Peer Victimization and Risky Alcohol Use among Sexual Minority Youth: Understanding Mechanisms and Contexts

State University Of New York At Buffalo, Buffalo NY

Investigators

Abstract

Administrative supplemental funds are being sought to maintain staffing support needed to extend recruitment efforts and secure sufficient sample size and power to achieve the stated aims of the parent grant (R01AA028810). Due to unforeseeable challenges, recruitment and data collection goals have been significantly delayed. The parent study seeks to address the etiology of high-risk substance use among non-heterosexual adolescents. Early, risky alcohol use may result in later alcohol problems and related deleterious health consequences (e.g., other substance abuse, victimization, poor physical and mental health) that contribute to health disparities among non-heterosexual adults. The harmful effects of peer victimization (PV; i.e., bullying and sexual harassment) on adolescent psychosocial functioning may be one pathway through which non-heterosexual youth become involved in high-risk alcohol use. In adolescence, social norms are strongly enforced through bullying and sexual harassment. Alarmingly high numbers of non-heterosexual youth experience peer aggression that can lead to stress, which has been strongly associated with increased alcohol use across the lifespan in non-heterosexual populations. Cross-sectional studies have shown that PV is positively associated with alcohol and other substance use among non-heterosexual adolescents. However, the mechanisms through which PV contributes to negative outcomes and the protective factors that ameliorate those outcomes among these youth are not well understood and have been hampered by a dearth of longitudinal and mixed methods studies that include these youth in their samples. Using a mixed methods design, the parent study investigates the acute daily and longitudinal effects of PV on non-heterosexual adolescent risky alcohol use and identifies potential buffers and risk factors for these outcomes. Data collection is underway (target N=400; ages 14-18 years) using four longitudinal surveys (6 months apart), two bursts of daily reports (4-weeks each), and qualitative interviews. The parent study aims to: (1) identify the psychosocial mechanisms linking PV with alcohol and other substance use among non-heterosexual youth; (2) understand the daily associations between PV and alcohol and substance use; and (3) examine the contexts in which non-heterosexual youth experience PV and gain insight into the psychosocial factors related to PV and substance use. We utilize a novel integrative conceptual model that incorporates extant knowledge from research on PV and substance use conducted with heterosexual youth with factors identified in the non-heterosexual literature (i.e., stress, psychological mediation framework) to clarify the processes implicated in substance use by non-heterosexual youth. Results from this study are critically needed to understand the pathways, processes and contexts through which non-heterosexual adolescents become involved in high-risk alcohol use to inform the development of primary prevention programs that improve and preserve the health of these youth. Modified Specific Aims/Strategy Section INTRODUCTION TO CURRENT ADMINISTRATIVE SUPPLEMENT REQUEST In this application, we seek administrative supplemental funds to support ongoing recruitment to achieve the stated aims from our funded parent study (R01AA028810). We seek these funds to support ongoing staffing that is needed to extend recruitment efforts that were hampered due to several unforeseeable challenges that significantly delayed our recruitment and data collection goals by over a year. Our initial plan was to recruit 500 youth who identify as non-heterosexual. We are currently in Year 4 of the grant and have 225 participants enrolled. After several setbacks, we have successfully enrolled participants by modifying our approach in a few ways to account for the delays. We have pivoted from our initial plan to recruit via social media and outreach to community centers to conduct in-person and virtual presentations to school-based clubs in high schools across New York State. This method has been successful but is labor intensive and has resulted in a longer recruitment timeline than we originally estimated. We also have recalculated our power analysis to evaluate the possibility of achieving our aims with a smaller sample. Our original target sample was N = 500 and the revised power analysis indicates that we will have sufficient power to achieve our aims with a smaller sample of N = 400 (see power analysis below and revised enrollment table). We also have expanded the age of eligibility to include 14 to 18-year-old participants to further bolster our recruitment efforts (the original application proposed eligibility at 15-17 years). We are confident that we can achieve our goal of recruiting 400 participants, but we lack sufficient resources to do so given the challenges that we will explain in detail later in the application. Thus, we are seeking administrative funds to extend staff salaries to achieve those goals. An extension of staffing will allow us to continue to recruit participants via high-school clubss and complete longitudinal, mixed methods data collection from those participants to account for attrition and ensure adequate power to successfully