Gathering insights about sexual health promotion from adolescents’ adoption of COVID-19-protective behaviors

Abstract

This study aimed to understand how adolescent adoption of behaviors protective against different health risks relate to each other, as well as the role family and friends might play in the adoption of protective behaviors. We used data from adolescents (N=1803; age 13-17 years) who were recruited between April 2022 and June 2023 for QuaranTeen Health, an online, longitudinal survey assessing the impact of the COVID-19 pandemic on adolescent sexual health. We used data from the baseline questionnaire to calculate weighted correlations between friend and family norms about masking and vaccination and adolescents’ COVID-19 protective behaviors, defined as 1) participant masking in the past two weeks and 2) receiving at least one COVID-19 vaccination shot. We used weighted logistic regression models to estimate associations between COVID-19-protective behaviors and sexual health-protective behaviors (i.e., condom use, and hormonal contraception use at most recent sex). We found that increased friend and family norms for COVID-19-protective behaviors were associated with participant masking (friends: ρ=.40, p<.01; family: ρ =.34, p<.01) and vaccine uptake (friends: ρ=.40, p<.01; family: ρ=.56, p<.01). Both behaviors were positively associated with condom use (masking: OR=1.63, 95% CI: 1.05-2.52; vaccination: OR=2.20, 95% CI: 1.32-3.69). Neither were associated with hormonal contraception use. Our findings suggest that family and friend norms for health protective behaviors appear to be associated with health protective behaviors among adolescents, at least when examining COVID-19-protective behaviors. Integrating friend and family norms may increase the effectiveness of future sexual health interventions designed to increase sexual health protective behaviors among adolescents.

Introduction

Adolescents in the United States (U.S.) are substantially burdened by sexually transmitted infections (STIs) and unintended pregnancy (Leichliter et al., 2021; Rossen et al., 2023). In fact, STI rates have increased in the U.S. since 2000, according to a 2021 report from the National Academies of Sciences, Engineering, and Medicine, disproportionately affecting adolescents and marginalized groups (Vermund et al., 2021). This ongoing crisis necessitates renewed efforts to advance sexual health promotion, with particular emphasis on moving beyond individual-level behavior and into a socioecological framework (Boyer et al., 2021; National Academies of Sciences et al., 2021; Rietmeijer et al., 2022).

Insights from the COVID-19 pandemic have potential to inform health promotion efforts in other areas of public health, including adolescent sexual health. In 2020, COVID-19 upended life for people across the globe. Early public health measures recommended substantial changes to daily activities, including maintaining physical distance from other people and wearing masks in shared spaces, to limit community spread (Cowling & Aiello, 2020). Enacting these behavioral changes required frequent and dynamic messaging from a variety of sources, including federal and local governments, across diverse media platforms (Anwar et al., 2020; Cooks et al., 2022). Additionally, community norms, especially the perception of friends’ and community-members’ beliefs about health behaviors, played a crucial role in the individual adoption of COVID-19-protective behaviors in adults (Hensel et al., 2022; Latkin et al., 2022). This widespread and rapid rollout of behaviors that were previously not normative in most situations (e.g., masking and physical distancing) offered a poignant opportunity to study the adoption of health-promoting behaviors.

Sexual protective behaviors have many similarities to COVID-19-protective behaviors, including the use of physical barriers (masks and condoms) and routine testing. Elucidating factors related to the successful adoption of COVID-19-protective behaviors therefore has strong potential to apply to sexual health promotion. Existing research suggests that social norms, including family and peer behaviors and values, are particularly salient to adolescent health behavior (Kotchick et al., 2001; Viner et al., 2012), though limited research has specifically evaluated the role of social norms in COVID-19-protective behaviors in adolescents. A handful of studies have demonstrated a positive association between perceived community norms related to COVID-19 vaccination and adolescent vaccine intentions (Euser et al., 2022; T. Li et al., 2023; Rogers et al., 2021) and one study reported a positive association between higher perceived religious norms supporting COVID-19 vaccination and an increased likelihood of adolescent vaccination (Zilhadia et al., 2022). Though these preliminary studies suggest important connections between social norms and COVID-19-protective behaviors, more research is needed to elucidate a complete picture of this relationship (Agnew et al., 2025).

