Efficacy of behavioral interventions to increase engagement in sexual health services among LatinX youth in the United States: A meta-analysis for post-pandemic implications

Abstract

Introduction

LatinX youth in the U.S. are disproportionately affected by HIV and STIs, commonly attributed to a lack of diagnostic testing and regular physician consultations to address sexual health. These disparities have been exacerbated by the COVID-19 pandemic. This meta-analysis seeks to assess the efficacy of behavioral interventions among LatinX youth in the U.S. that aim to increase engagement in sexual health services (i.e., STI/HIV testing, physician consultations).

Content

Following PRISMA guidelines, seven electronic databases were searched. We systematically extracted data with a coding form, and effect sizes were obtained from each study on HIV/STI testing outcomes and physician consultation. Moderator analyses were run for demographic and intervention characteristics.

Summary and Outlook

Of nine included studies, the interventions created a small-to-moderate effect on increased engagement of sexual health services (d ₊=0.204, 95 % CI=0.079, 0.329). Moderator analyses showed that interventions including the following characteristics were most efficacious at facilitating care services: community-based or online setting, access to diagnostic testing, social media/remote components, parental involvement, and longer session duration. This meta-analysis provides informative results regarding behavioral interventions that have proven efficacious in facilitating engagement in sexual health services among LatinX youth. Most prominently, interventions that are remote or through social media, community-based, and incorporated parents had large positive effects. These findings prove useful for the ongoing COVID-19 pandemic situation and provide guidance for targeting LatinX youth to engage them in sexual health services as primary and secondary STI and HIV prevention.

Introduction

Youth aged 14 through 25 comprise only 13 % of the United States (U.S.) population but result in at least 50 % of sexually transmitted infection (STI) diagnoses [1], [2], [3]. Ethnic minority youth such as LatinX youth in particular are disproportionately affected by STIs and human immunodeficiency syndrome (HIV) [4], as disparities among ethnic minority groups such as socioeconomic status contribute to the inaccessibility of youth to receive adequate healthcare for their sexual health needs due to decreased access to and utilization of sexual health services [2, 5]. In 2018, cases of chlamydia and gonorrhea among LatinX youth were nearly two times higher than those of non-Hispanic whites [6]. While there are several sexual health prevention interventions that exist for youth, very few are aimed at increasing HIV and STI testing or emphasizing availability of resources and services available, especially among ethnic minority youth [7]. In fact, existing interventions have left youth with feelings of uncertainty regarding the services that are available to them, and a lack of confidence in the need to routinely check on their sexual and reproductive health [8].

The National Institute of Health (NIH) currently recommends regular STI testing for sexually active adolescents and for health providers to collect a complete sexual history from them, but fulfilling this recommendation poses major challenges if adolescents are unable to speak freely or alone with their providers. Concern over confidentiality between the provider and their parents is a commonly cited barrier to engagement in sexual health services for LatinX youth [2, 5, 9, 10]. All 50 states in the U.S. allow minors to receive STI testing and treatment without parental permission, although 18 of those allow physicians to inform parents about testing or treatment, which can serve as a major barrier to LatinX youth [11]. In 2015, the National Survey of Family Growth found that 12.7 % of youth who reported insurance coverage from a parental plan stated they would not seek sexual and reproductive health checkups due to confidentiality concerns; however, 71.1 % of youth who did have direct physician care received sexual health assessments [12]. Other commonly cited barriers to seeking and receiving sexual health services for LatinX youth include cost of healthcare use and treatment, confidentiality breaches specific to being on their parent’s health insurance, low health literacy, service hours not matching adolescent availability (i.e., 9 am–5 pm), and perceived stigma to STI testing and care [5, 13, 14].

Another factor contributing to the spread of STIs and HIV among youth is that they are commonly experienced asymptomatically. In fact, 60 % of youth are unaware that they are infected with an STI, as many are only tested based on their own – often inaccurate – perceived risk and are not routinely tested for asymptomatic STIs [15, 16]. STI and HIV testing is crucial and recommended by the Centers for Disease Control and Prevention (CDC) as untreated infections may cause long term complications as well as other serious health consequences [17]. Lack of effective treatment for common STIs such as chlamydia, gonorrhea, and syphilis include pelvic inflammatory disease and miscarriages in women, infertility in both men and women; and in some untreated syphilis cases, internal bleeding, seizures, loss of motor function, dementia, or death, among others [18, 19]. Thus, the necessity for youth to visit their physician regularly for STI and HIV testing and preventive exams is paramount for both their and their partners’ long-term health.