achieve the study aims. In the remainder of the application, we will outline the aims and approach of the parent grant, including the power analysis for the new target sample size, and detail the nature of the challenges and resulting delays underlying this request for administrative supplemental funds. PARENT STUDY SPECIFIC AIMS Rates of risky alcohol use and related adverse outcomes are alarmingly high among non-heterosexual youth1-4 and these disparities persist across the lifespan.5-7 non-heterosexual youth initiate alcohol use at earlier ages and use more heavily relative to their heterosexual peers.8, 9 This raises concerns that—like patterns found in the general population10—risky alcohol use by non-heterosexual youth may strongly predict later alcohol problems and related deleterious health consequences (e.g., other substance abuse, victimization, poor physical and mental health8, 11). Given that 3 million non-heterosexual adolescents reside in the U.S.,12 understanding the mechanisms associated with risky alcohol use in this population aligns with NIH NOT-AA-20-017 as a public health priority. Non-heterosexual people are subject to unique and pervasive stressors due to the perceived violation of social norms surrounding heterosexuality.13-15 Stress has been strongly associated with increased alcohol use across the lifespan in this population.2, 15, 16 In adolescence, social norms promoting heterosexuality are strongly enforced through bullying and sexual harassment. Alarmingly high numbers of non-heterosexual youth report experiences of peer victimization (PV).17-19 A national study of non-heterosexual students found that 82% reported verbal harassment, 56% were physically harassed or assaulted, and 55% were cyber-bullied in the past year due to their sexual orientation.20 The harmful effects of PV on adolescent psychosocial functioning may be one pathway through which non-heterosexual youth become involved in high-risk alcohol use.21, 22 Cross-sectional studies have shown that PV is positively associated with alcohol and other substance use among non-heterosexual adolescents.21, 23, 24 However, the mechanisms through which PV contributes to negative outcomes, as well as protective factors that ameliorate those outcomes among non-heterosexual youth, are not well understood. Further, most of what we know about the role of PV in alcohol and other drug use has come from samples of predominantly heterosexual youth. The timing of PV, especially sexual harassment, may be even more critical for non-heterosexual youth as they grapple with disclosure of their status. Non-heterosexual youth also experience unique stressors that may further exacerbate the harmful impact of PV. Advances in understanding the effects of PV on non-heterosexual adolescents have been hampered by a dearth of longitudinal studies focused on these youth, as well as a lack of qualitative research to identify unique aspects of their experiences.25 Mixed methods studies are vitally needed to understand the pathways, processes and contexts through which non-heterosexual adolescents become involved in early and high risk alcohol use and to inform the development of primary prevention programs that preserve the health of non-heterosexual youth.8 In our funded parent study (R01AA028810), we are using a mixed methods approach to investigate the acute daily and longitudinal effects of PV on non-heterosexual adolescent risky alcohol use, as well as identify potential buffers and risk factors for these outcomes. Data collection from 400 adolescents (ages 14-18 years, 50% female, diverse racial composition) is actively underway using a combination of four longitudinal surveys (baseline, 6-month, 12-month, 18-month follow-ups), two bursts of daily reports (4-weeks each), and qualitative interviews. The parent study is informed by an integrative conceptual model that merges stress13 and psychological mediation models26 (see Figure). Our model is novel in that it incorporates extant knowledge about PV and alcohol use among heterosexual youth, along with stressors unique to non-heterosexual youth. The aims include the following: Aim 1: Identify psychosocial mechanisms linking PV with alcohol and other substance use among non-heterosexual youth. Longitudinal survey data are being collected to examine prospective, cumulative effects of PV on psychosocial functioning and substance use over time. Survey data collected at six-month intervals over an 18-month period will allow us to identify concurrent and prospective associations between PV and alcohol and other substance use via pathways that include risky and protective social relations, perceptions of social acceptance, and emotion regulation/coping. Moderating risk and protective factors will be considered. Aim 2: Understand daily associations of PV with alcohol and substance use. Daily diary data is being collected to identify the acute, proximal effects of PV on alcohol and other drug and nicotine use. We expect that reports of negative affect, emotional dysregulation, alcohol use, and other substance use will be higher on days that PV occurs compared with non-victimization days. Lagged effects into subsequent day(s) will be examined. Aim 3: Examine the contexts in which non-heterosexual youth experience PV and gain insight into the psychosocial factors related to PV and substance use. Qualitative interviews are being conducted to elucidate the