervention (t = −6.67, p < .001). Overall, patients found the intervention highly acceptable, easy to use, and understandable. |
| Aronson et al., 2017 [45] | Intervention study to examine the feasibility and acceptability of a tablet-based video designed to provide opioid overdose prevention and response and facilitate HIV and HCV testing in community settings. | 18+-year-old clients of a community-based syringe exchange program | Of the 11 participants enrolled in the HIV module, 91% accepted HIV tests offered by the tablet. All participants in the HCV module accepted the HCV test, but only 60% of the overdose prevention module agreed to receive additional naloxone training. Overall, participants in the three different modules accepted the intervention well with acceptability scores no less than 6.8 out of 10. |
| Tang et al., 2016 [46] | Non-inferiority RCT to compare first time HIV self-test rates among participants who viewed HIV test promotion videos. Two groups: (1) group that received crowdsourced HIV test video and (2) group that received health marketing HIV test promotion video. | 16+-year-old MSM and transgender individuals | HIV testing for the crowdsourced group was 37% compared to 35% for the health marketing group. The estimated difference between the groups was 2.1% (95% CI −5.4 to 9.7%). Investigators conclude that crowdsourcing may be a feasible tool for effective HIV test promotion messaging. |
| Electronic health record | |||
| Ahlström et al., 2019 [47] | Investigators examined the feasibility of using electronic registry data to predict HIV status using machine learning algorithms. | Members of a national hospital registry who met study criteria and at least 16 years old | Machine learning algorithms can learn from electronic health records to predict HIV cases and distinguish them from non-HIV cases with high accuracy. Further integration of prediction models may be complementary to condition-guided HIV testing and identification of patients suitable for PrEP. |
| Kershaw et al., 2018 [48] | Retrospective cohort analysis of electronic medical records before and after the implementation of a reminder in the electronic medical record for HIV screening. | 18–65-year-old patients of a medical left | Overall, HIV testing increased to 30.7% after the reminder, up from 15.3% (RR 2.02, CI 1.95–2.09, p < 0.0001). The significant increase (p<0.0001) was seen in all age groups. |
| Marcus et al., 2019 [49] | Investigators examined the feasibility of using electronic registry data to predict incident HIV diagnosis using absolute shrinkage and selection operator logistic regression to identify potential candidates for PrEP. | 18+-year-old patients of a medical left who met study criteria | Prediction models using electronic health record data were able to identify incident HIV cases. Simpler models identified fewer patients and had lower sensitivity for incident HIV cases compared to the full model (6.0–28.9% vs. 38.6% of incident HIV cases). |
| Sha et al., 2019 [50] | Investigators examined the feasibility of a screening program of electronic health records to increase HIV testing, diagnosis, and linkage to care in the emergency department. | 13–64-year-old patients of a medical left who met study criteria | Overall, HIV testing increased to 17.1% after electronic health record screening was implemented and 21.1% when HIV testing was linked to blood test orders, up from 2.5% at baseline. |
| Vending machine | |||
| He et al., 2019 [51] | Intervention to examine the acceptability, feasibility, and effectiveness of a vending machine that dispensed HIV urine specimen collection kit for HIV testing. | University students | 11 vending machines were placed in seven universities during an AIDS/STD educational campaign. Of 957 urine collection kits distributed by the campaign during the study period, 197 were sold through the vending machines and 760 were freely distributed. 64.5% of kits purchased from the vending machines were returned for analysis. 72.2% of kits purchased were from vending machines in dormitories and 27.8% from instructional buildings. |
| Lee et al., 2020 [52] | Cross-sectional structured survey to identify acceptability, preferences, and concerns about HIV test kits dispensed from vending machines. | 16+-year-old Black African, Black Caribbean, and Latin American | Rates of willingness to use HIV self-test kits from vending machines were high in all groups except Black Caribbeans (Black Africans 77.9%, Black Caribbeans 53.6%, Latin Americans 81.8%, other non-white ethnicities 80.0%). Participants preferred vending machines located in healthcare settings over community settings. |
| Stafylis et al., 2018 [53] | Intervention study to increase HIV testing through delivery of free HIV self-test kits dispensed in vending machines. | MSM | Vending machines were placed in two commercial sex venues in the on-site STI testing room and social lounge. A total of 1398 kits were dispensed through vending machines, and 110 surveys were completed by patrons who obtained a free test kit. Of those surveyed and reported that the kit was used, 95% reported using the test kit themselves, and 17.7% of these participants reported a reactive result. |
| Vera et al., 2019 [54] | Intervention to examine the feasibility and acceptability of vending machines to dispense HIV self-test kits in community settings. Two comparative methods: (1) test kits dispensed through vending machines and (2) test kits distributed through community workers. | 18+-year-old clients of a sauna frequented by MSM | A total of 265 test kits were dispensed through the vending machine. HIV testing was 34 tests per month through the vending machine compared with 6 tests a month through the outreach testing by community workers in the same venue. The vending machines were accepted by the participants with 94% of survey respondents stating they would use the vending machine or recommend it to others. |
| Young et al., 2014 [55] | Qualitative study to examine the usability and acceptability of using vending machines to dispense HIV self-test kit. | 18+-year-old African American and Latino MSM | Overall, participants interviewed stated that obtaining an HIV self-test kit from a vending machine was acceptable. Participants differed in opinion regarding aesthetics of the vending machine but agreed that confidentiality and customization made using a vending machine acceptable. Participants felt confident that the test kit they obtained from the vending machine was an accurate medical device and sanitary. Other factors for acceptability of using a test kit from a vending machine included: ease of use, time efficiency, and confidentiality. |
| Chatbots/conversational agents | |||
| Brixey et al., 2017 [56] | Implementation paper on designing a chatbot using Facebook application programming interface within its Messenger platform. | Developers used a variety of credible sources to create a body of questions and answers about HIV/AIDS which served as the foundation of knowledge for the chatbot. Developers used a response classifier and dialogue management system to drive responses to questions. | |
| van Heerden et al., 2017 [57] | Experimental study to explore the feasibility of using a conversational agent embedded in a messaging mobile application to provide pre-test counseling prior to an HIV test. | Six of the 10 testers preferred a quick reply button instead of typing out responses to the chatbot and thought that both the aesthetics and appropriate responses made the interaction feel like a conversation with a human. Other testers thought the interaction was not realistic, and that the language was too formal. Seven testers felt that the interaction was private and anonymous and therefore, they would be willing to take a test with the chatbot. | |
| Vermey et al., 2019 [58] | Assessment of an online advice application with a chatbot feature that provides tailored advice based on user-provided information regarding personal characteristics, sexual behavior, sexual risk, and symptoms. | The online application was visited 337,736 times in 2018 with 113,257 visitors started the questionnaire, and 17,449 utilizing the chatbot. Around 60% of visitors finished the questionnaire, which led to 65,736 advice given, and 8700 clicks to online self-test providers. | |
RCT randomized control trial, MSM men who have sex with men, PrEP pre-exposure prophylaxis, (A)OR (adjusted)odds ratio, CI confidence interval, (A)RR (adjusted)relative risk, HCV hepatitis C virus