A thorough understanding of the role of social norms in the adoption of COVID-19-protective behaviors has broad implications for health promotion and could inform future interventions geared towards increasing the adoption of other health protective behaviors, including sexual-protective behaviors. As such, we sought to assess how family and friend norms relate to the adoption of COVID-19-protective behaviors, and how COVID-19-protective behaviors relate to sexual-protective behaviors, among adolescents in the U.S.

Materials & Methods

We used baseline data from the QuaranTeen Health study, an online, longitudinal survey assessing the impact of the COVID-19 pandemic on the sexual behaviors and romantic relationships of U.S. adolescents. The study protocol was reviewed and approved by the Pearl Institutional Review Board. A waiver of guardian permission was granted.

Participants were recruited using advertisements on Instagram and Facebook and completed the baseline survey between April 2022 and June 2023. We used both photo and video advertisements designed to appeal to diverse groups of youth. Advertisements were created in both English and Spanish and invited youth to “make a difference” and “help teens like you.” Eligible participants lived in the U.S., were able to complete a survey in English or Spanish, were 13-17 years of age, had not completed high school, and demonstrated the capacity to assent by completing an assessment in which they answered a series of questions confirming their understanding of study procedures, risks, and benefits. Eligible participants who provided informed assent moved on to the baseline survey. We designated diversity targets to ensure that different groups of youth were well-represented in the sample: 50% racial/ethnic minority youths, 20% youth living in rural areas, 40% sexual and gender minority youth, and even distributions of sex assigned at birth. When a target for a particular group was met, additional participants from that group were ineligible for the survey. We were not able to meet all targets, due to the increasing costs of social media advertisements and restrictions on targeted advertisements. Figure 1 illustrates the flow of participants from clicking on a study advertisement to the final baseline sample.

Figure 1: Participant flow for the QuaranTeen Health Study.

Note: Percentages are calculated based upon the sample size in the box above. *Reasons for ineligibility are not mutually exclusive.

On average, adolescents completed the survey in 38 minutes (standard deviation (SD) = 10 minutes). Survey questions covered a broad range of issues including COVID-19 related experiences, sexual relationships and experience, and health behaviors. As this was a longitudinal cohort study, multiple forms of participant contact information (e.g., email, phone number, social media handles) were collected at the end of the baseline to facilitate follow-up and fraud checks. Participants were offered a choice of incentives for completing the baseline survey: a $15 Amazon gift card, $15 donation to a pre-specified list of non-profits, or to forgo the compensation altogether.

The research team followed set protocols to detect fraudulent responses. Participants who input a date of birth during the survey process that did not match the age that they provided in the screening survey were flagged and unable to complete the survey until the discrepancy was reconciled. Participants who could not confirm their age and date of birth were removed from the study. Additionally, study staff cross-referenced the participants’ IP addresses on a third-party IP fraud check website (Scamalytics Ltd, 2024). If an IP address had a fraud score ≥60%, the participant was contacted by telephone to confirm their date of birth and email address. Participants that could not confirm these details were removed from the study.

Measures

Sociodemographic variables.

We captured sociodemographic variables via self-report on the baseline. Participants provided their birthdate, which we used to derive age at baseline, and reported their current grade in school. We assessed sex assigned at birth using the question “What sex were you assigned at birth on your original birth certificate?” Participants could select male, female, or intersex. Gender identity was assessed separately using the question “Which of the following best describes your gender identity?” Participants could select all that apply from a list of options (which included cisgender girl, cisgender boy, transgender boy, transgender girl, non-binary, etc.), and could also use an open-ended text box to describe additional identities that were not listed. To determine whether participants identified as having Hispanic identity we used the question: “Are you of Spanish or Latino origin, such as Latin American, Mexican, Puerto Rican, or Cuban?” which was separate from a question asking “What is your race?” For racial identity, participants could select all that applied from a list of options (which included Black or African American, White or Caucasian, Asian, etc.), with an additional option of writing in a racial identity in an open-ended text box. We also asked for participants’ residential ZIP code to determine state of residence.

