America’s HIV epidemic among youth aged 12 to 24 years has significantly increased in the last 15 years [1,2]. Young people now represent 26% of the HIV epidemic [1,2] despite investments in evidence-based behavioral interventions (EBI) and more recent scale-up of innovative antiretroviral treatments that can stop acquisition of HIV, which are known as preexposure prophylaxis (PrEP) and postexposure prophylaxis (PEP) [3,4]. Adolescents continue to become infected at disproportional rates [4]. It is critical to intervene with youth at high risk (YAHR) of acquiring HIV before they become infected. This study aims to intervene with YAHR with a set of interventions, which could be easily mounted, tailored, adapted over time, and broadly disseminated.

YAHR are those in urban epicenters and increasingly in the southeastern United States, particularly men who have sex with men (MSM) and transgender youth [1]. Homeless youth are also at elevated HIV risk, yet the last HIV seroprevalence study was conducted in 1991, showing 5.3% prevalence [5]. Youth in the criminal justice system may also be at elevated HIV risk [6,7]. YAHR are difficult to identify and intervene with in medical clinics because most youth (about 60% of general adolescent population) do not access health care [8-10]. Adolescents typically fail to disclose their sexual behaviors to their families or their physicians, most often because they are never asked about risk [11-13]. YAHR are likely to be encountered at agencies serving lesbian, gay, bisexual, transgender, and queer (LGBTQ) youth; homeless shelters; in the criminal justice system; and through in-person or Web spaces associated with sexual networking [14]. In each geographic epicenter of HIV, African American and Latino youth are at the highest relative risk of contracting HIV [1,15]. YAHR also typically seek peers, economic opportunities, and social services in the neighborhoods associated with the highest prevalence of HIV such as Hollywood in Los Angeles and the French Quarter in New Orleans [16]. This study uses community-based recruitment and screening to identify YAHR (see details in Rotheram-Borus et al [17]).

The current paradigm for reaching global HIV prevention goals (ie, increased uptake of the HIV prevention continuum [18]) is far more complex today than it was in the first 25 years of the epidemic. The US Centers for Disease Control and Prevention (CDC) recommends repeat and routine testing for HIV and sexually transmitted infection (STI) for YAHR every 3 or 6 months, as well as concurrently linking youth to HIV prevention and health care services and retaining them in care over time [19]. This study aims to operationalize and evaluate the impact of these guidelines using community-based recruitment and implementation of rapid diagnostic testing, referral and linkage to services, and easily scalable and tailorable technology-mediated interventions. The possibility now exists to implement biomedical and combination biobehavioral prevention for YAHR, which requires that youth know their serostatus (ie, be repeatedly tested for HIV over time), be linked to medical care, and consistently adhere to a strategy to protect themselves from HIV (eg, high adherence to PrEP, PEP, and condom use [18]). Textbox 1 summarizes the multiple endpoints for operationalizing uptake of HIV prevention continuum. These are particularly challenging tasks because the developmental challenges of youth evolve with age and may be more difficult for gay, bisexual, and transgender youth; homeless youth; and youth involved in criminal justice because of discrimination and stigma, which are often exacerbated for African American and Latino youth [11,20-23].

Traditional EBIs for HIV prevention rely on highly scripted and manualized protocols, delivered in-person, and that are time limited and became increasingly brief in numbers of and durations of sessions over the past 20 years [24]. Although EBIs for youth HIV prevention have been selected and diffused by the CDC on the basis of demonstrated efficacy in trials, they have been challenging to scale up and lack evidence for effectiveness [24]. These EBIs are highly structured and scripted manual-based protocols, which have been noted as being difficult to implement with fidelity and are not tailorable to intervention facilitators’ and youths’ varying styles, preferences, hierarchies of needs, development stages, and HIV risks. They also typically rely on in-person visits at community-based organizations (CBOs) or health care settings. Disruptive innovations of massively scalable mobile and social media technologies may be able to implement and broadly reach youth with prevention messages and linkage to services [24-26]. Bower and Christensen [27] defined the concept of disruptive innovations in Harvard Business Review as simpler, cheaper, and good enough solutions to meet the majority of consumers’ needs and preferences relative to incumbent products and services that are often designed for the highest-need consumers, for example, ATM machines versus bank tellers, US $2 reading glasses versus specialized prescriptions, or minute clinics in pharmacies. Rotheram et al (2012) applied this concept to EBIs for youth behavioral and mental health problems by positing that interventions based on common practice elements identified across manualized EBIs, and technology-mediated modalities could be disruptive innovations that might be more scalable, adaptable, and amenable to providers’ and patients’ needs and preferences [24].

