Youth experience extensive hormonal, behavioral, and psychological changes in transitioning from adolescence to adulthood (Spear, 2000). During adolescence, many young people engage in sexual risk behaviors (e.g., multiple sexual partners) that contribute to negative health outcomes that impact their life course (Anda et al., 1999; Bernat et al., 2012; Felitti et al., 1998; Norman et al., 2012; Whitfield et al., 2003). From a socio-ecological perspective, many factors influencing adolescent sexual behaviors operate at individual, interpersonal, and institutional levels and thus may be best addressed by multilevel and systemic interventions (Cordova et al., 2020; Ethier et al., 2011; Mayberry et al., 2009; Perrino et al., 2000; Wight et al., 2006; Yildiz et al., 2019).

From 2013 to 2018, the Centers for Disease Control and Prevention ([CDC], 2018)’s Division of Adolescent and School Health (DASH) funded school districts to provide resources and assistance to schools to implement evidence-based approaches through a number of multilevel, multicomponent approaches to prevent student risk behaviors, with a focus on sexual behaviors. The program centered on health education, health service continuum, and safe and supportive environments to address student health from all levels; from individual-level factors (such as teaching skills to help adolescents access health care providers) to policy-level factors (such as policies that govern which services are available and how youth may access them). Although core activities were required of school districts (see Supplemental Table S1 for detailed activities and levels of implementation), districts had broad discretion in which activities were implemented and how they were implemented based on local needs and resources.

Robin et al. (2022) evaluated the effectiveness of the CDC DASH program, and evidence indicated that exposure to the program was associated with significant decreases in ever having sex, having four or more lifetime sexual partners, and being currently sexually active. However, exposure to the program was also associated with decreases in the use of effective hormonal birth control and with no change in condom use, use of both effective hormonal birth control and condoms at last intercourse, and HIV testing. We also evaluated how levels of program implementation affected the program effects and found that implementing more activities to enhance safe and supportive school environments was associated with enhanced program effects in improving student sexual health outcomes (Li et al., 2022).

By design, the DASH multilevel health program aimed to impact student health through implementing the program at every level of the educational system (e.g., school districts, schools, students). Decisions regarding health programs, education, and services are often made at these levels, and each of these levels of influence impacts individual student health behaviors significantly in the form of service and resource allocation. For the multilevel program to be successful, it is important for activities at all levels of the educational system to be coordinated to achieve maximum intervention effects.

To date, little evidence exists to examine the influence of each level of educational systems on student health behaviors (Acosta et al., 2019; Bohanon et al., 2016; Lindberg et al., 2020; Shackleton et al., 2016). As part of a broader evaluation of the CDC DASH program, our aim is to measure the amount of influence of the school and district levels on student sexual health behaviors. Understanding the influence of schools and school districts on student sexual behaviors may help develop or refine interventions for these settings. A secondary aim of this study is to examine any school or district level impacts on variance in these outcomes by racial/ethnic subgroups and by sex. Understanding differences in variance among these groups may allow tailoring of programs to better address their specific contexts.

Understanding the role of each level of the educational system in adolescent risk behaviors is key to effectively allocating intervention efforts and resources to affect risk behaviors at the appropriate levels.

Multilevel interventions are much more effective in achieving desired outcomes than single level interventions, particularly when there is coordinated design and implementation of programmatic elements across levels (Anderson & O’Donnell, 1994; Stokols, 1996). A prominent question when evaluating multilevel interventions is whether the intervention at one level facilitates or hinders the intervention at another level. Assessing this is important because without understanding the effect of each level on intervention outcomes, it is difficult to understand why the intended outcomes were or were not achieved. Additionally, a multilevel intervention is more likely to be transferred successfully to other settings when we know how each level of the intervention affects the intended outcomes. Hence, a major task when evaluating a multilevel intervention is researching multiple levels of influence of the intervention setting.