COVID-19 Protective Behavioral Norms.

To assess perceived friend norms, adolescents responded to a series of statements starting with: “How many of the friends your age who you respect…”: 1) Think you should get vaccinated; 2) Think you do not need the booster to stay healthy; 3) Think you should wear a mask when you are inside public places, like the grocery store and restaurants; and 4) Think you can go to concerts and large sports events without worrying about getting COVID-19. For each of the four items participants responded on a 5-point Likert scale from ‘None of my friends’ (assigned a score of 1) to ‘All of my friends’ (assigned a score of 5). Items 2 and 4 were reverse scored so that higher scores corresponded to greater endorsements of protective behaviors. This 4-question battery was repeated for family norms starting with the prompt: “How many of the people in your family who you trust…” Survey items were created for the study based upon similar scales for perceived peer and caregiver norms for other preventive behaviors (Misovich et al., 1998). Given the significant correlations between items, we generated separate sum scores for friend and family norms, ranging from 4-20. A higher score indicated COVID-19-protective behaviors were more normative. Alphas indicated acceptable reliability (α=0.82 for friends and α=0.87 for family).

COVID-19 Protective Behaviors.

Per the Centers for Disease Control and Prevention’s (CDC) recommendations for mask wearing at the time of baseline survey development and launch (Centers for Disease Control and Prevention, 2021), participants were asked to report how often they “wore a mask or face covering, keeping it over your nose and mouth” in three different settings: 1) “In an indoor public place (this includes school, stores restaurants and other public places),” 2) “Doing something outside when you are in close contact with other people,” and 3) “On public transportation, including a bus, train, or plane,” during the prior 2 weeks. The 5-point Likert response scale ranged from none (1) to all the time (5). Responses were collapsed to create a binary indicator of wearing a mask at least some of the time (i.e., a response of 4 or 5) in any setting. Participants were also asked how many, if any, vaccination shots for COVID-19 they had received. Responses were collapsed to create a binary variable indicating receiving any COVID-19 vaccinations versus none.

Sexual Protective Behaviors.

All participants were asked at what age they first had sex involving a penis going into a vagina when they wanted to, as well as sex involving a penis going into an anus when they wanted to, accompanied by a check box to indicate “I have never had this type of sex.” Those who reported engaging in penile-vaginal sex were asked whether they or their partner used the following during the most recent time they had sex where a penis went into a vagina: female condom, male condom, birth control pills, shot (e.g., Depo-Provera), patch (e.g., Xulane), ring (e.g., NuvaRing), intrauterine device (IUD, e.g, Skyla, Mirena or Paragard), or implants (e.g., e.g., Implanon or Nexplanon). [Though frequently used, the terms “female condom” and “male condom” are not inclusive. Language such as “internal” or “condoms that can be used inside a vagina or anus,” and “external” or a “condom that covers the penis” can differentiate between the two without gendered language.]

Those who reported engaging in penile-anal sex were asked “Thinking about the most recent time you had sex where a penis went into an anus (or butt), did you or your partner use a condom?”. Any condom use, either internal or external, during penile-vaginal or penile-anal sex was collapsed into a single category indicating condom-protected sex for analysis. Use of hormonal contraceptives during penile-vaginal sex was collapsed to indicate any type of hormonal contraceptive use.