Advances in mobile phone and social media technologies have created opportunities to engage and intervene with large numbers of youth at relatively low costs, using technologies that permeate their daily routines [25,28]. This study will use 2 primary technology platforms, that is, text messaging and social media, in addition to telephone and in-person visits based on youths’ preferences. Text messaging, email, and social media use are nearly universal among youth [29-31]. Approximately 90% of youth report having a mobile phone [31-33], 90% of them text about 30 times each day [31]; rates are similar for homeless youth but with some inconsistency in maintaining service [29,30,34]. Similarly, over 90% of youth access the Web daily, and smartphones have become the primary mode of accessing the internet for the majority of youth [31].

This study’s automated and interpersonally delivered technology-mediated interventions are based on the shared features of existing EBI—we do not aim to create a new smartphone app or an EBI with a manual to be replicated with fidelity [35,36]. In the last 25 years, over 100 HIV EBIs and 36 adolescent sexual health EBIs have been identified by the CDC and other review bodies as efficacious [37-39] and supported for diffusion. Members of this study team rated the manuals of 5 of the CDC’s most popular, behaviorally oriented primary prevention EBIs for adolescents [36,40,41,42], finding that each of them incorporated common processes, principles, and factors. Each of the EBIs are also based on cognitive behavioral theories, even though some researchers cited a more specific iteration (eg, social cognitive theory, theory of reasoned action, and the AIDS risk reduction model) or meta-theory (eg, information-motivation-behavior and social action theory). The EBIs had much more in common than different [36,40-42]. The intervention approach for this study is to focus on the common elements that many different interventions share, delivered by the most cost-efficient, scalable, and adaptable delivery strategies that permeate youths’ daily routines: mobile and social media technologies.

This hybrid implementation-effectiveness study examines alternative models for implementing the CDC’s guidelines for routine HIV/STI testing for YAHR of acquiring HIV and for delivering EBIs in modular elements instead of scripted manuals, while also monitoring implementation, costs, and effectiveness [43]. A randomized controlled trial (RCT) evaluates the efficacy and cost-effectiveness of the 4 intervention strategies of variable intensities and costs. This study aims to inform future prevention programs implemented by communities to avert the acquisition of HIV among young people by monitoring outcomes at 4-month intervals over 24 months.

All procedures in this study have been approved by the institutional review board (IRB) of the University of California, Los Angeles, which serves as the single IRB of Record for researchers at collaborating institutions. This is 1 of 3 studies in a National Institute of Child Health and Human Development (NICHD)-funded U19 Cooperative Agreement for the Adolescent (HIV Medicine) Trials Network (ATN) as well as a Management Core and Analytic Core [17,44-47]. This study began recruitment in May 2017. From June 2018 to December 2018, the protocol underwent a process of review and revision in collaboration with NICHD project scientists, the study’s scientific monitoring committee, and an external statistical expert, to meet budget constraints and align more closely with ATN scientific priorities. The major changes were (1) to reduce STI testing frequency from every 4 months routinely to only at baseline, 12, and 24 months and rectal testing only (unless requested by youth or indicated by symptoms); (2) stopping follow-up and intervention at 12 months with youth who are not MSM or transgender; and (3) changing randomization allocations for person-mediated intervention arms to have statistical power for MSM and transgender participants with 70% retention rates.

Figure 1 shows the study design overview. The randomized controlled factorial design assesses efficacy of 3 intervention strategies (arms 1-3) and their combination (arm 4). The person-mediated interventions in arms 2 to 4 have a sample size of 270 per arm with 220 MSM and transgender youth, whereas the larger sample of 690 in the automated intervention only (arm 1) is because of the broader U19 Cooperative Agreement’s goal to identify youth acquiring HIV infection during follow-up to refer to a sister protocol on acute HIV infection [45], and earlier expectations to recruit youth that could be referred to other ATN protocols. Participants are followed longitudinally over 24 months (12 months for youth who are not MSM or transgender) and assessed at 4-month intervals by interviewers. Participants also complete a brief 7-question survey every week by text message (or email when nonresponsive to text messages). The overviews of assessments are provided below and described in detail in linked publications for the U19 [17,44-47].