Additionally, the COVID-19 pandemic has caused major disruptions in youth sexual health and access to sexual health services [20]. As parental monitoring increased, stay-at-home guidelines were enforced, and other pandemic-related restrictions made accessing healthcare more challenging – regular wellness visits became infrequent, HPV vaccines among youth decreased by 68 %, and STI and HIV routine testing for youth diminished leading to untreated STIs [20]. Further, sexual education which is already limited in many parts of the U.S. was not often included in the national shift to online learning in 2020, and any missed sexual education content was likely left out when in-person learning resumed [20]. In the absence of educational interventions during this time, many youth reported using online sources to help answer their sexual health-related questions, though inaccurate and limited information may have been more harmful than beneficial [20, 21]. Telehealth and virtual interventions became increasingly common and utilized during the pandemic years, though pose several challenges for youth populations including high cost, limited insurance coverage, potential lack of privacy, virtual learning fatigue, and disorganization of services or programs [22], [23], [24], [25], [26], [27], [28]. Further, telephone consultations with providers during the pandemic decreased 100 % for those 13 and under, 52 % for those 14 and 15, and 31 % for those 16 and 17 [27]. Even with these remote options for sexual health services and intervention, youth during the pandemic reported that their preferred sources of learning about sexual health remain doctors, teachers, and family [29]. Thus, even though much has changed since the start of the COVID-19 pandemic, exploring the interventions pre-pandemic remains of critical importance.

Finally, it is important to acknowledge that there are many behavioral interventions for youth aimed at reducing risky sexual behaviors or increasing knowledge regarding sexual health, but fewer aimed solely at increasing HIV and STI testing or engagement in sexual health services [7]. Prevention and treatment through HIV and STI diagnostic testing and routine checkups with providers are just as critical primary and secondary prevention techniques for controlling the high rates of STIs and HIV among LatinX youth populations. Importantly, several of the behavioral interventions include components, information, or outcomes related to HIV and STI testing and/or engagement in sexual health services. These combination interventions can provide youth with a broader knowledge of prevention techniques including both behavioral and clinical techniques; although when taken individually, these interventions tend to have more impact on sexual risk behavior outcomes, but their efficacy for clinical outcomes has not yet been reviewed and synthesized. Thus, this systematic review and meta-analysis aimed to assess the efficacy of pre-pandemic behavioral interventions and programs among LatinX youth that resulted in seeking sexual health services (i.e., physician visit for a sexual health consultation, STI and/or HIV testing) to determine whether combination interventions have positive effects for clinical outcomes, or if these would be better addressed separately.

Materials and methods

Literature search strategy

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to obtain data sources and conduct the systematic review and meta-analysis [30]. With the assistance of a University of Connecticut Allied Health Sciences Librarian (J.L.), we conducted electronic literature searches using PubMed, CINAHL, PsycInfo, SocIndex, Scopus, ERIC, and Cochrane databases in November 2019. The search for all potential studies was comprised of keywords relating to STIs, HIV, sexual education and sexual health, various contraceptives, youth populations, behavioral interventions, and various study designs (including randomized, quasi-randomized, and pretest-posttest designs). To ensure databases were searched exhaustively, search entries included all conceivable variations of the relevant keywords and were adjusted to match the specific search engines of each database (e.g., indexing terms, MeSH terms, and appropriate truncation methods). As initial refined searches missed several relevant papers, we did not restrict our search to LatinX populations in order to avoid excluding studies that contained LatinX in the sample but were not particularly targeting LatinX populations, and developed an extremely thorough testing process to ensure accuracy of study inclusion (search strategy available in Appendix Table 1).

Study selection and criteria

Under the supervision of a senior researcher, the first author and a team of three researchers (A.H., M.M., G.M.) independently evaluated all literature results in order to determine which studies were fit for inclusion. We did not place time limits on included studies, and we included papers published in English or Spanish. Further, we included grey literature by searching within publications and databases for presentations, theses, or other types of published scientific reports. (PRISMA chart in Figure 1).

Figure 1:

PRISMA flowchart for screening and inclusion of studies included in this subset meta-analysis.

Participant demographics: We only included studies that were conducted in the U.S. Based on the Center for Disease Control and Prevention’s age cut-off for youth and young adults being 24 years old, participants in the included studies had to fall between the ages of 11 and 24 years [17]. To ensure exclusive results among LatinX populations, we included studies with exclusively LatinX samples; or if mixed samples, we included studies that reported quantitative results specifically on the LatinX participants as long as there were more than 10 % in the sample. We excluded studies if they targeted unique subpopulations such as individuals with chronic illness, mental illness, homelessness, disabilities, HIV, addiction; and those who are incarcerated, alternative school students, emergency department patients, recent immigrants, engage in sex work, or are men who have sex with men. These subpopulations commonly require uniquely tailored interventions to address their specific risk factors; therefore, they were excluded in order to avoid potential confounders to the results.