Analysis

To approximate the population of U.S. adolescents, we developed survey weights. Weighting was done using Sample Balancing, an iterative sample weighting program that simultaneously balances the distributions of all variables using a statistical technique called the Deming Algorithm (Deming & Stephan, 1940). To do so, we used US Census Bureau data to weight the sample on demographic characteristics, and the Youth Risk Behavioral Surveillance Survey to weight the sample on youth behaviors so that the data approximate the demographic and behavioral characteristics of US youth (CDC, 2025; United States Census Bureau, 2023). Specifically, we derived targets for age, gender, race, Hispanic ethnicity, number of children in the household, and number of adults in the household for 13-17 year olds from the United States Census Bureau Current Population Data for 2023 (United States Census Bureau, 2023). We also used data from the Youth Risk Behavior Surveillance System (YRBSS) (CDC, 2025) to identify weight targets for three health-related behaviors: physical activity (assessed using the question “During the past 7 days, on how many days were you physically active for a total of at least 60 minutes per day?”), texting and driving (assessed using the question “During the past 30 days, on how many days did you text or e-mail while driving a car or other vehicle?”), and alcohol consumption (assessed using the question “In the past 30 days, on how many days did you have one or more drinks of alcohol?”). We calculated weighted descriptive statistics on the aggregate sample including socio-demographic characteristics, sexual behavior, and COVID-19-protective behaviors. We reported the response distributions for individual COVID-related norm items separately for friends and family as well as their sum scores (participants were included in the calculation of family and friend sum scores if they provided responses to at least 3 of the 4 norm questions).

Our first analysis investigated the relationship between friend and family COVID-19-protective norms with participants’ COVID-19-protective behavior. We calculated separate survey-weighted correlation coefficients to estimate non-parametric relationships between the friend and family COVID-19 norms (sum scores) and individual-level COVID-19 protective behaviors: dichotomous indicators of masking and vaccine uptake. Next, we assessed the relation between individual-level COVID-19-protective behaviors and sexual protective behaviors. We used separate survey-weighted logistic regression models to generate unadjusted odds ratios (ORs) and corresponding 95% confidence intervals (CIs) quantifying the association between individual-level COVID-19-protective behaviors and sexual health protective behaviors: condom use and contraception use at most recent sex. As the COVID-19 pandemic and related restrictions shifted during data collection, we conducted a sensitivity analysis to explore whether effects were conditional on time from the onset of the pandemic (March 13, 2020; (CDC, 2024)) to completion of the baseline survey and found that the effects were not changed. Analyses were conducted in R Studio 3.6.0 with significance level set at .05 a priori. We used a likelihood-based approach to estimation, thus making use of all available data without directly imputing missing values.

Results

Table 1 displays weighted sample characteristics of the 1,803 participants. The mean age was 15.3 years (SD=1.34, Range: 13-17 years). One quarter (25%) reported being Hispanic ethnicity, with the most frequently endorsed racial identities being White (63%), Black/African American (13%), and Multiracial (6%). Reflective of the study design, one-quarter (25%) identified as sexual minority and 7% as gender minority. Fifty-four percent of participants reported any masking in the past 2 weeks and 78% reported receiving at least 1 COVID-19 vaccine shot. Participants resided in all 50 U.S. states, the District of Columbia, and Puerto Rico. Due to small cell sizes by state, we are not able to assess meaningful differences by state.

Table 1.

Characteristics of the QuaranTeen Health weighted baseline sample (n=1,803)