In both Los Angeles and New Orleans, the Recruitment, Engagement, and Retention Centers in the Management Core [17] are responsible for recruiting, enrolling, and following up with youth. Youth are primarily recruited from CBOs and clinics serving gay, bisexual, and transgender youth; homeless youth; and youth on probation or released from incarceration. Youth are also recruited through dating apps (eg, Grindr, Jack’d, and Scruff), including while present in social venues (eg, bars, clubs, and community events). We anticipate the study sample to be predominantly MSM and transgender, African American and Latino, and aged between 18 and 24 years. Youth are approached with verbal assent to complete brief screener questions and a rapid HIV test; however, youth aged between 12 and 14 years are asked to provide written voluntary informed consent to screen because of IRB requirements. Details on recruitment and screening are provided in other publications from the U19 [17,47] and briefly summarized below.

To screen as eligible for enrollment, youth must test seronegative on a rapid HIV testing at screening and report at least 3 of the following criteria: self-reporting as gay, bisexual, or transgender; African American or Latino race/ethnicity; having unprotected anal sex, sharing needles for injecting drugs, or an HIV-positive partner in the last 12 months; having been homeless (defined as not having a regular place to sleep for 3 or more months); illicit substance use (not including marijuana) in the last 12 months; having been hospitalized for a mental health disorder; having been in jail or on probation; having an STI in the last 12 months. Transgender and MSM youth are always eligible. Eligible participants are invited to enroll in the study with written informed consent.

Following enrollment, participants complete a baseline assessment, which includes a questionnaire and a series of rapid diagnostic tests. These assessments are repeated at follow-ups. Participants receive a US $50 cash incentive for each baseline and 4-month follow-up assessment completed. Other assessments include weekly monitoring surveys, staff implementation time and processes, and costing data for cost-effectiveness analyses. These assessments are briefly described below and in detail in Rotheram-Borus et al [17].

Questionnaires are administered by an interviewer in a private room at partner sites or study offices, using Android tablets and take approximately 45 min to complete (see Rotheram-Borus et al for details [17]). Briefly, baseline questions cover lifetime and past 4 months; follow-up assessments cover the past 4 months. Questionnaires assess sociodemographic factors; health care access, insurance, and utilization; substance use; sexual behaviors; PrEP and PEP use; mental health; social media use; and locator information and consent to access medical records and to use social security numbers for locating efforts. Interviewers enter the participants’ responses and rapid diagnostic test results in the CommCare mobile-Web electronic assessment and case management system that is cloud based and Health Insurance Portability and Accountability Act (HIPAA) compliant, by Dimagi Inc (see Comulada et al for details [44]). CommCare also manages and sends intervention and assessment text messages as part of the automated text-messaging and monitoring intervention (AMM).

The monitoring strategy is based on a weekly Check-In survey, which can be completed via text message or a Web link sent via email with a HIPAA–compliant RedCap version of the survey. Surveys assess the previous week time period with 7 questions on potential symptoms of acute HIV infection (yes/no), STI symptoms (yes/no), and number of days of feeling sad/depressed, sex without condoms, drug or alcohol use, not having a place to sleep or enough food; and miss taking medications (if taking medications). Participants receive a US $1 incentive per weekly survey completed, either in cash at their next 4-month follow-up assessment or on demand, including via electronic transfer (via Paypal, Venmo, or Zelle). Yes responses to HIV and STI symptom questions are reported to interviewers for follow-up risk assessment and rapid testing within 2 weeks. Reports on sadness/depression are monitored for potential follow-up and referral.

Our primary outcome is uptake and adherence to the HIV prevention continuum, according to the following measures, which will be analyzed individually and in sum (see Statistical Methods below):

Secondary outcomes include mental health symptoms, substance use, and housing security, which are hypothesized to impact primary outcomes as mediators or moderators of intervention effects.