Study focus: Each included study had to have an intervention with at least one goal of addressing sexual behavior. Studies reporting quantitative data on outcomes related to getting an STI or HIV test or seeking sexual health services with a physician or at a clinic were included. Due to a small quantity of studies reporting each outcome, and the similar nature of these outcomes (i.e., both related to prevention with medical care, and sexual health service visits leading to or including an STI or HIV test), the outcomes of the studies were combined to the primary outcome of engagement in sexual health services (See below for a more detailed description of outcomes, and Appendix Table 2). As the topic of this research is STI and HIV prevention in the context of healthcare, we focused only on studies related to testing or sexual health service engagement in interventions aimed at preventing STIs or HIV, and did not include studies that focused on reproductive health. We excluded studies if the primary focus of the intervention was related to vaccination, or adherence to Pre-Exposure Prophylaxis (PrEP), antiretroviral therapy, or Post-Exposure Prophylaxis (PEP); as interventions and programs for these biomedical options each contain their own bodies of literature and are not relevant to our primary outcome.

Review methods and statistical analysis

We developed a comprehensive coding form and corresponding manual to uniformly and systematically extract data from the included studies to gather information on study, sample, and intervention characteristics and to complete a study quality assessment. We conducted a pilot test of the coding form, and added any necessary variables for complete coding. The first author and research team (A.H., M.M., G.M., E.K.) used the coding form to extract data from the included studies. The coders discussed any coding discrepancies with another expert until resolved (TBH). All original outcomes were analyzed and tested for interrater reliability using Kappa for categorical variables and Pearson’s r for continuous variables with IBM SPSS Statistics Version 25 [31]. All descriptive statistics extracted with the coding form were calculated using SPSS.

We conducted risk of bias assessment following the guidelines from the Cochrane Collaboration’s tool for assessing risk of bias [32]. Raters scored items on the basis of high (coded by “+”), low (coded by “−“), or unclear (coded by “?”) risk of bias in regards to selection bias, blinding bias, attrition bias, and reporting bias. With these scores, descriptions of level of bias summary ratings were reported categorically. We used the Begg and Egger techniques to test for asymmetries in effect size (ES) distribution which may represent reporting bias [33, 34].

The primary outcome was engagement in sexual health services, which included outcomes related to HIV or STI testing, or physician visits. Across the studies, eight measured these outcomes via surveys, while one was measured via whether an HIV/STI self-test kit was ordered; all outcomes were measured dichotomously. Three studies measured a sexual health service visit exclusively, four studies measured HIV and/or STI tests exclusively, and one study measured both service uptake and testing outcomes. Please refer to Appendix Table 2 for more measurement details across the studies. ESs were calculated for this outcome by calculating the standardized mean difference, d, for each outcome and subgroup using an effect size coding calculator [35, 36]. The calculator uses a factor that controls for small sample sizes when applicable [37]. Then, the ds were combined for the standardized mean difference, which allows results using different metrics to be compared or combined directly, thus allowing this meta-analysis to include studies across a wide array of sites and settings [35, 38]. In some studies, various subsamples or related outcomes were assessed from the same dataset, for example among sexually experienced vs. sexually inexperienced participants (Appendix Table 2), though STI and HIV testing outcomes were only reported for participants who had been sexually active. Further, while 49 % of our total sample was sexually active at baseline, the majority of each sample was sexually active at follow-up. Sensitivity analyses were run eliminating those studies and the results did not change, and there were not enough comparisons from each study to require additional control for dependence.

Weighted mean effect sizes by the inverse of the variance of each study were calculated across all studies under random-effects assumptions, with restricted maximum likelihood variance estimation [39]. We did not use fixed-effects as the random-effects is a more conservative and robust model [40, 41]. Data from the final data collection point, or only available data collection point was utilized. In one case of unusable data due to the format reported in the paper, we contacted the first author, who kindly sent the data in an appropriate format in order for us to calculate the weighted mean effect [42]. To test for heterogeneity, Cochrane’s Q was calculated for significance of heterogeneity, and I ² was calculated for magnitude of heterogeneity [43, 44]. To evaluate the sources of heterogeneity of the ESs, moderator analysis using weighted mixed-effects models with restricted maximum likelihood estimation of the variance was performed, testing each variable for study, intervention, and participant characteristics independently. Statistical models, including moderator analyses, were conducted using the R package Metafor [45]. The moving constant technique was used to produce estimates of the ES (d+) at meaningful levels of the moderators and their confidence intervals at different levels of interest [46]. This technique was used to demonstrate results at the maximum and minimum values of significant moderators. Two-sided statistical significance was p<0.05.

