Associations between state-level high school human immunodeficiency virus education policies and adolescent risk behaviors

Abstract

BACKGROUND:

School-based HIV education can reach most adolescents, but inconsistencies exist in state-level content policies. The purpose of this study was to evaluate the associations between state-level high school HIV education policies and adolescent HIV risk behaviors.

METHODS:

This was a cross-sectional analysis of the 2019 Youth Risk Behavior Survey linked to the Guttmacher Institute Sex and HIV Education report. Logistic regression models examined the associations of state-level HIV education mandates and content policies with three HIV risk behaviors: 1) four or more lifetime sexual partners; 2) substance use before last sex; 3) condomless last sex.

RESULTS:

Across 33 states, 128,986 high school students were included. Multivariable adjusted models demonstrated no associations between mandated HIV education and risk behaviors. Covering abstinence along with other safe sex options was associated with lower odds, whereas stressing abstinence was associated with higher odds of at least four lifetime sexual partners and condomless last sex. Discriminatory sexual orientation content was associated with increased condomless last sex; associations for all HIV risk behaviors were stronger among sexual minority youth (SMY).

CONCLUSIONS:

Increased HIV risk behaviors associated with state policies stressing abstinence or requiring discriminatory sexual orientation content supports the need for comprehensive and inclusive HIV education.

HIV education delivered in school has the ability to inform the majority of United States youth about HIV risk because 96% of adolescents attended public or private high school in 2018.¹ However, 21% of new HIV diagnoses in the United States occur among adolescents and young adults.² Of those, 97% acquired HIV through sexual contact.² Congruently, a 2019 survey of high school students found 46% of sexually active students did not use a condom during their last sexual intercourse encounter—a 9% increase since 2003.^3,4

The benefits of school-based HIV education include reduction of a range of sexual risk behaviors among adolescents: delaying age at first sexual intercourse,⁵ reducing number of sexual partners,^5,6 decreasing drinking or drug use before sexual intercourse,^5,7 and increasing condom use.^5–7 However, these benefits can be attenuated depending on the HIV education curriculum content and delivery; for example, abstinence-only education has been associated with greater incident HIV among adolescents.⁸

Content and delivery of HIV education curricula are governed by state policies. For example, state policies govern how content surrounding condoms, abstinence, and sexual orientation is taught. While some states require condom use to be discussed, others do not.⁹ Regarding abstinence, some states require abstinence-only education while others cover abstinence among other safe sex options.⁹ When it comes to discussing sexual orientation, some states require an inclusive tone regarding sexual orientation while other states require a positive emphasis on heterosexuality and/or a negative tone regarding homosexuality.⁹ It is important to understand the role of state policies in shaping HIV education and HIV risk behaviors that result.

It was the purpose of this study to evaluate the associations between state-level high school HIV education policies and adolescent HIV risk behaviors in a 2019 United States nationwide sample of high school students. The overall hypothesis was that state policies that mandated HIV education would be associated with a decrease in three HIV risk behaviors: 1) four or more lifetime sexual partners; 2) drug or alcohol use before last sexual encounter; 3) no condom use at last sexual encounter. Among states with mandated HIV education, state policies that did not support comprehensive content including condom education, covering more than abstinence as a safe sex option, and inclusive sexual orientation content we hypothesized would be associated with increased HIV risk behaviors.

METHODS

Participants

The Youth Risk Behavior Survey (YRBS) is a national survey that is conducted biennially by the Centers for Disease Control and Prevention (CDC) to collect health-related behavior data from students in grades nine through twelve attending United States public and private schools.¹⁰ The version of the 2019 YRBS used in this study was delivered on a state, territorial, tribal, and large urban school district level by health or education departments using a two-stage, cluster sample design to include a representative sample of high school students in each jurisdiction.¹⁰ The YRBS provided demographic information including age, sex, race, and grade as well as sexual orientation and sexual risk behavior information on each student. State-level policy data including mandates on HIV education, condom content, abstinence content, and sexual orientation content were retrieved from the 2018 Guttmacher Institute Sex and HIV Education report.⁹

Procedure

Identifying states

Domains of state-level policies from the Guttmacher Institute Sex and HIV Education report were linked to the YRBS by state.¹⁰ Eight states did not have 2019 YRBS data available and were excluded. Of these, four states (Washington, Oregon, Wyoming, and Minnesota) did not conduct a YRBS in 2019; two states (Delaware and Indiana) collected survey data that were not representative of their state student populations per CDC criteria for response rates and nonresponse bias analyses, and two states (Massachusetts and Ohio) did not give CDC permission to distribute their files. Of the 42 states that gave the CDC permission to distribute their data files, 38 states (excluding Georgia, Louisiana, Tennessee, and Utah) gathered data on at least one of three possible YRBS questions on HIV risk behaviors. Five additional states (Alaska, Idaho, Kansas, Montana, and South Dakota) did not collect sexual identity data and were excluded (Figure 1, Figure 2).

