HIV/sexually transmitted infections (STIs) and drug abuse remain major public health concerns in the United States. Although youth in general (defined here as adolescents 13-25 years of age) are at increased risk of HIV/STI, African American and Hispanic youth are disproportionately affected [1,2]. African American and Hispanic youth disproportionately engage in sexual risk and licit/illicit drug use behaviors [3,4], which increase their risk for STI and HIV infection [5,6]. However, HIV testing—a National Institutes of Health HIV/acquired immune deficiency syndrome (AIDS) research priority [7]—among African American and Hispanic youth remains underutilized [8]. Furthermore, many youths are not routinely screened for asymptomatic STIs, as recommended by the Centers for Disease Control and Prevention [2]. Youth experience multiple barriers to access HIV/STI preventive services, and exploring the use of technology as a platform to improve the delivery and uptake of such services is warranted.

Given the ubiquity of technology among youth, technology-based interventions provide an ideal platform for the dissemination of HIV/STI and drug abuse preventive interventions to populations disproportionately affected by HIV/STIs, including African American and Hispanic youth [6,9,10]. However, a gap in knowledge persists in the literature with respect to the state of the science of technology-based HIV/STI and drug abuse preventive interventions. The purpose of this review is to identify randomized controlled trials (RCTs) of technology-based HIV/STI and/or drug abuse preventive interventions for African American and Hispanic youth.

As shown in Table 1, our search revealed three RCTs, all of which were published in 2013 or later. The studies examined the preliminary efficacy [41], efficacy [42], and comparative effectiveness [43] of a preventive intervention. Two studies used an intention-to-treat approach [41,43], one study used an attention control condition [42], and type of randomization was documented in two of the studies [41,42]. Sample sizes ranged from 72 [42] to 141 participants [43]. The longitudinal designs ranged from two weeks [43] to six months [42] follow-up, and two trials were registered with ClinicalTrials.gov [41,43]. The primary outcomes across the three studies included reductions in past 30-day sex partners [43], number of cigarettes smoked in the past 30 days [42], and rate of condomless anal sex acts [41].

Marsch et al [43] examined the comparative effectiveness of an interactive, customizable, Web-based intervention on sexual risk behaviors and the potential theoretical underpinnings by which change occurs (eg, skills acquisition), relative to a traditional educator-delivered intervention, among youth in an outpatient community-based substance abuse treatment clinic. Employing an intention-to-treat approach, 141 youth were randomized to either the experimental or control condition, and assessed at baseline and two-weeks postintervention.

Marsch et al [43] provide descriptive statistics whereby participants could indicate both their race and ethnicity, with 44.6% identifying as Black and 51.0% Hispanic, respectively. Participants in the experimental condition received a self-directed, interactive, Web-based intervention. Guided by a community reinforcement approach to behavior therapy [44], the Web-based intervention aims to prevent and reduce HIV risk behaviors by improving skills, including decision-making and negotiation skills. The Web-based intervention is comprised of a computerized risk behavior assessment, followed by a total of 26 tailored modules. Using a computer at the clinic, the intervention was delivered in 60-minute sessions and the time for completing each module ranged between 10 and 30 minutes. On average, participants completed the sessions in approximately three sessions. Participants assigned to the control condition received two group or individual sessions, each of which lasted approximately one hour. Based on the National Institute on Drug Abuse HIV prevention principles in drug-using populations, the sessions focused on the epidemiology of HIV, promotive strategies to prevent and reduce HIV risk behaviors (including sexual risk and drug use), and HIV testing information [45]. Additionally, participants received a 20-minute youth-centered HIV prevention video. Findings indicated that participants in both conditions reported reductions in past 30-day sex partners (P=.05), however no significant differences in condition x time were observed. Furthermore, findings suggested that participants in both conditions reported improvements in condom use skills (P<.001), and this difference did not significantly differ across conditions two-weeks postintervention [43].