Results

Systematic review

Our search for studies examining interventions that facilitate sexual health services among LatinX youth yielded nine studies, with 10 interventions (i.e., one study contained two different interventions compared to a control group). Publication years ranged from 1999 to 2019, with a median publication year of 2009. The total sample of participants at baseline was 11,990 youth with a mean retention rate of 81.7 % across the studies. Participants ranged in age from 14–22 years old, with a mean age of 17.3 (SD=7.5). Across the studies, there was a mean of 28.9 % (SD=24.3) LatinX participants, and a mean of 49.9 % (SD=30.3) were sexually active at baseline. A detailed depiction of sample demographic characteristics is reported in Table 1.

Table 1:

Demographic Characteristics of the sample reported in nine studies (N=11,900).

Characteristic	Mean (SD) or N (%)	Min	Max
% Female	54 (6.9)	46.2	68
Race/ethnicity
% Latino	28.9 (24.3)	11.9	88
% White	48.2 (28.4)	4	84
% Black	26.1 (25.3)	4	81
% Asian	9.2 (8.1)	2	18
% Other	6.7 (3.6)	2	13
% Sex. active at baseline	49.9 (30.3)	15	100
Age	17.3 (7.5)	14	22
Education of sample
Less than HS	1 (11.1 %)
Current middle or HS	6 (66.7 %)
Expected college completion	2 (22.2 %)

Six studies (67 %) included HIV and AIDS specific education, while only four (44 %) included pregnancy prevention information. A full summary of the studies can be found in Table 2. Nine of the ten (90 %) interventions provided risk reduction strategies for participants (Figure 2). Seven of ten (70 %) interventions incorporated games, quizzes, or cognitive work; sexual communication skills training; and general prevention strategies. Notably, four interventions (40 %) included a testing component where participants were able to acquire an HIV or STI test by choice. Very few interventions incorporated cultural relevance, practicing condom use skills, or condom distribution. A full summary of the interventions can be found in Table 2.

Table 2:

Intervention, demographic, and results summary from the nine studies.