Figure 1.

Flow diagram of state exclusion criteria for primary and secondary analyses.

Figure 2.

Map depicting states included in primary, secondary, and stratified analyses.

Exposures

Our first analysis (Analysis 1) exposure was a dichotomous variable representing if the state mandated HIV education; 33 states were included.

Our second analysis (Analysis 2) only included students in states with mandated HIV education; 29 states were included (Arizona, Colorado, Mississippi, and Nebraska did not mandate HIV education). Within states with mandated HIV education, three HIV content exposures were considered: 1) included condom education, 2) if abstinence was stressed, and 3) tone of sexual orientation content. Condom education was a dichotomous variable for mandated inclusion in HIV education or not. Abstinence content had three categories: no policy, covering abstinence along with other safe sex options, or stressing abstinence. Sexual orientation content had three categories: no policy, inclusive content, or discriminatory content (positive emphasis on heterosexuality and/or a negative tone on homosexuality).⁹

Outcomes

The outcomes of interest were coded from three HIV risk behavior questions asked in the YRBS: “During your life, with how many people have you had sexual intercourse?”, “Did you drink alcohol or use drugs before you had sexual intercourse the last time?”, and “The last time you had sexual intercourse, did you or your partner use a condom?” The responses to these questions were coded dichotomously with HIV risk defined as: (1) having at least four lifetime sexual intercourse partners, (2) used drugs or alcohol before last sexual intercourse, or (3) did not use a condom during last sexual intercourse. These three outcomes were analyzed individually and as a dichotomous composite outcome representing if YRBS respondents had indicated engagement in at least one of the HIV risk behaviors.

Covariates

Demographic covariates from YRBS included sex, race/ethnicity, grade, and sexual identity. Sex was coded as a dichotomous variable limited to YRBS response options of male and female. Race/ethnicity was coded categorically as White, Black/African American, Hispanic/Latino, and all other races. Grade in school was coded as grades nine through twelve. Sexual identity was coded as heterosexual, gay or lesbian, bisexual, and unsure.

Data analysis

Chi-square tests assessed differences in descriptive statistics by the composite HIV risk behavior variable. Logistic regression assessed the association of state mandate and content exposures with individual HIV risk behavior outcomes and the HIV risk composite outcome. We used stratified analyses to assess the associations between the tone of sexual orientation content and HIV risk behaviors at the different levels of sexual identity (heterosexual, gay or lesbian, bisexual, and unsure) because there was a significant interaction between sexual identity and sexual orientation content.

All statistical analyses were done using Stata 15.1 SVY procedures to account for the complex sampling design of the YRBS, and results were weighted to reflect estimates of the included states.

Sensitivity analysis

A post-hoc sensitivity analysis was done to study the more global relationship between HIV education content and HIV risk behaviors. A composite dichotomous variable was created for states with comprehensive and not comprehensive HIV education content. Comprehensive HIV education content was defined as that which mandated condom education, covered more than abstinence as safe sex options, and required inclusive sexual orientation content. Non-comprehensive HIV education was defined as that which did not mandate condom education, stressed abstinence, and required discriminatory sexual orientation content. Of the states included in this study, four (California, Connecticut, Maryland, New Mexico) had comprehensive HIV education. Two states (Florida and Oklahoma) had non-comprehensive HIV education. Logistic regression tested the association of comprehensiveness of HIV education with HIV risk behaviors in students within the included states.