Mason et al [42] examined the efficacy of a text-based motivational interviewing intervention, relative to an attention control condition, on reducing smoking among a sample of predominately African American (90.8%) youth. Using a respondent-driven sampling design, 72 youth were randomized to either the experimental or attention control condition via block randomization, and were assessed at baseline, 1, 3, and 6-months postbaseline. Participants in the experimental condition received a text-based intervention that delivered 30 interactive text messages and three booster text messages over a five-day period that focused on rapport building, tobacco use feedback, social network information and feedback, and summary and plans with respect to the prevention of tobacco use [42]. By integrating motivational interviewing and social network counseling, the intervention aims to prevent and reduce tobacco use by increasing psychological precursors of behavior change, and reducing risk in close peer social networks. The attention control condition consisted of 30 health-based texts, including diet, exercise, and study habits. Findings indicated that participants in the experimental condition reduced the number of cigarettes smoked in the past 30 days, relative to the attention control condition. The results yielded a large and significant effect size, F (1,55)=4.39, P<.01, η²=0.17 [42]. Furthermore, youth in both conditions increased their readiness to change and willingness to refuse cigarettes from their friends, and decreased the number of close friends who were daily smokers, however no significant differences across conditions were found [42].

Mustanski et al [41] examined the preliminary efficacy of the Keep It Up! intervention, compared to an online didactic HIV knowledge condition. Using an intention-to-treat approach, 102 predominately African American (12.7%) and Hispanic (47%) YMSM were randomized via a computerized algorithm stratified by race, and assessed at baseline, immediate postintervention, prior to a 6-week booster session, and 12 weeks postintervention [41]. Participants in the experimental condition received the Keep It Up! intervention. Informed by the information motivation behavioral skills model of HIV risk behavior change, Keep It Up! consists of 7 tailored, interactive modules delivered across three sessions and completed at least 24 hours apart. Keep It Up! aims to prevent and reduce condomless sex via improvements in HIV knowledge, safer-sex self-efficacy, beliefs and intentions to use condoms, and condom use errors.

Participants in the control condition received an HIV knowledge intervention consisting of 7 nontailored, noninteractive modules delivered across three sessions. The didactic intervention included HIV information and facts, including condom use, and the epidemiology of HIV and STIs in YMSM. Findings indicate that, relative to the control condition, participants in the experimental condition reported a 44% lower rate of unprotected anal sex acts at the 12-week follow-up (Risk Ratio=.056, P<0.05). Participants in both conditions showed a large increase in HIV knowledge at postintervention (Cohen’s d for Keep It Up!=0.75; control=0.87), however no significant differences were found across conditions [41]. As shown in Table 2, we computed Cochrane summary of findings to designate the score for each criterion in each article reviewed. All three articles that were reviewed received a high grade with respect to the quality of the evidence.

The present systematic review indicates that although advances in technology-based HIV/STI and drug abuse preventive interventions for African American and Hispanic youth have been made, only three studies that utilized an RCT design and included licit and illicit drug use or sexual risk behavior outcomes were identified. To the best of our knowledge, this is among the first systematic reviews focused on the state of the science on technology-based HIV/STI and drug abuse preventive interventions for African American and Hispanic youth. From this review, we identified four important gaps in the literature. First, there remains the need to employ methodological and statistical rigor, including the use or RCTs, to develop a fuller understanding of the effects of technology-based interventions on HIV/STI risk behaviors. Second, studies examining the long-term effects of technology-based HIV/STI and drug abuse preventive interventions are limited. Third, the identification of how, and for whom, technology-based prevention interventions do and do not work are needed. Finally, a dearth of rigorously tested technology-based HIV/STI and drug abuse preventive interventions for African American and Hispanic youth exist. Addressing these important gaps in the literature may be helpful in moving the field forward.

There remains a need to employ rigorous methodological and statistical analytic approaches to test the effects of technology-based HIV/STI and drug abuse preventive interventions. The RCT remains the standard when examining the impact of interventions [46], yet we identified only three studies that employed an RCT design. Therefore, future research should use RCT designs when examining the effects of technology-based interventions on HIV/STI and drug use among African American and Hispanic youth. Although studies examining the effects of technology-based HIV/STI and drug abuse preventive interventions using RCT designs in the United States are limited, perhaps we can draw from international studies aimed at moving the field forward. In fact, during the review process, we identified five studies conducted internationally that have employed an RCT design when examining HIV/STI risk behavioral outcomes [47-51]. For example, Ybarra et al [47] examined the effects of CyberSenga, a five-hour online healthy sexuality program on past 90-day condom use and abstinence over a six-month period. A total of 366 youth aged 12 years and older from Mbarara, Uganda, were randomized via parallel group with adaptive randomization to CyberSenga, CyberSenga with booster sessions, or a control group [47]. Although no statistically significant results were found among the main outcomes, study findings may help inform the development and testing of future technology-based interventions. For example, it seems that abstinent youth in the CyberSenga group benefited most in the short-term, whereas sexually active participants in the CyberSenga with booster group benefited long-term [47]. Future research to examine whether (and extent to which) these findings hold true in other youth populations is needed.