There are 2 types of costs: costs of delivering the intervention and the additional costs incurred by participants for their use of health care services and services from other agencies (eg, use of health care services). Intervention costs are classified into 1 of 4 categories [48]: (1) capital equipment (eg, computers); (2) recurring supplies and services; (3) facility space; and (4) personnel, including fringe benefits. Costs in the first 3 categories are obtained from project records. Personnel costs include hours and wages of staff to design and deliver the interventions, including peers, coaches, supervisors, facility charges, software costs, and short messaging service (SMS) and other social media costs, messaging and mobile app data costs, additional time in coaching and supervision, and server hosting. Personnel time is estimated from time reported on time sheets for hourly employees and budgeted time for other staff and investigators. Time spent on specific activities for hourly staff (eg, coaches and interviewers) is assessed in detail over 1-week periods quarterly using the Time It app [49] on their study-issued Android smartphones. Recorded time over 1-week periods is extrapolated to cover total time over the study period. The costs of additional services are derived from respondent reports on utilization and medical records and are estimated using publicly available data. Research-specific costs, such as incentive payments, informed consent, assessments, and software adaptation for survey tools, are excluded from total costs. All cost data are price-adjusted back to the first year of the study, using the medical care component of the consumer price index.

The goal of this study is to test the efficacy and examine the cost-effectiveness and implementation of alternative models for delivering the CDC’s guidelines for routine HIV/STI testing for YAHR of acquiring HIV and for delivering evidence-based interventions in modular elements instead of scripted manuals and in flexible technology-based delivery formats. The technology-mediated interventions for YAHR in this study aim to improve HIV prevention continuum of engaging in medical care, adopting PEP after HIV exposure or PrEP before HIV exposure, or a behavioral protection strategy, as well as repeatedly testing for HIV and STIs. Consistent access and utilization of medical care is a common challenge for adolescents and young adults [114], particularly African American and Latino youth [115]. Text messaging and social media technologies offer relatively low-cost modalities to scale interventions for adolescents nationally. The study design provides opportunities to assess the efficacy, potential synergistic or cumulative effects, and cost-effectiveness of the proposed automated and person-mediated strategies. The study assessments will also enable examination of time trends in onset and periodicity of risk, and the relationships between primary and secondary outcomes in bivariate outcome analyses. In the evaluation of each intervention condition’s cost-effectiveness for the primary outcomes, we hypothesize that each intervention of increasing intensity (AMM, online peer support groups, and coaching) will have greater efficacy but that the added costs may not justify use at scale.

The mobile and social media intervention arms in this study build off the relatively nascent evidence base of internet, mobile, and social media interventions for populations at high-risk for acquiring or transmitting HIV [116-120]. Although technology-based assessments (eg, Web- and text-message surveys) [92,117] have demonstrated success with large samples in the thousands [121], mobile and social media technology intervention studies have tended to be in small scale by comparison [122]. Although several large-scale RCTs are currently underway to address HIV, most focus on a single technology-based strategy or a bundle of strategies in a single intervention arm, instead of a comparison of multiple strategies as in this study [120,123,124].

Youth present with a wide variety of issues that affect their risk for HIV infection. In addition, they can be extremely labile in terms of their emotional and behavioral reactions to life events as they undergo developmental changes, which affect their risk. The interventions in this study are available to youth over the duration of their study follow-up of 24-months for YAHR of acquiring HIV in the US (ie, MSM and Transgender youth) and 12-month of follow-up for other youth at elevated risk. This approach acknowledges variability and unpredictability of risk behaviors as adolescents experience developmental transitions. The interventions are available to prepare youth for transitions or to be available when transitions, crises, and risks occur. This is in novel contrast the last 30 years of the EBI movement for HIV prevention, which incentivized highly structured and increasingly brief interventions that demonstrated only short-term impacts on behaviors.

A limitation of this study is that study procedures and enhanced standard of care are relatively intensive interventions in and of themselves. Routine or indicated HIV and STI testing several times per year follows the CDC’s recommendations for people at high risk of acquiring HIV; however, this study operationalizes this with active and incentivized follow-up, community-based testing, and immediate treatment provision and partner therapy for STIs without a clinical visit. Daily text messaging and weekly monitoring with follow-up add another layer of intervention that cannot be directly assessed in this study. These limitations result from several factors, including utility for study assessments, ethical standards for research with vulnerable adolescents and the ATN to provide some base level of prevention support, goals for the broader U19 to identify youth with acute HIV infection from this study’s participants over time for a sister protocol [45], and to operationalize and examine costs and impacts of implementing CDC guidelines with community-based strategies [125]. Therefore, the potential efficacy and cost-effectiveness of the online peer support group and coaching interventions examined in this study should be considered within this context.

The study findings will be invaluable to inform future adolescent prevention interventions, not only for HIV but also in many other areas. This study will have policy implications for the allocation of resources to HIV testing resources in local communities, the uptake and scalability of interventions for youth, and innovative approaches for designing and diffusing EBI globally.