RESULTS

Analysis 1: Comparing HIV risk behaviors among states with mandated and not mandated HIV education

There were 128,986 respondents representing 8,267,520 high school students from 33 states (Table 1). Just over half were female (n=67,735, 51.2%). Almost half identified as White (n=73,087, 47.6%), 12.6% (n=14,160) identified as Black/African American, 29.1% (n=25,227) identified as Hispanic/Latino, and 10.7% (n=19,298) identified as other races/ethnicities. While each grade accounted for approximately 25–26% of the study sample, there were significant differences in HIV risk behaviors between grades. Larger proportions reported at least one HIV risk behavior with each subsequent higher grade (Grade 9 n=3,687 [12.0%], Grade 10 n=5,945 [20.2%], Grade 11 n=8,455 [28.4%], Grade 12 n=9,496 [39.5%]; p<0.01). Among the study population, 84.5% of students identified as heterosexual (n=108,216), 2.5% identified as gay or lesbian (n=3,229), 9.1% identified as bisexual (n=12,115), and 4.0% were unsure (n=5,426). Of the study population, 21.4% (n=27,583) reported engaging in at least one HIV risk behavior.

Table 1.

Descriptive statistics of baseline demographic characteristics of study population and exposures by HIV risk

HIV Risk Behavior
	Total (n=128,986)		No (n=101,403)		Yes (n=27,583)
Demographic characteristics	na	%b	n	%	n	%	P
Sex
Female	67,735	51.2	53,319	51.1	14,416	51.5	0.68
Male	61,251	48.8	48,084	48.9	13,167	48.5
Race/ethnicity
White	73,087	47.6	57,818	47.6	15,269	47.7	<0.01
Black/African American	14,160	12.6	10,855	12.1	3,305	14.4
Hispanic/LatinX	24,035	29.1	18,248	28.9	5,787	30.1
All other races	17,704	10.7	14,482	11.5	3,222	7.8
Grade
9th	36,102	26.1	32,415	30.4	3,687	12.0	<0.01
10th	34,565	24.8	28,620	26.2	5,945	20.2
11th	31,859	24.5	23,404	23.3	8,455	28.4
12th	26,460	24.6	16,964	20.1	9,496	39.5
Sexual orientation
Heterosexual	108,216	84.5	86,307	85.6	21,909	80.7	<0.01
Gay or Lesbian	3,229	2.5	2,058	2.0	1,171	4.0
Bisexual	12,115	9.1	8,518	8.1	3,597	12.2
Not sure	5,426	4.0	4,520	4.3	906	3.1
Exposures
Mandated HIV education
No	5,196	8.0	4,100	8.1	1,096	7.5	0.24
Yes	123,790	92.0	97,303	91.9	26,487	92.5
Condom educationc
No	39,324	40.2	30,743	40.1	8,581	40.8	0.55
Yes	84,466	59.8	66,560	59.9	17,906	59.2
Abstinence contentc
No policy	6,189	7.0	1,522	6.7	4,667	7.1	<0.01
Coverd	68,603	25.4	13,875	19.4	54,728	27.2
Stress	48,998	67.6	11,090	74.0	37,908	65.6
Sexual orientation contentc
No policy	66,681	37.2	52,226	36.8	14,455	38.7	0.05
Inclusive	44,112	24.4	34,887	25.2	9,225	21.5
Discriminatory	12,997	38.4	10,190	38.0	2,807	39.8

^aSample size

^bAll percents are weighted

^cAmong those with mandated HIV education = yes (n=123,790)

^dState policy mandates covering abstinence along with other options

A majority of students were from states that mandated HIV education (n=123,790, 92.0%). Table 2 shows that after controlling for sex, race/ethnicity, grade, and sexual identity, whether or not states mandated HIV education had no significant association with the composite outcome of reporting at least one HIV risk behavior nor with any of the individual HIV risk behaviors (four or more lifetime sexual partners, substance use before last sexual intercourse, and condomless last sexual intercourse.)

Table 2.

Associations between HIV education/HIV education subgroups and HIV risk behaviors

	HIV risk behavior			≥4 lifetime sexual partners			Substance use at last sex			No condom use at last sex
	AORa	95% CI	P	AOR	95% CI	P	AOR	95% CI	P	AOR	95% CI	P
Mandated HIV education
No	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
Yes	1.10	(0.98, 1.23)	0.11	0.99	(0.86, 1.15)	0.91	0.92	(0.77, 1.09)	0.33	1.07	(0.97, 1.19)	0.16
HIV education content
Condom education
No	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
Yes	1.02	(0.93, 1.13)	0.66	1.07	(0.92, 1.25)	0.35	0.96	(0.84, 1.10)	0.57	1.02	(0.93, 1.12)	0.63
Abstinence content
No policy	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
Cover	0.78	(0.65, 0.93)	0.01	0.76	(0.54, 1.07)	0.11	0.94	(0.70, 1.26)	0.66	0.73	(0.62, 0.85)	<0.01
Stress	1.21	(1.08, 1.35)	<0.01	1.20	(1.03, 1.39)	0.02	1.18	(0.97, 1.42)	0.09	1.19	(1.06, 1.34)	<0.01
Sexual orientation content
Inclusive	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
No policy	1.21	(1.01, 1.46)	0.04	1.16	(0.85, 1.57)	0.35	1.16	(0.84, 1.62)	0.37	1.30	(1.11, 1.51)	<0.01
Discriminatory	1.21	(0.99, 1.48)	0.07	1.33	(0.98, 1.81)	0.07	1.19	(0.86, 1.64)	0.29	1.28	(1.07, 1.53)	<0.01