circumstances surrounding PV and how non-heterosexual youth interpret their experiences. A randomly-selected, diverse sample of participants (n = 90) are asked to describe a recent PV experience (e.g., setting, abuse type) and post-PV consequences (i.e., barriers to obtaining support, disclosure, substance use, emotional distress). Psychosocial aspects of the experience, including the impact of the event on their social relations, perceptions of social acceptance, substance use beliefs and attitudes, emotion regulation, and coping will be assessed. Thematic differences between subgroups will be explored. Scientific Premise & Significance of Parent Study For non-heterosexual youth, PV may be part of a larger pattern of discrimination and victimization that compromises their physical and psychological well-being, ultimately contributing to alcohol and other substance use (AOSU). Given that the underlying mechanisms may operate differently for non-heterosexual youth due to stress and barriers to support seeking,27 an understanding of the developmental and reciprocal influences of these risk factors in relation to PV among non-heterosexual youth is needed. The scientific premise of the parent study is to advance understanding of the prospective and situational conditions under which PV contributes to AOSU, as well as identify potential buffers and risk factors for these adverse outcomes among non-heterosexual youth. This study will build upon extant etiological models of adolescent AOSU by incorporating unique risk and protective factors relevant to non-heterosexual adolescents. non-heterosexual adolescents, ages 14-18 years, will be targeted because this is a particularly vulnerable developmental stage when PV rates are high28 and AOSU and sexual exploration commences.29 Risky AOSU during adolescence can produce long-term, sometimes irreversible consequences that disrupt healthy transitions into adulthood.30 Most studies conducted with non-heterosexual youth are cross-sectional,25 and do not offer prospective and proximal daily insights,31 thereby limiting our ability to establish temporal ordering and mechanisms that explain the association between PV and AOSU. For this reason, it is critical that we have the staffing resources to recruit and retain our target sample to conduct longitudinal analysis. Four survey data points will permit the longitudinal examination of PV consequences to identify intervening factors that may elevate risk or buffer against AOSU outcomes. Daily reports will provide new and critical insights about the proximal factors linking PV and AOSU among this high-risk group. Given the unique vulnerabilities and stressors faced by these youth, it is possible that their victimization experiences differ in ways that are not yet well-understood and that may not be adequately assessed using survey methods. Qualitative approaches are highly effective for gaining insight into social contexts and understanding how these individuals make meaning of their experiences in ways that cannot be adequately captured through quantitative approaches.32 The collection of mixed methods data for the parent study positions our team to fill these critical gaps in the literature. INNOVATION OF PARENT GRANT In the parent study, we are utilizing a novel integrative conceptual model to understand the pathways through which PV impacts AOSU among non-heterosexual youth. Existing theoretical models pertaining to heterosexual youth fail to account for unique factors impacting non-heterosexual youth, while well-established models of non-heterosexual health theorize how PV contributes to alcohol use among non-heterosexual youth (e.g., psychological mediation model), but predominately utilize cross-sectional study designs that fail to account for complex developmental dynamics of these relationships. Furthermore, these models oftentimes fail to agree on the nature of key mechanisms associated with health outcomes. The first innovative feature of our study is our integrative model, which incorporates extant knowledge from research on PV and AOSU conducted with heterosexual youth along with factors identified in the non-heterosexual literature, including stress13 and the psychological mediation framework.26 Results will clarify the processes implicated in the AOSU of non-heterosexual youth and will provide prospective, empirical support for an integrative model of stress and psychological mediation models. The second innovative aspect of our study is our multi-method design, which is best suited to addressing critical gaps in the literature. We are collecting four waves of data over an 18-month period to consider how these relationships evolve over time, thus building on the prior work that has been largely cross-sectional in nature. Daily diary data will allow us to examine proximal processes related to PV and AOSU that likely vary from day to day. Finally, we will use qualitative interviews to gain insights into the descriptive contexts of specific PV events among this population and explore the psychosocial aspects of these experiences (e.g., impact of the event on their social relations, perceptions of social acceptance, emotion regulation, coping). While mixed methods are not novel in-and-of-themselves, no known studies have utilized our parent study’s combination of surveys, daily diaries and interviews to understand the nuances of the PV-alcohol connection among these