	N(%) or mean (SD)
Age, years	15.28(1.34)
Sex
Female	1215(46.94%)
Male	571(52.27%)
Intersex	17(0.79%)
Grade
7th Grade	37(5.34%)
8th Grade	133(13.24%)
9th Grade	304(19.10%)
10th Grade	430(22.18%)
11th Grade	567(23.91%)
12th Grade	310(15.69%)
Currently enrolled in a GED Program	10(.34%)
Dropped Out	12(.20%)
Race
White	1179(63.08%)
Black or African-American	123(13.20%)
Asian	134(7.71%)
Native Hawaiian or other Pacific Islander	5(.21%)
Native American, American Indian or Alaskan Native	61(3.25%)
More than one race	136(5.89%)
Other	120(5.30%)
Do not want to answer	45(1.35%)
Ethnicity
Hispanic ethnicity	410(24.86%)
Not Hispanic ethnicity	1371(74.57%)
Do not want to answer	22(.57%)
Gender Identity
Cisgender girl	885(44.47%)
Cisgender boy	442(49.53%)
Transgender girl	40(.64%)
Transgender boy	91(.86%)
Indigenous/Cultural Gender minority	6(.25%)
Non-binary	146(2.39%)
Genderqueer	43(.68%)
Pangender	7(.06%)
Genderfluid	102(1.55%)
Agender	33(.57%)
Questioning	121(3.65%)
Other gender identity	35(1.42%)
Do not understand the question	62(4.71%)
Do not want to answer	18(1.27%)
Sexual Identity
Heterosexual	875(75.17%)
Gay	113(3.73%)
Lesbian	98(1.99%)
Bisexual	385(11.51%)
Pansexual	205(4.72%)
Demisexual	60(1.24%)
Queer	123(2.91%)
Asexual	119(6.32%)
Questioning/Unsure	198(8.86%)
Another Sexual Identity	66(2.07%)
Do not understand the question	262 (16%)
Do not want to answer	15(1.26%)
Sexual Behaviors
Any vaginal or anal sex	339(13.94%)
Any condom use at last vaginal or anal sex	195(57.90%)
Hormonal contraceptive use at last vaginal sex	98(40.21%)
COVID-19 Protective Behaviors
Any masking	975(53.80%)
≥ one COVID-19 vaccine shot	1398 (77.53%)
COVID-19 Protective Norms Scores (range 4-20)
Friends	12.62(4.34)
Family	12.63(4.09)

Abbreviation: SD, standard deviation. Note: Counts are observed counts and proportions are weighted.

Item-level COVID-19 protective norms are depicted in Figure 2 (separately for friends and family norms). Response distributions skewed slightly in favor of COVID-19-protective behaviors. Although sum scores were not different (family: mean 12.62, SD=4.09; friends: mean 12.63, SD=4.34), in general, item-specific scores for perceived family norms were slightly more favorable towards COVID-19-protective behaviors as compared with friends (see Figure 2). We found that higher norms around COVID-19-protective behavior among friends and family were each significantly associated with participants’ past 2-week masking behavior (friends: ρ=.40, p<.01; family: ρ =.34, p<.01) as well as reporting at least 1 COVID-19 vaccination shot (friends: ρ=.40, p<.01; family: ρ=.56, p<.01).

Figure 2. Friend and Family COVID-19-protective norms: Summary of Weighted Proportions at the Item Level.

*Items were reverse coded for the sum scores.

Fourteen percent of the sample reported ever having had penile-vaginal or penile-anal sex. Of those, 61% of participants reported condom use at last penile-vaginal or penile-anal sex and 46% reported hormonal contraceptive use at last penile-vaginal sex. Figure 3 displays observed associations between COVID-19-protective behaviors and sexual protective behaviors. Participants who reported ever having at least one COVID-19 vaccination shot had over 2 times the odds of using a condom at most recent sex (OR=2.20, 95% CI: 1.32-3.69). Participants who reported using a mask within the prior two weeks also had an increased likelihood of using a condom at most recent sex (OR=1.63, 95% CI: 1.05-2.52). There was no association between vaccine uptake or masking with contraception use (masking: OR=1.1, 95% CI: 0.6, 1.9; vaccine uptake: OR=0.7, 95% CI: 0.4, 1.2).

Figure 3. Relation between COVID-19-protective behaviors and sexual protective behavior among those who reported vaginal or anal sex (N=339).