Note: AOR = adjusted odds ratio; CI = confidence interval

^aAdjusted for sex, race/ethnicity, grade, sexual identity

Analysis 2: Analyzing HIV risk behaviors by HIV education content policy domains in states that mandate HIV education

Of the total sample, 92% (n=123,790; representing 7,606,139 high school students) were from the 29 states that mandated HIV education. Almost two-thirds (n=84,466, 59.8%) were in states where condom education was required. Looking at abstinence content, 25.4% (n=68,603) were from states that covered abstinence among other safe sex options, 67.6% (n=48,998) were from states that stressed abstinence, and the remaining 7.0% (n=6,189) had no policy on abstinence content. Lastly, 24.4% (n=44,112) of students were from states that required inclusive sexual orientation content, 38.4% (n=12,997) were from states that required discriminatory content, and 37.2% (n=66,681) were from states with no policy (Table 1).

Condom education

Logistic regression showed there were no significant associations between mandated condom education and the composite HIV risk behavior variable or any of the individual HIV risk behaviors (Table 2).

Abstinence content

Content covering abstinence along with other safe sex options was associated with 22% lower odds of reporting at least one HIV risk behavior (AOR 0.78, 95% CI 0.65–0.93, p=0.01), while content stressing abstinence was associated with a 21% higher odds of at least one HIV risk behavior (AOR 1.21, 95% CI 1.08–1.35, p<0.01) compared to having no abstinence content policy. Looking at individual HIV risk behaviors, covering abstinence in addition to other safe sex options was protective of not using condoms during last sexual intercourse (AOR 0.73, 95% CI 062–0.85, p<0.01), but was not associated with having at least four lifetime sexual partners nor substance use before last sex. Requiring HIV education to stress abstinence was associated with 20% higher odds of having at least four lifetime sexual partners (AOR 1.20, 95% CI 1.03–1.39) and 19% higher odds of no condom use at last sex (AOR 1.19, 95% CI 1.06–1.34, p>0.01). There was no association between policies stressing abstinence and substance use before last sex (Table 2).

Sexual orientation content

Compared to inclusive sexual orientation content, having no policy was associated with 21% higher odds of at least one HIV risk behavior (AOR 1.21, 95% CI 1.01–1.46, p=0.04) and 30% higher odds of no condom use at last sex (AOR 1.30, 95% CI 1.11–1.51, p<0.01). Mandating discriminatory sexual orientation content was associated with 28% higher odds of no condom use at last sex (AOR 1.28, 95% CI 1.07–1.53, 95% CI <0.01). There was no significant association with sexual orientation content and having four or more lifetime sexual partners nor substance use before last sex (Table 2).

Effect modification of sexual identity with the tone of sexual orientation content

There was a protective effect of discriminatory sexual orientation content on substance use before last sex in students who identified as gay or lesbian (AOR 0.30, 95% CI 0.15–0.61, p<0.01).

Among students who identified as bisexual, there were significantly higher odds of at least one HIV risk behavior among students from states that had no sexual orientation content policy or that mandated discriminatory sexual orientation content compared to students from states that mandated inclusive sexual orientation content (no policy: AOR 1.46, 95% CI 1.08–1.96, p=0.01; discriminatory: AOR 1.39, 95% CI 1.00–1.94, p=0.05). The largest association with HIV risk behaviors among students who identified as bisexual were seen with policies mandating discriminatory sexual orientation content; there was 97% higher odds of at least four lifetime sexual partners (AOR 1.97, CI 1.22–3.18, p=0.01), 76% higher odds of substance use before last sex (AOR 1.76, 95% CI 1.10–2.82, p=0.02), and 51% higher odds of no condom use at last sex (AOR 1.51, 95% CI 1.04–2.20, p=0.03).