The long-term effects of technology-based HIV/STI and drug abuse preventive interventions are not well understood. Of the three studies included in this review, a relatively short-term follow-up assessment (ie, range from 2 weeks to 6 months) was used. Therefore, developing a more complete understanding of the long-term effects of technology-based HIV/STI and drug abuse preventive interventions on behavioral outcomes is essential to moving the field forward. Importantly, variations in the types of drug use behaviors and vulnerability to HIV infection exist. For example, alcohol use behaviors alone are not a risk factor for HIV infection. However, alcohol use behaviors increase the risk of engaging in sexual risk behaviors, which in turn enhances vulnerability to HIV infection [52]. Injection drug use (IDU), conversely, is a direct and efficient mechanism for the transmission and acquisition of HIV [53]. Despite this knowledge, we did not identify one single study with IDU as a behavioral outcome. Given the recent rise in opioid use in the United States and links to HIV infections [54], the development and testing of technology-based interventions focused on opioid use behaviors may be warranted.

Beyond the need to establish the long-term efficacy of technology-based interventions, there also remains the need to understand the pathways through which such interventions work. Indeed, scientists have underscored the necessity of developing and testing theory-driven interventions, which can aid in identifying the mechanisms through which preventive interventions have an impact on HIV/STI testing and risk behaviors among youth. Although our review suggests that investigators have developed theory-driven technology-based preventive interventions, the testing of the theoretical underpinnings that guide these interventions is limited. In fact, our review did not reveal a single study that tested for mediation. Therefore, future research should examine the pathways through which technology-based interventions may have an effect on HIV/STI risk behaviors and testing.

There also remains a need to develop a more complete understanding of the populations for which preventive interventions do and do not work [55,56]. That is, investigators examining both malleable and nonmalleable sample and contextual study characteristics may have great utility in working toward optimally efficacious technology-based interventions. For example, some studies indicate that the effects of technology-based interventions varied as a function of gender, suggesting a potential need for gender-specific interventions [48,51,57,58].

Although stark racial and ethnic disparities exist in the incidence of new STI and HIV infections among youth [1,2], our review revealed that the majority of studies did not include samples comprised of >50% racial and ethnic minority populations. Given the facts that African American and Hispanic youth disproportionately engage in HIV/STI risk behaviors [4,5], are disproportionately affected by STI and HIV infections [1,2], and underutilize HIV and STI testing [5], more research on the development and testing of technology-based HIV/STI interventions focusing on African American and Hispanic youth is warranted. Moreover, because African Americans (12%) and Hispanics (13%) are more likely to report being smartphone-dependent compared to non-Hispanic whites (4%) [22], mobile-technology interventions utilizing smartphones may be an effective approach to engaging and retaining African American and Hispanic youth in prevention services.

Findings from the present review may have implications for practice in at least two ways. First, improving clinician-delivered HIV/STI preventive services, including screening for HIV/STI risk behaviors and status, remains a federal priority [18]; however, clinicians face numerous challenges, including interpersonal and structural-level factors. Technology-based interventions have the potential to overcome many of the challenges associated with providing HIV/STI preventive services, including improving youth-clinician HIV/STI risk communication [27], providing access to HIV/STI prevention programs [28], and supporting and enhancing traditional prevention models by supplementing face-to-face intervention efforts [59]. Second, technology-based interventions have the potential to improve clinicians’ delivery of targeted, tailored, and cost-effective HIV/STI preventive services. For example, some technology-based interventions include a youth risk assessment [43,60], which can be used to deliver targeted and tailored HIV/STI prevention services based on the youth’s specific risk behaviors.

Findings from the present review should be interpreted in light of an important limitation. Specifically, we did not include quasi-experimental studies in our review. RCTs may not always be a feasible intervention design, and well-designed quasi-experimental studies may have added to the evidence-base of the present review. We did not include quasi-experimental studies in our review for several reasons. First, the key distinction between quasi-experimental studies and RCTs is the random assignment of participants to either the treatment or control group. This distinction is particularly important during analyses given the increased potential of nonequivalence between groups, yielding challenges in comparisons. Second, the aim of the present review is to establish the efficacy of technology-based HIV/STI and drug abuse preventive interventions for African American and/or Hispanic youth. As such, it was pertinent that our review focus on RCTs, which has been established as the primary method that allows for better causation implications. Finally, in our systematic review, we followed the methodological approach of including only RCTs, as employed in other studies in medical Internet research [61,62].