Note: Error bars represent 95% confidence intervals

Discussion

The influence of social norms is well-documented across an array of adolescent health behaviors (Pedersen et al., 2015; Rice & Klein, 2019; Spijkerman et al., 2007). However, the COVID-19 pandemic represented an unprecedented, fast-moving, and urgent public health situation that can inform public health practitioners about how social norms relate to the uptake of health-protective behaviors among adolescents. Based upon data from a diverse sample of over 1,800 U.S. adolescents, we observed a positive association between perceived family and friend norms supporting COVID-19-protective behaviors, and participants’ own COVID-19-protective behavior. Specifically, as the proportion of trusted family members or friends perceived to support COVID-19-protective behaviors increased, so too did the likelihood of a participant reporting engaging in COVID-19-protective behaviors. These findings are in line with previous research. As noted earlier, previous studies among adolescents have reported positive associations between higher levels of perceived support for COVID-19 vaccination from friends and family (Euser et al., 2022; T. Li et al., 2023; Rogers et al., 2021) and religious communities (Zilhadia et al., 2022) with vaccine intentions; and in one case, vaccine uptake (Zilhadia et al., 2022). Our study builds upon this emerging literature by documenting a positive correlation with masking behavior, suggesting that social norms influence COVID-19-protective behaviors beyond vaccination.

We also found that COVID-19-protective behaviors were positively associated with condom use at most recent penile-vaginal or penile-anal sex. Interestingly, we did not observe similar associations between COVID-protective behaviors and contraceptive use. This could potentially be due to misclassification of contraceptive use when reported by the partner not using it. Many types of contraception that are used by females may not be obvious to a sexual partner (e.g., oral contraceptive pill or IUD). If partners did not discuss contraception, then one partner may not know whether contraceptives were used which could result in misclassification of the outcome. It is also possible that this reflects substantial differences in the availability and use cases for condoms as compared with hormonal contraceptives. Unlike condoms, hormonal contraceptives typically require engagement with the healthcare system and are often prescribed to adolescent females for non-contraceptive reasons, including the treatment of dysmenorrhea, irregular or heavy bleeding, endometriosis, or acne (Bitzer, 2013). The decision to use a condom, akin to wearing a mask, can be made more spontaneously (provided that condoms are available), as opposed to the decision to use hormonal contraception.

The association between COVID-19-protective behaviors and condom use may represent underlying health consciousness of certain individual adolescents or the clustering of health-promoting norms across a wide range of health behaviors in families and friend groups. Our findings support the influence that perceived norms play in adolescence, as well as the ongoing influence that perceived caregiver norms can have on adolescent decisions, including health-protective decisions. We believe that these findings, in the context of existing research, highlight the potential for interpersonal-level sexual health promotion for adolescents. Although interpersonal-level health interventions have shown promise at improving health outcomes, including sexual health outcomes (Hunter et al., 2019; Mason-Jones et al., 2023), they remain uncommon. For example, perceived peer norms have been shown to influence adolescent condom use across a wide range of populations (Diiorio et al., 2001; H. Li et al., 2017; McCarthy et al., 2022; Schaalma et al., 1993; Whitaker & Miller, 2000). Prior research also suggests that communication between parents and adolescents can influence teenage condom use, particularly if caregivers are perceived as open and comfortable with the topic (Leland & Barth, 1993; Whitaker & Miller, 2000), and that family closeness is positively associated with consistent condom use (Gillmore et al., 2011; Kao & Manczak, 2013). Perceived family support for condom use has been positively associated with condom use during first sex (Tarkang, 2014). Taken together, parent/caregiver communication and norm-setting may have immense potential as an intervention to improve adolescent sexual health, and peer-level interventions may have the opportunity to increase perceived perceptions of peer support for preventive behaviors. Future programs could explore opportunities to support parents in having productive and positive discussions with adolescents, as well as peers talking with each other about these important health topics.

Our study had many strengths. We recruited a large sample of adolescents from across the U.S. and used weighted analyses so that the data would behave as if it were generalizable to the U.S. adolescent population. We also asked about both penile-vaginal and penile-anal sex to get a more complete picture of sexual behaviors. In terms of COVID-19-protective behavior, we asked participants about masking behavior as well as vaccination, which extends prior research that has focused exclusively on adolescent vaccination (Euser et al., 2022; T. Li et al., 2023; Rogers et al., 2021; Zilhadia et al., 2022).