Sensitivity analysis of composite HIV education content and HIV risk behaviors

Six states included 48,965 respondents representing 2,544,454 high school students. Compared to states that that had non-comprehensive HIV education policies, states with comprehensive HIV education policies had 21% lower odds of at least one HIV risk behavior (AOR 0.79, 95% CI 0.64–0.98, p=0.03), 36% lower odds of four or more lifetime sexual partners (AOR 0.64, 95% CI 0.45–0.91, p=0.01), and 20% lower odds of condomless last sex (AOR 0.80, 95% CI 0.66–0.96, p=0.02). There was no association with substance use before last sex.

DISCUSSION

Our study sought to establish the association between state-level high school HIV education policies and adolescent HIV risk behaviors. We found that despite the majority of students being from states that mandated HIV education, simply having a mandate for HIV education was not associated with HIV risk behavior, as we hypothesized. Rather, we found that the content of the HIV education was associated with HIV risk behaviors. In particular, sexual orientation content that is discriminatory differentially affected sexual minority students, and, as has been previously found, abstinence only education meant HIV risk behavior was more likely. In aggregate, HIV risk behaviors were less likely among students from states with more comprehensive HIV education.

There was considerable variation in the HIV education content regarding condom education, abstinence and other safe sex options, and the tone of sexual orientation content. Previous studies have found that HIV and condom education either reduce or have neutral relationships with sexual risk.^{5–7,11–15} Similarly, we found that students in states mandating HIV and condom education were neither more nor less likely to report engaging in any of the studied HIV risk behaviors. Our findings, consistent with previous evidence, do not support the concern that discussing HIV or condom use may increase risky sexual behaviors among adolescents.

Additionally, our study found notable cross-sectional associations between abstinence education policies and the studied HIV risk behaviors. In line with our hypothesis and similar to previous evidence,⁸ our study found that stressing abstinence in high school HIV education was associated with increased risk of engaging in at least one of the studied HIV risk behaviors – specifically, having four or more lifetime sexual partners and condomless last sex. However, if abstinence is not stressed and simply covered as one of several safe sex options, we saw a decrease in the odds of reporting at least one HIV risk behavior and specifically a significant reduction in the odds of condomless last sex. Our findings are complementary to previous studies looking at teen pregnancy, which found that covering the benefits of abstinence along with condom use and contraception was more effective than abstinence-only curricula at delaying initiation of sexual activity and reducing teen pregnancy risk.^12,13 To that end, we also found that students exposed to comprehensive education had significantly lower odds of HIV risk behaviors compared to students from states that did not. This supports the recommendation for more comprehensive HIV education as opposed to education that stresses abstinence or teaches only abstinence in an effort to decrease HIV risk behaviors among high school students.

Also consistent with our hypothesis, we found that compared to policies requiring an inclusive approach to sexual orientation content, students from states that had no policies or that required a discriminatory approach to sexual orientation were more likely to demonstrate at least one HIV risk behavior—specifically condomless last sex. Moreover, our findings of significantly higher odds of all HIV risk behaviors among students who identified as bisexual are especially important considering the limited available evidence focusing on youth who self-identify as bisexual; recent evidence has demonstrated results consistent with our findings with youth who have sex with both males and females having higher odds of both substance use before last sex and condomless last sex.⁷ These findings highlight the need to focus on addressing the disproportionate HIV incidence and prevalence among SMY.^2,16 It is known that SMY, particularly young men who have sex with men, have significantly higher rates of HIV risk behaviors and HIV incidence.^2,16 However, previous studies have also shown that SMY are less likely to report receiving school-based HIV education, indicating a missed opportunity for HIV education to meet the specific needs of SMY.^6,7,11 To that end, our findings support the benefits that can result from having HIV education with inclusive sexual orientation content and that reaches and resonates with SMY. Failure to provide inclusive content may lead to further exacerbation of the already existing disparities in HIV risk.² Furthermore, given that our findings also demonstrated increased condomless last sex among students who identified as heterosexual (84.5% of respondents), it is essential to avoid discriminatory content in order to reduce youth HIV risk behavior and HIV incidence among both sexual majority and SMY.

Our stratified analysis also demonstrated a protective association between discriminatory sexual orientation content and substance use before last sex among students who identified as gay or lesbian. This finding should be interpreted with caution due to the small sample size of students who identified as gay or lesbian from states with discriminatory sexual orientation content, making up less than 0.3% of the study sample size.