Main author (Pub year)	Intervention setting	Sample	Study design	Intervention name	Recruitment method	Session details	Intervention details	Control condition	Relevant results
Borawski et al. (2009)	Urban Midwest, School	N: 1,357 Mean age: 15.2, (SD=NR) % Latino: 11.9 % Female: 51.8	Group-randomized replication study	Be Proud! Be responsible!	Mandatory health education class	# Sessions: 6 Session length: 50 min Duration: NR	Focus: HIV, STI, pregnancy prevention, safer-sex decision making and practices, abstinence-first philosophy. Activities: Role play, group discussions, speaker session. Information: Drug and alcohol, general condom info, risk-reduction and risk avoidant strategies, refusal and negotiation skills.	General health promotion, topics including nutrition, physical activity, stress reduction.	Students in the intervention group were significantly found to have talked to a health professional about a sex-related matter at the 4-month follow-up compared to controls. Retention=92 %
Coyle et al. (1999)	Northern California, Southeast Texas, Schools	N: 3,677 Age range: 14-16 % Latino: 27 % Female: 53	Randomized trial	Safer choices	Through School	# Sessions: 20 Session length: Class duration Duration: 2 school years	Focus: HIV, STI, and pregnancy prevention. Addresses school and home environments. Activities: Role playing, personal story modeling, parent-student involvement, HIV-positive speaker session, resources provided. Information: Targeted school cultural norms, provided list of resources off-campus for HIV, STI, and pregnancy related services.	A standard, knowledge-based HIV prevention curriculum comparison program	At 7-month follow-up, HIV and STI testing had small, positive effect sizes, though not significant. Retention=95 %
Coyle et al. (2001)	Northern California, Southeast Texas, Schools	N: 3,058 Age range: 13-18 % Latino: 27.5 % Female: 52	Randomized trial	Safer choices	Through School	# Sessions: 20 Session length: Class duration Duration: 2 school years	Focus: HIV, STI, and pregnancy prevention. Addresses school and home environments. Activities: Role playing, personal story modeling, parent-student involvement, HIV-positive speaker session, resources provided. Information: Targeted school cultural norms, provided list of resources off-campus for HIV, STI, and pregnancy related services.	A standard, knowledge-based HIV prevention curriculum comparison program	At 31-month follow-up, HIV and STI testing had larger effect sizes than earlier follow-ups, but still not significant. Retention=79 %
Eastman-Mueller et al. (2019)	Social marketing campaign broadcast on TV, school (high school and college), doctor’s offices, local events, internet; location not specified	N: 1,101 (high school), 1,228 (college) Age range: 15–17 (high school), 18–25 (college) % Latino: 17.5 (high school), 20.8 (college) % Female: 49.4 (high school), 70.8 (college)	Survey of social marketing campaign	Get Yourself tested	Online panel survey	NR	Focus: Promote STI testing and open sexual health communication with health providers and partners. Information: “Get yourself talking” messaging to partners and healthcare providers, STI and HIV testing messages.	NA	HS and college students who saw the campaign were significantly more likely to have spoken with healthcare provider. College students were more likely to test for STI and HIV, HS students were more likely to test for STIs. Retention=32.9 % (high school), 51.2 % (college)
Lemieux et al. (2008)	Urban schools, region not specified	N: 306 Mean age: 16, SD:(NR) % Latino: 37 % Female: 49	Quasi-experimental nonequivalent control group design with pretest and posttest	NA	Enrolled in school health class	Duration: 3 months	Focus: HIV prevention Activities: Listen to HIV prevention songs created and recorded by influential peers (MOLs). Students received CDs and informative HIV-prevention gear from the musical group. Information: Perceptions of social norms, perceived vulnerability to HIV, attitudes towards abstinence and condom use, general HIV prevention information.	Standard-of-care condition of health class. Curriculum included similar content (primarily HIV transmission and prevention info) but did not include the MOL component.	Sexually active participants were more likely to obtain an HIV test in the experimental group compared to controls. Retention=72.5 %
Lustria et al. (2016)	Southeastern University, web-based	N: 1065 Age range: 18–26 %Latino: 18.8 %Female: 46.2	Randomized controlled trial	RU@Risk	Undergraduate courses	Session length: participants encouraged to browse the intervention site for a minimum of 10 min. Duration: one college semester	Focus: STI prevention and testing. Information: STI information. Content tailored to participant’s responses to STI risk assessment, gender, relationship status, sexual activity, and perceived risk. Participants offered free at-home STI test kit at intervention completion.	Site containing general STI information from the CDC website.	Participants in the tailored condition were more likely to order a test kit. Hispanics and females ordered more test kits than white participants, non-Hispanic participants, and males. Retention=NR
Marsiglia et al. (2013)	Southwest, Public University	N: 190 Mean age: 22, SD: (NR) % Latino: 14 % Female: 68	Quasi-experimental pretest and posttest comparison-group design	Course name: HIV/AIDS: Science, Behavior, and Society	University course	Session length: University course time Duration: 15 weeks (semester course)	Focus: HIV prevention. Information: HIV-themed college course that focused on HIV information and risk behaviors, and discussed relationships, sexuality, and experiential learning techniques.	Sociology course (Sociology of Marriage and Family), addressed similar themes as HIV/AIDS course, but did not incorporate experiential learning techniques.	Although positive, there were no significant effects on the odds of getting HIV or STI tested in either group. Retention=53 %
Rohrbach et al. (2015)	South and East Los Angeles, California schools	N: 1447 Mean age: 14.2, SD: (0.58) % Latino: 87.9 % Female: 51	Cluster-randomized trial	The Sexuality education initiative	Through School	# Sessions: 12 Session length: 50 min (class period) Duration: 7.5 weeks	Focus: STI and pregnancy prevention, improving students’ ability to manage their sexuality respectfully. Information: Focus on gender norms and power dynamics in relationships; media messages; sexual and reproductive anatomy; HIV, STI, and pregnancy prevention; contraception.	3 session, 9-day basic sexual education curriculum. Covered anatomy and prevention of unintended pregnancy and STIs.	Students in the intervention group were more likely to report the use of sexual health services compared to controls. Retention=78.3 %
VanDevanter et al. (2005)	Harlem, New York; Prince George’s County, Maryland; community-based organizations	N: 222 Age range: 12–21 % Latino: 12.3 % Female: 65.5	Randomized controlled trial	Check out that body	Recruited from community-based organizations	# Sessions: 3 Session length: 90 min Duration: 3 months	Focus: Increase sexual health care seeking. Activities: Discussions, brainstorming games, role-play, homework assignments, practice communication skills, physician led discussion, formulate detailed plan for scheduling health check-ups. Information: Mental health, STIs, substance use, STI testing, benefits of primary care/reproductive health visits, increase self-efficacy, skills and intentions to seeking health care.	1-Session self-esteem building condition	Females in the intervention group were significantly more likely to have scheduled a healthcare appointment or undergone a checkup compared to controls. Retention=71 %

NR, Not Reported; NA, Not Applicable.