youth. APPROACH OF PARENT GRANT Participants Eligible adolescents (N = 400, 50% female) are 14-18 years old, self-identify as non-heterosexual, enrolled in high school (public or private), and reside in NYS. To facilitate data collection/retention and compensation for participation, eligible youth must have a personal email address and be able to receive personal text messages on a mobile device. Youth are asked which of the following best describes how they identify themselves: strictly/exclusively heterosexual, mostly heterosexual, bisexual, mostly gay or lesbian, gay or lesbian, queer, pansexual, or other (open-ended). Exclusively heterosexual youth are ineligible for participation. Recruitment, Enrollment & Retention Efforts Recruitment efforts include distribution of referral incentives and the use of recruitment videos, flyers, and virtual presentations. The PIs have sought advisement from a community provider partner agency (GLYS of Western New York) to help achieve our recruitment goals. Our community partners and consultant Newcomb have provided advice on recruitment strategies (including the filming of a video and development of a website) and study design (e.g., delivery, language, structure), as well as how to facilitate networking opportunities to support recruitment across NYS. Recruitment materials contain a brief study description and a link to a study ‘marketing’ page featuring the recruitment video and details about the study. Recruitment materials seek non-heterosexual youth to participate in a paid study about health behaviors and adolescent peer relationships. Interested youth are led through a series of online steps to determine eligibility, facilitate assent (which is obtained under an IRB-approved waiver of parent consent to protect the confidentiality of youth), and provide contact information (i.e., cell phone number, email address, zip code). Eligibility and assent are verified by study staff, participant information is screened and verified to prevent fraudulent participants from enrolling, and a unique link to the baseline survey is sent to each participant’s email address to initiate enrollment. The same participants are retained for the duration of the study to complete assessments at multiple time-points using surveys, daily diaries and qualitative methods described in greater detail in the following sections. Our initial recruitment plan was based heavily (though not solely) on advertising via targeted social media campaigns (e.g., Facebook, Instagram, TikTok). However, as detailed in the Recruitment Challenges section below, this approach was of limited success due to an influx of fraudulent actors at project outset, and later, due to the institution of federal policies that restricted the targeting of youth through advertising on social media platforms. We have pivoted to increase our efforts to recruit participants via flyers, and online promotion through community agencies serving non-heterosexual youth in NYS, including school-based clubs across the state. We also have had some success with using an incentivized peer recruitment approach built on the principles of Respondent-Driven Sampling (RDS). Participants each receive three electronic referral coupons to distribute to their 14- to 18-year-old non-heterosexual friends. Friends contacting the study provide the unique coupon number, allowing staff to compensate the referrer $10 for each coupon. Procedures: Web-Based Longitudinal Surveys Participants complete four surveys at 6-month intervals over the course of 18-months (Baseline, 6-, 12-, and 18-months). Surveys are administered online using a secure server. We are utilizing web-based data collection because this method has been found to result in high-quality data that are available more quickly and cost-effectively than traditional methods,35 and supports our efforts to seek a diverse sample from across NYS. We strategically chose NYS as our target recruitment venue (versus a national sample) in order to control for potential confounding factors, particularly differing state laws regarding the rights and benefits of non-heterosexual individuals.36, 37 Our online data collection strategy also is well-suited for assessing sensitive topics because it is perceived as more confidential, thus yielding higher reports of socially undesirable behaviors than other methods.38, 39 Research indicates that nearly 73% of non-heterosexual youth (compared to 42% among heterosexual youth) report that they are more honest online than during face-to-face interactions.36 Furthermore, nearly 95% of 14-17 year olds in the U.S. access the internet.40 Participants provide electronic assent prior to initiating the survey (i.e., parental consent is not required). Participants are encouraged to complete the survey within a 2-week period, privately, in one sitting; however, programming allows for survey completion in multiple sittings, if desired. A 6-month assessment interval was selected to capture changes in the time-varying measures (e.g., peer victimization, alcohol use). Each survey is approximately 1 hour in length. Power analysis: longitudinal survey data (N = 400). We calculated power using a Monte Carlo simulation approach.207 We set α at 0.05 and generated and analyzed 1,000 datasets for each effect of interest. Based on Thompson et al.,11 we estimated effect sizes using a path model including bully victimization and alcohol use, assessed at