Our findings should also be interpreted in the context of study limitations. Specifically, the interpretation of our findings is limited by the cross-sectional nature of the data. We do not know the temporality of our observed associations, and it is possible that participants’ own beliefs about COVID-19-protective behaviors shaped their perception of their friend and family norms or influenced who in their networks they viewed as trustworthy. Further, we did not assess participants’ perceptions of friend and family norms towards sexual behaviors, which is likely to also be an important factor in their sexual behavior choices. Our vaccination question also did not differentiate between an initial vaccination shots and boosters, nor did it include when the participant received their shots. As such, we are unable to assess whether the number of shots a participant received was aligned with CDC recommendations at the time they received their shots. Vaccination rates in our sample may also have been influenced by other structural factors, including vaccination mandates at the municipal, state, and federal level. Future research assessing factors at other levels of the socioecology that may have impacted the COVID-19 protective behaviors of adolescents are warranted. While we took steps to ensure data quality, including fraud detection methods, it is still possible that our survey contained fraudulent entries that were not captured by our fraud detection protocol. Additionally, while over 18,000 individuals completed the screener survey and appeared eligible at screening, only 1,803 completed the survey and were ultimately deemed eligible. Although this is common for online surveys and we weighted the sample to approximate the adolescent population in the U.S., there may be differences between adolescents who complete an online survey and those who do not. Also, 14% of our sample reported ever engaging in penile-vaginal or penile-anal sex, compared with 32% of respondents in the CDC’s YRBSS who reported ever engaging in “sexual intercourse” (undefined) (Center for Disease Control and Prevention (CDC), n.d.). This may suggest that the current sample is less sexually active than the national population of adolescents; or that there was under-reporting of sexual behaviors; or that there is some underlying difference in the samples that is driving the difference that may reduce the generalizability of our findings. It is also possible that the undefined reporting of ever having “sexual intercourse” in the YRBSS captures a wider range of sexual behaviors, including behaviors beyond penile-vaginal and penile-anal sex. That said, our findings align with recent studies reporting declining trends in sexual activity in U.S. adolescents (Lindberg et al., 2021). There are also other sexual-protective behaviors that we did not assess, including use of dental dams and condom use during sex with sex toys, that are worthy of additional research as they may have their own relations with family and peer norms and COVID-19-protective behaviors that were not captured in this study. Finally, although we did not assess for sociodemographic differences in the current analyses, due to the deleterious effects of racism, sexism, heterosexism, and cisgenderism in the U.S. it is likely that there are sociodemographic differences in perceived family and friend health protective behavior norms and engagement in COVID-19- and sexual-protective behaviors. Future research assessing the potential influence of these harmful upstream factors may have on both health protective behavior norms and engagement in health protective behaviors could be useful.

Our findings suggest that perceived family and friend norms related to COVID-19-protective behavior are associated with individual vaccine uptake and mask-wearing in a diverse sample of U.S. adolescents. We also report a positive association between COVID-19-protective behaviors and sexual-protective behaviors. Taken together, these results suggest that family and friend norms are likely influential to the uptake of health-promoting behaviors, including sexual-protective behaviors. Given the initial success of public health campaigns at rapidly shifting COVID-19-protective behaviors as well as family and peer norms around said behaviors, it is possible that a similar approach could be used for the promotion of sexual-protective behaviors. Future research assessing the impact of frequent and dynamic public health messaging about sexual-protective behaviors on perceived sexual-protective behavior norms and adoption of sexual-protective behaviors is warranted. Further, given that findings from this study suggest that interpersonal-level health promotion efforts may be an effective mechanism by which to positively influence adolescent health behaviors, including sexual behaviors, it will be important to identify other modifiable interpersonal-level factors (e.g., interpersonal communication with family, friends, and sexual partners) that may be related to uptake of health protective behaviors. Overall, this research suggests that moving sexual health interventions from the individual-level to the interpersonal-level may be a fruitful path to increasing the adoption of sexual-protective behaviors among adolescents and, ultimately, decreasing the persistent sexual health concerns they experience.