Figure 2:

Distribution of the frequency of intervention components across the 10 interventions.

For all content coding, interrater reliability was 0.88 (k=0.81). Additionally, the Egger’s and Begg’s test for all outcomes suggest no significant publication bias (Egger’s p=0.198, Begg’s p=0.548). The study quality assessment of the nine studies can be found in Figure 3. Across studies, the majority had some bias concerns, and only one had low concerns. The studies that had some or high levels of bias concerns typically raised concerns regarding selection bias and blinding which is not incredibly concerning based on the nature of the studies (e.g., youth interventions conducted in school typically have difficulty blinding the participants; likewise, selection bias tends to be high when participants are convenience samples). Attrition bias was low in most studies, and reporting bias was mainly low or with some concerns.

Figure 3:

Risk of bias assessment.

Meta-analysis

As noted in the forest plot (Figure 4), there was a small to moderate significant overall effect of the interventions on participants’ engagement in sexual health services (d ₊=0.204, 95 % CI=0.079, 0.329). There was significant heterogeneity across the interventions (Q=181.69, p<0.0001) and a great amount of variability (I ²=92.1 %).

Figure 4:

Forest plot displaying the overall effect size and individual effect sizes for related sexual health care service outcomes among 10 interventions. Each horizontal line represents the 95 % confidence interval of the overall effect size for each corresponding study, outcome, and subgroup (where applicable). The black boxes represent the point estimate of the study; the larger the box, the larger the sample size of that study. The null line (dotted line) through the forest plot separates whether the effect favored the control or intervention groups (e.g., black boxes that fall to the right of the null line represent an effect that favors the intervention group). Horizontal lines that do not cross the null line represent a significant effect of the corresponding intervention. The diamond represents the point estimate and confidence intervals when you combine and average all of the individual studies together. The horizontal points of the diamond represent the limits of the 95 % confidence intervals; thus, since the diamond does not touch or cross the null line, it can be interpreted as a statistically significant overall effect of the studies.

Moderator analysis (Table 3) found that the program site of the intervention significantly influenced whether participants sought sexual health services (β=0.126, 95 % CI=0.044, 0.209). Interventions conducted in a community-based setting (d ₊=0.441, 95 % CI=0.124, 0.758) or online (d ₊=0.321, 95 % CI=0.069, 0.573) had the greatest positive significant effects compared to interventions conducted in schools (d ₊=0.07, 95 % CI=−0.052, 0.192). Additionally, the geography of the intervention influenced effect size; interventions conducted in exclusively urban areas had a moderate significant effect on facilitating sexual health visits (d ₊=0.295, 95 % CI=0.068, 0.522), compared to those conducted in rural areas or a combination of both urban and rural areas. Finally, interventions had the largest positive effect among participants who had discontinued high school education (d ₊=0.437, 95 % CI=0.114, 0.761), compared to current middle or high school students (d ₊=0.072, 95 % CI=−0.060, 0.205), or college students (d ₊=0.224, 95 % CI=−0.007, 0.454).

Table 3:

Moderator analysis of interventions that facilitate seeing a healthcare physician or getting HIV or STI tested among Latino youth aged 14–22.

Moderator		Category	n*	d+ (95 % CI)	β (95 % CI)
Setting and demographics	Geography				0.005 (−0.097, 0.107)
		Urban (ref)	8	0.295 (0.068, 0.522)*
		Rural & urban	10	−0.106 (−0.406, 0.194)
	Program site				0.126 (0.044, 0.209)**
		School (ref)	13	0.070 (−0.052, 0.192)
		Community-based	4	0.441 (0.124, 0.758)**
		Online	6	0.321 (0.069, 0.573)*
	Education				−0.063 (−0.201, 0.075)
		Middle or HS student (ref)	13	0.072 (−0.060, 0.205)
		Less than HS	4	0.437 (0.114, 0.761)**
		In college	6	0.224 (−0.007, 0.454).
Intervention components	Presence of art in intervention				−0.269 (−0.496, −0.043)*
		Yes	11	0.075 (−0.320, 0.470)
		No	12	0.344 (0.176, 0.512)***
	Technology focus				0.226 (−0.042, 0.493).
		Yes	7	0.365 (−0.042, 0.772)
		No	16	0.139 (0.000, 0.279)*
	Exposure to people living with HIV				−0.298 (−0.522, −0.074)**
		Yes	10	0.056 (−0.330, 0.443)
		No	13	0.354 (0.191, 0.517)***
	Presence of testing				0.356 (0.156, 0.557)***
		Yes	12	0.399 (0.071, 0.727)
		No	11	0.043 (−0.085, 0.170)
	HIV/AIDS education present				−0.350 (−0.556, −0.145)*
		Yes	12	0.054 (−0.314, 0.422)
		No	11	0.404 (0.242, 0.566)*
	Session length (minutes)				0.013 (0.002, 0.025)*
		50 min (min)	4	−0.045 (−0.311, 0.221)
		90 min (max)	4	0.487 (0.126, 0.848)**