four time points. To calculate power, we allowed bully victimization to be prospectively associated with alcohol use (b = 0.40), and alcohol use to be prospectively associated with bully victimization (b = 0.30). We set autoregressive effects to be 0.65. We estimate patterns of missingness (completely at random) as 1 pattern (p = 1.00), where alcohol use is missing on 25% of cases. Estimates suggest that the proposed final sample size (N = 400) with four assessment points will have power greater than 0.97 to detect significant effects of cross-lagged paths. We also estimated the power necessary to test moderation by entering a moderator variable and the interaction term at T3 to predictor T4 outcome. Calculating using a main effect of 0.40 and an interaction of 0.20, results showed that we would have power of greater than 0.97 (cross-lagged paths) and 0.97 (interactions) t0 detect significant effects. Procedures: Daily Diaries All participants (N = 400) will complete a total of 8 weeks of brief (~5 minutes) daily assessments, administered in two separate 4-week bursts timed to begin 3- and 15-months from baseline. Currently, there are 115 participants who have completed or are active in the first wave of daily diary bursts and 16 participants completed or active in the second wave of bursts. A text message containing a unique link to the daily report is sent to participants with a prompt to complete the report each day. Web-based technology is being rigorously applied by (1) validating it against written measures; (2) providing brief, individualized instruction to youth during enrollment; and (3) pilot testing.50 To ensure high retention, we: (1) offer monetary incentives (e.g., nominal electronic gift cards sent during daily assessment period), (2) continuously monitor compliance and prompt youth by phone, text, or email for completion; and (3) contact a youth-selected advocate directly if these attempts fail. To encourage consistent reports, compensation in the form of e-gift cards is provided, which include: $0.50 for each completed daily report, plus a weekly bonus ($10) for 4-5 reports OR $20 for 6-7 consecutive days of on-time reports, and a $25 bonus for providing 25-28 reports during each 4-week burst. Power analysis: daily data (N = 400). Power was calculated using a Monte Carlo simulation approach207 involving the generation and analyses of 1,000 datasets for each effect of interest (α at 0.05). Analyses followed a nested structure, with daily (i.e., within-subject) observations modeled at Level 1, and dispositional (i.e., between-subject) observations modeled at Level 2. We estimate patterns of missingness (missing completely at random) as one pattern (p = 1.00) where outcome variables are missing on 25% of cases. We conservatively estimated effect sizes corresponding to Cohen’s d = 0.15 for Level 1 main effects (i.e., within subject daily associations between bully victimization and alcohol use) and d = 0.30 for Level 2 main effects (i.e., between subject associations). According to simulations using these effect sizes, these tests would have power of greater than 0.99 (Level 1 main effects) and 0.99 (Level 2 main effects) to detect significant effects among 400 individuals over 56 days. Procedures: Qualitative Interviews Qualitative interviews (n = 90; 45 each male and female) are being conducted with participants reporting at least one bullying or sexual harassment incident during one of the 4-weeks of daily diary bursts. Interviewees are selected randomly each month to participate in a semi-structured video interview (i.e., via Zoom) about the details of their most recent PV experience.52 Interview questions, modeled after those used by our team in prior studies,57-60 explore unique situational factors associated with PV experiences among non-heterosexual youth, and provide insight into the factors hypothesized to mediate the relationship between PV and alcohol use (see Figure; Aim 3). Participants are asked to provide their account of the target PV incident, followed by prompts to fill in missing details (e.g., timing, setting, perpetrator characteristics [sex, relationship to, age, if known], precipitating factors (e.g., prior attacks), use of aggression, severity of incident, presence and involvement of others). The role of AOSU, if any, is assessed (e.g., quantity, timing of alcohol consumption; perceived intoxication of self/perpetrator[s]). Participants are asked to describe any conflict resolution and help-seeking following the incident. Perceptions of the perpetrator’s motives, including the role of participant’s sex in the incident, are assessed. Youths’ perceptions about the impact of the incident on their social relationships (e.g., peer and parent-child interactions/relationship, resulting disclosure [if any] of non-heterosexuality to friends and family) and how this experience shaped social relationships, behaviors, and coping in subsequent contexts (e.g., changing peer groups, increased AOSU, increased concern with rejection, fear of disclosure) are explored. Interviews last 30-45 minutes and youth are compensated with $25 e-gift cards. Currently, we have completed 47 interviews. Rationale for Requesting Additional Funding under PA-20-272 To successfully achieve the parent study aims, we request supplemental funds (PA-20-272) to extend our staffing and complete our longitudinal mixed methods data collection with our revised target