Search results
The final searches were performed on 16 September 2020, and 5213 papers were identified in total. The removal of 3490 duplicates left 1723 papers to be reviewed for eligibility, 1539 papers were removed following title and abstract screening, with 137 removed following full text screening. In total, 47 papers from 46 trials met our inclusion criteria with the PRISMA flow diagram (Fig. 1) showing details of reasons for exclusions, mainly being the wrong age range (for example, an age range that went above 25). The 47 included papers were spread over the high-risk groups as follows: three focused predominantly on AOD, six on juvenile justice, two on homelessness, five on YMSM, 26 on ethnic minorities, two on mental ill-health, three on out of home care, and finally no trials were identified in LGBTQI + groups outside of YMSM. The 46 included trials had a combined total of 21,543 participants and results are detailed in Table 1.
Nearly all (45/46) trials took place in the USA and tested interventions that were delivered in group settings, ranging from 60 minutes to eight hours, over one week to seven months. Six of the interventions were delivered at one time only. Some included trials had intersectional populations; where this was the case they are grouped according to the primary target population as stated by the authors, with secondary groups noted in relevant sections.
Included trials
Alcohol and other drug (AOD) use
Our review identified three papers describing interventions for youth with AOD-related issues. One was conducted inside a residential treatment facility (24), one in outpatient clinics (25), and one via text message (26) Only Letourneau (25) reported longer-term outcome data (over 12 months), focusing purely on behavioural outcomes. Their group intervention focused on reducing substance use and sexual risk-behaviours for young people who had been referred to a juvenile drug court. It incorporated caregiver involvement in a contingency management program, including elements of cognitive-behavioural therapy to help teens identify the antecedents of their risk behaviours. This intervention had no significant impact on sexual risk behaviours, with the authors highlighting the challenge of addressing common co-occurring ‘problem behaviours’ with one broad approach.
Five other papers identified in this review included young people with substance use problems as secondary groupings, given the commonality of comorbidities between these at-risk groups (Juvenile Justice and Homelessness groups) (27–31). These will be discussed under their primary grouping.
Justice-involved youth
We identified six papers reporting trials that targeted justice-involved youth (27, 28, 31–34). Four of the six trials took place in detention centres or prisons, while one (32) was conducted in foster care homes adolescents had been placed in as part of their ‘treatment’. Two of the trials were single gender only (one each for males (33) and females (32)), and two others also included content aimed at reducing drug and alcohol consumption (27, 28). The papers mostly reported behavioural outcomes such as condom use, pregnancy and STI incidence. One study (27) reported the reduction in STI incidence as a result of their sexual risk reduction group with additional content on alcohol and cannabis use. Kerr and colleagues (32), in a trial with girls aged 13–17, found multidimensional treatment foster care reduced the odds of pregnancy over the subsequent 24 months. The two other trials (28, 33) did not have a significant long-term impact on risky sexual behaviours despite post intervention effects on knowledge, attitudes and condom skills. No data were available to extract for meta-analyses within this high-risk group.
Four additional papers described three trials where some of the participants were involved with the justice system (25, 35–37).
Homelessness
Two papers focusing on homeless youth are included in this review; both were conducted at drop-in centres providing other services for homeless youth, and also included content to reduce AOD use (29, 30). Both trials reported significant findings when an unplanned post-hoc analysis was conducted. Slesnick et al. (29) found their intervention only had a significant impact on condom usage when age was factored into the analysis: intervention group participants aged 14–18 used condoms more frequently at follow up than control group youths aged 19–22. While the intervention used by Tucker and colleagues (30) failed to have an impact on knowledge and attitudes, it did decrease unprotected sexual acts for participants with multiple sexual partners.
LGBTIQ+ (including YMSM)
Five papers focused on YMSM (38–42). Two trials were conducted remotely, one of them online (40) and one via text message (41) as novel approaches to engaging this potentially hard to reach population. The remaining trials occurred in HIV clinics and LGBTIQ + community health centres. One occurred in Thailand (42). All five trials reported behavioural outcomes, typically rates of engagement in protected sex and the number of sexual partners. Overall, results were mixed with some, but not all, trials reporting a significant improvement in sexual health behaviours for participants who took part in an intervention. For example, Chen et al.’s (38) intervention increased participant’s likelihood of using condoms (p < .001) and Rongkavilit et al’s (42) intervention reduced frequency of engaging in anal sex (p = .04) but did not significantly improve condom use. Two of the five trials (39, 40) also reported knowledge and attitude outcomes, however neither intervention produced significant changes in these outcomes.
Mental ill-health
Two papers targeting young people with mental illness were identified; one recruited from mental health outpatient clinics (43), and the other (44) recruited high school students with “emotional or behavioural problems”. Both were three-arm trials and produced mixed results. In Brown et al 2017 (44), the HIV prevention plus affect management (AM) intervention appeared to have more of an effect on sexual behaviours than a skills-based HIV prevention (SB), although both active interventions significantly improved HIV knowledge and condom attitudes at six months follow-up (p < .05). There was no impact on engaging in sexual intercourse with concurrent substance use in either arm of the trial. Brown et al.’s (2014) (43) family-based HIV prevention and adolescent-only HIV prevention interventions both improved sexual behaviours at three months compared to control (p < .05), but the family-based intervention also improved HIV knowledge and parent-teen sexual communication (p < .01).
Two papers included in other high-risk groups also address youth with a history of abuse and/or mental ill-health (36, 45).
Ethnic minorities
There were 26 trials identified that tested a sexual health intervention targeting young people who were considered to be an ethnic minority. All were undertaken in the USA across a variety of settings, from sexual health clinics (46, 47) to community-based youth organisations (48–50), and one provided in the context of a summer basketball camp for American Indian adolescents (51). Most (19/26) were delivered as group-based interventions, two as one-off interventions (47, 52), three interventions were delivered as one-to-one case management support (53–55) and one was a mass media intervention across different cities (56).
Fourteen of the interventions focused specifically on HIV prevention, four on HIV/STI prevention, two focused on preventing pregnancy, three had a focus on intervening at the family level, and the final three focused more broadly on reducing unprotected sex. Newer trials were less likely to focus specifically on HIV prevention, often recognising the need to address and measure sexual health more broadly than the prevention of one disease.
Two papers from other high-risk groups also included ethnic minorities, one targeting homelessness (30), and one for YMSM (40).
Out of Home Care
Three papers were identified from two trials that focused on delivering an intervention to young people in out of home care (35–37), with both recruiting a sample that were also involved in the juvenile justice system. The two trials tested intensive (at least twice weekly) group-based interventions delivered within the out of home care setting. Whilst the older trial (36) focused on HIV/AIDS prevention, (35)’s trial focused on pregnancy as well as HIV and STI prevention. Both trials found long-term (up to 12 months post intervention) significant change in knowledge and attitudes, but non-significant changes in sexual behaviours. (35) also assessed the sexual wellbeing concept of ‘ability to communicate with partner’ with the intervention group reporting a significantly higher ability over the 12 months follow up period.
Meta-analysis
Data were only available to undertake meta-analyses within one high-risk population; ethnic minorities. Data were pooled when available resulting in meta-analyses being completed on the following outcomes; STI/HIV/AIDS knowledge, condom use, frequency of unprotected sex and STI incidence).