*n=number of comparisons across interventions; *=p<0.05, **=p<0.01, ***=p<0.001.

For the intervention components, those that included some type of art or theater (i.e., music-based intervention, HIV prevention art show at school, school play about sexual situations), exposed participants to people living with HIV (e.g., videos or speeches from people living with HIV about their experiences), or had HIV and AIDS education had smaller effects on outcomes compared to those that did not, and the differences were statistically significant (β=−0.269, 95 % CI=−0.496, −0.043; β=−0.298, 95 % CI=−0.522, −0.074; β=−0.350, 95 % CI=−0.556, −0.145; respectively). Interventions that primarily focused on technology (i.e., social media or online interventions) had a larger positive effect compared to those that did not, although not significantly (d ₊=0.365, 95 % CI=−0.042, 0.772). Interventions that had a testing component (i.e., provided details or resources on where to get tested, provided the option for free testing) had a moderate effect on the outcome (d ₊=0.399, 95 % CI=0.071, 0.727). Lastly, session length ranged from 50 to 90 min in duration. Interventions that were 90 min had a larger effect than those that were shorter (d ₊=0.487, 95 % CI=0.126, 0.848).

Discussion

This meta-analysis quantitively assessed the clinical outcomes of behavioral STI and HIV prevention interventions and programs to determine their efficacy to engage LatinX youth in sexual health services such as HIV or STI testing, or a physician consultation. Results showed that interventions were effective in creating a small to moderate overall effect on service engagement. Interestingly, moderator analyses showed that interventions conducted in community-based or online settings were more effective than those conducted in schools. We also found that certain intervention components, such as those with testing components present, technology utilization, and longer session lengths, were most efficacious.

One of our most applicable findings is that interventions conducted online, with a technology focus, or through media resulted in larger effect sizes for our outcome. Five of our included studies had technologic components and elicited positive outcomes, ranging from online sexual health marketing campaigns to highly tailored online STI testing interventions [47, 48]. Our findings suggest that use of technology, virtual methods, or social media may prove effective tools for engaging LatinX youth in sexual health services, particularly in the current global climate during and post-COVID-19 pandemic. However, as telehealth, home-based STI testing, and contraceptive delivery programs were developed during the pandemic, complications with these strategies for youth highlighted the urgent need for more effective organization and development of youth-friendly and easily accessible sexual health services [23, 49]. While virtual interventions and learning may often be a great alternative to in-person methods, a survey on ethnic minority youth in the U.S. found that only 13 % actually prefer to learn about sexual health via online methods [29]. Consistent with other research among youth in general, our results also support efficacy in social media campaigns to increase STI testing [5, 50], [51], [52], although this has not yet been tested for this outcome among LatinX youth.

Our findings also suggest that longer session duration (90 vs. 50 min) may be more efficacious at facilitating testing and healthcare visits, posing the question of whether the longer session length is truly necessary for a positive effect among youth. Once examining the studies with the longer sessions, it is clear this may not be the case, and that the positive outcomes may be attributed to the content and techniques of the longer interventions as opposed to the actual length of the session. For example, in one of our included studies, the researchers conduct a community-based intervention that included 90 min skill-building workshops for youth [53]. Local youth were recruited to assist in the planning of the sessions and the sessions were gender-specific, highly interactive, and included valuable information about testing and wellness visits (e.g., confidentiality laws, insurance eligibility, and expectations for the preventive health visit). Before intervention completion, participants were asked to formulate a detailed plan for scheduling check-ups with a healthcare provider, and results were significantly promising for making a sexual health check-up appointment. These topics were not explicitly addressed in other interventions, especially those conducted in schools; thus addressing insurance barriers and expectations for physician visits may be the cause for the larger effects of this intervention. Indeed, a study among LatinX teens found that participants cited confidentiality and insurance as the biggest barriers to seeking preventive care [54]. Participants noted that if there were clinics available exclusively for teens, with friendly staff, and ensured confidentiality in a private setting, they would feel more comfortable seeking care. Further, those who are uninsured and health illiterate are more likely to have never accessed healthcare, therefore an intervention that addresses these barriers and provides solutions to overcome them may be efficacious among LatinX youth [54]. The fact that this intervention was community-based as well may have contributed to higher efficacy, as community-based STI testing and treatment programs for youth have been previously identified as a promising approach [2].