sample (N = 400, original target N = 500). This request results from several unforeseeable challenges encountered by our team that significantly delayed our recruitment and data collection goals. Recruitment Challenges Delayed release of funds for participant incentives: Our grant funds were released September 9, 2021; however, the funds that had been budgeted for participant incentives had been cut from the award. Successful execution of our study aims is dependent on data collected from 400 youth who complete a series of longitudinal mixed methods assessments over an 18-month period. Thus, the participant incentives were critical to our successfully achieving the stated aims. After alerting our grants administrator, Dr. Frann Gallogly, of this oversight, funds for incentives were reinstated on January 17, 2022. This delayed initial project startup and recruitment. University administrative delays: Upon receiving the funds to support payment of incentives to our participants, we discovered that the only university-approved vendor for paying research participants was inappropriate for our study due to high administrative costs associated with re-loading e-gift cards. US Bank charges $1.00 for each time a new e-payment is made; given the large number of payments to be distributed to each participant across the longitudinal study, these fees were entirely cost-prohibitive for our study. We proposed an alternate vendor but the university required extensive review before approval. This approval process required edits to the vendor’s terms and agreements by University legal, which resulted in several months of additional recruitment delays for our study. The vendor was finally approved as a ‘pilot project’ for our study in January 2023 and we began actively recruiting participants in February 2023. Fraudulent participants: We launched our recruitment campaign in social media and at local venues as proposed in our grant application. Soon after launching (April 2023), we discovered that numerous fraudulent participants had infiltrated our study via social media advertisements. In response, we immediately stopped recruitment and enrollment in order to revamp our screening process and investigate and remove fraudulent cases from our dataset. This was an enormously labor-intensive process that took several months to complete. After ensuring that our procedures rigorously screened-out fraudulent participants before enrollment, and that prior suspect cases were purged from the dataset, we began recruitment again. Incentives lost to fraudulent participants were reimbursed to the study by the School of Nursing dean. Internal University audit: Shortly thereafter, in December 2023, we were informed that our procedures would be audited by the university to ensure that no additional incentives would be lost to fraudulent participants. This audit required intensive administrative oversight by our team, which resulted in slower recruitment and processing of eligible participants. The audit was completed in February 2024 after which we reinvigorated our recruitment efforts. Social media advertising restrictions: Unfortunately, the timing of our renewed recruitment efforts coincided with federal oversight concerns about the dangers of social media exposure to children, resulting in historic changes made across social media platforms to manage and reduce young peoples’ exposure to potentially inappropriate content. Our plan to utilize social media as our primary recruitment vehicle was derailed as a result of these changes. Our paid IRB-approved advertisements, which had previously been accepted for dissemination on Instagram, were repeatedly denied and were ineffective when they were launched (likely due to background algorithms screening content for our original target age group). Pivoting to new recruitment methods: Faced with extremely slow recruitment via these social media outlets, we pursued alternate recruitment strategies. The most successful of these efforts was recruitment via school-based clubs in high schools across New York State in conjunction with incentivized snowball recruitment. Our staff are actively attending in-person and virtual meetings in school-based clubs to recruit participants for the study. As a result, our recruitment has improved substantially but enrollment is at a slower pace than anticipated. Therefore, we continue to lag behind our planned recruitment timeline. In summary, we require 400 eligible and enrolled participants to have sufficient power to achieve our study aims. The administrative delays have set our recruitment efforts back by over a year. We are currently in Year 4 of the grant and have 225 participants enrolled. We have pivoted from our initial plan to recruit via social media to conduct in-person and virtual presentations to school-based clubs in high schools across New York State. This method has been successful but is labor intensive. We are confident that we can achieve our goal of recruiting 400 participants, but we lack sufficient resources to do so given the prior challenges, during which our staff were being paid. Thus, we are seeking administrative funds to extend staff salaries to achieve those goals. An extension of staffing will allow us to continue to recruit participants via school-based clubs and complete longitudinal data collection from those participants to ensure adequate power to successfully achieve the study aims.