This study is a first step in understanding how district and school levels affect the CDC DASH program in addressing students’ sexual health outcomes. We examined school- and district-level variance in students’ sexual health outcomes among 16 urban school districts that received CDC DASH funding and found the use of hormonal birth control has considerable variance (10.1%) at the district level while ever having sex, having four or more lifetime sexual partners, and being currently sexually active show considerable variance (around 10%) at the school level. We also found noticeable heterogeneity in those variances across race/ethnic subgroups and sexes. These findings confirm the importance of multilevel interventions to attend to school- and district-level influences that may impact sexual health programming and sheds light on future intervention program design.

In our previous evaluation of the CDC DASH school-based adolescent health program, we found that exposure to the program was associated with reduced use of effective hormonal birth control and not associated with having ever tested for HIV (Robin et al., 2022). The current study provides the evidence that the district-level variance had considerable influence on these two student outcomes (10.1% and 6.1%, respectively). This finding indicates that district-level influences that may include district and community characteristics and state education and health policies and guidance (Guttmacher Institute, 2021; Westbrook et al., 2022) may impact students’ uptake of these protective behaviors to a considerable degree. To date, little research has explored district-level influences on student-level outcomes and few school-based interventions address these influences. Future research is warranted to determine which specific district-level factors influence students’ hormonal birth control use and HIV testing. Such research could provide insights to help strengthen the development of district-level strategies for a more effective school-based multilevel health program.

We also found notable school-level variance in having ever had sex (11.5%), having four or more lifetime sexual partners (12.7%), and being currently sexually active (9.2%). This finding is consistent with previous evidence that school environments are important influences on student sexual risk behaviors (Elkington et al., 2011; Holt et al., 2013; Marschall-Lévesque et al., 2014; Ramirez-Valles et al., 1998; Ritchwood et al., 2015; Teitelman et al., 2008). This finding reinforces the need to strengthen school-level activities including health education, health and mental health services, and increasing student connectedness to their schools to better address students’ sexual risk behaviors (Boyer et al., 2000; Cordova et al., 2020; Fletcher et al., 2008; Li et al. 2022; Mayberry et al., 2009; Perrino et al., 2000; Poteat et al., 2013; Shackleton et al., 2016).

Our findings show noticeable racial/ethnic and between-sex differences in most outcomes, indicating that school and district levels differentially influence students’ sexual health outcomes. For example, the district level had more influence on use of hormonal birth control among non-Hispanic Black, Hispanic/Latino, and other race students than on non-Hispanic White students. Similarly, the district level had more influence on female students’ use of hormonal birth control than on male students’ use. At the school level, there was some evidence of differential impact by race and gender. For example, the school level had more influence on non-Hispanic Black and other race students’ likelihood of four or more sexual partners than for their White and Hispanic peers. Similarly, the school level had more influence on male students’ likelihood of having four or more sexual partners than for female students. Further research on these differential impacts by race and sex could facilitate the identification of effective intervention elements in schools, districts, communities, and state health and education agencies that could help to reduce health disparities by race, ethnicity, and sex.

As part of a larger evaluation of the CDC DASH program, this study examined levels of influence of the educational system on student sexual health outcomes that might be overlooked in a single-level analysis. Specifically, students’ use of effective hormonal birth control and HIV testing showed sizeable district-level variance and their sexual behaviors showed considerable school-level variation. We also found notable heterogeneity in these variances across racial/ethnic groups and sexes.

This study sheds new light on the findings of a previous evaluation of the CDC DASH program and has important implications for future intervention programming. It reveals the need to investigate the specific drivers of district-level influences on use of effective hormonal birth control and having ever tested for HIV and reinforces the needs to incorporate school-level activities to address school-level influences on student sexual risk behaviors. Additionally, depending on the outcome and depending on the racial/ethnic group, school and district administrators, teachers, and staff may need to consider the specific impact of school and district contexts on intended populations when tailoring program efforts.