Finally, it is important to note that consistent with prior findings from a 2012 systematic review on sexual health behavioral interventions among U.S. LatinX youth, our updated review still found that few interventions were culturally tailored to LatinX culture, or included components of parent-child activities or communication [55]. Among Latino cultures, the positive influence of family in healthy decision making has been widely cited. For example, a survey among LatinX teens found that they most frequently report going to their parents or the internet for sexual health questions, and that getting the correct information on sexual health and related topics from a trusted source such as parents was very important [54]. Since the COVID-19 pandemic and even with the many advances in remote learning, this preference for parental guidance has not changed [29]. Two of our included interventions had a parent education portion of the intervention that included student-parent homework activities to facilitate communication regarding HIV, STIs, and pregnancy [56, 57]. The intervention did increase the extent to which students talked with parents in terms of prevention, but we could not include this in our moderator analysis because the other papers did not include similar components, and two interventions did not provide sufficient power. While confidentiality and privacy from parents can act as a barrier to sexual health services engagement, approaches that incorporate parents in adolescent healthcare delivery have been proven useful and are encouraged and supported by adolescent-focused organizations and current national guidelines. Given the family value in LatinX cultures, this may be a particularly useful strategy for this population, as long as youth still have the privacy to discuss their sexual health alone with their provider [2, 58–65].

While this meta-analysis is the first to empirically analyze the effects of behavioral interventions on clinical outcomes such as LatinX youth’s engagement in sexual health services, it does have some limitations. First, we excluded papers that focused on specific population subgroups that, due to their unique risk factors, often require their own tailored interventions. These criteria excluded a large quantity of interventions among LatinX youth that could have provided additional data; however, the exclusion of these studies allowed us to provide a more accurate synthesis of the interventions that would best serve LatinX youth in general. It is important to note that we did not search for studies that were exclusively created for and tested on LatinX populations; rather, sought to test the efficacy of general youth interventions on LatinX youth populations, to see if culturally tailored interventions would be more beneficial for LatinX populations. Thus, not all included studies had exclusively LatinX populations. Additionally, as nine studies fit our inclusion criteria, this adequate but smaller sample size limited our ability to run analyses on a variety of moderators. As a result, we were unable to present the results of several moderator analyses that were underpowered, inhibiting us from drawing reasonable conclusions. Additionally, each study’s outcomes were measured slightly differently, although this is common across meta-analyses and we were able to use these outcomes as the closest proxy to engagement in sexual health services. Lastly, one included study had a one-group design, compared to the rest that had two-group designs. To test any violations from this discrepancy, sensitivity analyses were run removing the one-group study and there were no changes to the main models.

Despite these limitations, this review has notable strengths. It is, to our knowledge, the first meta-analysis focusing on HIV and STI testing and physician consultation outcomes from behavioral interventions among LatinX youth, a vulnerable population disproportionately affected by STIs and HIV. This meta-analysis is unique in that it examined many different studies (evidenced by the high amount of heterogeneity), but with the coding calculator tool we were able to standardize the effect sizes and compare across a variety of different settings and sites for the interventions, as opposed to only being able to compare homogenous trials. Lastly, our meta-analysis is methodologically sound, as we followed the PRISMA guidelines, which presents more rigorous science.

Conclusions

This meta-analysis provides informative results to the scientific community regarding behavioral interventions that have proven efficacious in facilitating clinical outcomes, such as engagement in sexual health services among LatinX youth. Most prominently, interventions that are remote or through social media had large positive effects; interventions that were community-based, and that incorporated parents in some way were also efficacious. These findings are particularly useful for the ongoing COVID-19 pandemic situation and can provide guidance for targeting at-risk LatinX youth and engaging them in sexual health services. Our findings emphasize the importance of STI and HIV testing and sexual health service engagement as primary and secondary prevention techniques among youth, and the importance of not solely focusing on behavioral risks and prevention. Healthcare providers, educators, policy makers, clinicians, and researchers may find these results beneficial in creating novel interventions or adapting existing interventions that utilize identified effective components to increase testing and sexual health service engagement.

Supplementary Material

Supplementary Material

Supplementary Material

This article contains supplementary material (https://doi.org/10.1515/ijamh-2022-0113).