Limitations

Though we tried to ensure the validity of our study to the best of our ability, there are several limitations that warrant consideration. Due to the cross-sectional nature of our study, causal inferences cannot be made. It also remains possible that our associations were affected by unmeasured variables. The mechanisms between state-level policies and their effects on HIV risk behaviors cannot be determined by this study. Additionally, our generalizability is limited to the states included in each analysis by the necessity to exclude states that did not collect data on the variables included in the models. The sexual identity variable used in this study does not fully capture the details of sexual attraction or sexual behavior. However, sexual identity was chosen to be utilized in this analysis to yield a higher sample size and allow for inclusion of more states nationwide that included the YRBS sexual identity question but did not ask about sex of sexual partners. Although correlated, differences in operationalization of sexual and gender identity, sexual attraction, and sexual behavior have effects on HIV risk behaviors not captured in this study. Future studies should be done with larger sample sizes of different levels of SMY to examine more robust within-group associations also taking into account gender identities.

Conclusions

This study supports the need for state-level policies that mandate comprehensive and inclusive HIV education content because of the significant benefit they confer on adolescent HIV risk behaviors. It is essential that this content is taught in a way that resonates with high school students and supports the development of healthy, safe practices regardless of sexual identity.

Table 3.

Odds of HIV risk behaviors by tone of sexual orientation content stratified by student-reported sexual orientation

	HIV risk behavior			≥4 lifetime sexual partners			Substance use before last sex			No condom use at last sex
	AOR	95% CI	P	AOR	95% CI	P	AOR	95% CI	P	AOR	95% CI	P
Heterosexual
Inclusive	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
No policy	1.16	(0.95, 1.43)	0.14	1.07	(0.79, 1.44)	0.67	1.12	(0.79, 1.59)	0.54	1.24	(1.06, 1.45)	0.01
Discriminatory	1.17	(0.93, 1.48)	0.17	1.23	(0.92, 1.66)	0.17	1.18	(0.82, 1.68)	0.37	1.23	(1.01, 1.50)	0.04
Gay or lesbian
Inclusive	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
No policy	1.25	(0.71, 2.20)	0.43	1.24	(0.51, 3.02)	0.64	0.66	(0.33, 1.33)	0.25	1.33	(0.74, 2.40)	0.34
Discriminatory	1.05	(0.56, 1.94)	0.89	1.35	(0.50, 3.68)	0.55	0.30	(0.15, 0.61)	<0.01	1.27	(0.60, 2.67)	0.53
Bisexual
Inclusive	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
No policy	1.46	(1.08, 1.96))	0.01	1.61	(1.00, 2.59)	0.05	1.61	(1.04, 2.48)	0.03	1.58	(1.20, 2.06)	<0.01
Discriminatory	1.39	(1.00, 1.94)	0.05	1.97	(1.22, 3.18)	0.01	1.76	(1.10, 2.82)	0.02	1.51	(1.04, 2.20)	0.03
Not sure
Inclusive	Ref	--	--	Ref	--	--	Ref	--	--	Ref	--	--
No policy	1.50	(0.86, 2.62)	0.16	2.07	(0.92, 4.67)	0.08	1.72	(0.74, 3.99)	0.20	1.61	(0.82, 3.14)	0.16
Discriminatory	1.57	(0.88, 2.78)	0.13	2.02	(0.94, 4.36)	0.07	1.50	(0.63, 3.58)	0.36	1.60	(0.84, 3.05)	0.16

Note: AOR = adjusted odds ratio; CI = confidence interval

^aAdjusted for sex, race/ethnicity, grade, sexual identity

IMPLICATIONS FOR SCHOOL HEALTH.

The results of this study have implications for educators and policy makers to advocate and take action to support the provision of HIV education in schools that is supported by empirical evidence to decrease HIV risk behaviors in adolescents. Our findings demonstrate that policies simply mandating HIV education to be taught is not enough. Within HIV education, content should be comprehensive in providing instruction about safe sex practices in addition to abstinence. Additionally, HIV education materials must be inclusive of SMY and not be designed or delivered with the assumption that students are heterosexual and cisgender. Educational materials should be reviewed and redesigned to be taught in a way that resonates with youth regardless of sexual orientation, or gender identity. In addition to revisiting the content of HIV curricula, schools can also explore ways to integrate technology into HIV education to increase youth engagement through digital platforms, gamification, and social media. A comprehensive, accessible, and acceptable approach to HIV education is necessary in order to empower adolescents to make safe, informed, and positive decisions about their sexual practices and continue the efforts to decrease HIV incidence among youth at risk.