A virtual randomized controlled trial of an alcohol exposed-pregnancy prevention mobile app with urban American Indian and Alaska Native young women: Native WYSE CHOICES rationale, design, and methods

Abstract

Background:

Fetal Alcohol Spectrum Disorders (FASD) result in lifelong disability and are a leading cause of preventable birth defects in the US, including for American Indian and Alaska Natives (AIANs). Prevention of alcohol exposed pregnancies (AEPs), which can cause FASD, is typically aimed at adult women who are risky drinkers and have unprotected sex. Among AIANs, AEP prevention research has been primarily conducted in reservation communities, even though over 70% of AIANs live in urban areas. Culturally appropriate AEP prevention for urban AIAN young women, regardless of current drinking or sexual behaviors, may maximize the potential for primary prevention at the beginning of the reproductive years for this underserved population.

Methods:

We developed a virtual randomized controlled trial (RCT) – fully implemented through technology – to evaluate Native WYSE CHOICES, a culturally tailored mobile app, with urban AIAN young women ages 16–20 nationally. While virtual RCTs are not new, this is the first engaging a solely urban AIAN population, historically excluded from research. Participants are recruited on a rolling basis through the project social media community, organizational partnerships, and in-person events. Eligible participants complete a baseline survey and are randomized to either the app’s intervention or comparison arm – each of which provide about 3 hours of content. Follow-up data are collected at 1-, 6-, and 12-months post-baseline.

Results:

Our study offers a template for building trust and extending reach to this underserved population while also providing important lessons and insights on advances in virtual or hybrid research approaches.

1. Introduction

Fetal Alcohol Spectrum Disorders (FASD) result in neurodevelopmental deficits and lifelong disability; they are a leading cause of preventable birth defects in the U.S.^{1, 2} Any sexually active woman of reproductive age who drinks alcohol and does not use effective contraception is at risk for an alcohol exposed pregnancy (AEP) that could cause FASD.² While AEP prevention efforts typically target adult women who are behaviorally at risk for AEP, teens and young women may be an especially important group to reach regardless of current behaviors because of the potential for primary prevention at the beginning of their reproductive years.³ Considering the deep cultural values that center children’s health and wellbeing, American Indian and Alaska Native (AIAN) communities have advocated for such primary prevention efforts.⁴ Available data suggest that AIAN teens and young women are an important group to reach for AEP prevention to protect both their health and the health of future generations. Compared to their peers, AIAN high school youth report the highest levels of ever having had sex (50%) and being currently sexually active (39%) while at the same time, 80% of sexually active youth reported not using effective contraception at last sex.^{5, 6} In addition, compared to other race groups, AIANs aged 12–17 and 18–25 reported the highest levels of heavy alcohol use in the past month except for Whites.⁷ Together, these factors may make AIAN teens and young adult women vulnerable to AEPs. Indeed, one study estimated that 25% of AIAN teen births in one community setting were substance involved, with alcohol accounting for more than half of the substance involved pregnancies.⁸

Efforts to address AEP risk among AIANs include adaptations of CHOICES, an evidence-based brief AEP intervention supported by the CDC.⁹ Research demonstrates CHOICES is effective in reducing AEP risk across varied settings, modalities, and populations—all with women aged 18–44 who are sexually active, not using effective birth control, and engaged in risky drinking.^10–13 In the Oglala Sioux Tribe, at-risk AI women recruited at a tribal health clinic who received CHOICES sessions by telephone ¹⁴ showed a significant decrease in AEP risk.¹⁵ A culturally tailored, clinic-based CHOICES is now being evaluated in a randomized clinical trial (RCT) with at-risk AI women in a rural reservation community.¹⁶

While AEP prevention continues to advance for adult AIAN women who are at risk for an AEP, a substantial gap still exists for AIAN youth. AIAN youth have not been included in most AEP risk prevention research, despite demand supported epidemiologically and by community members.^{4, 17, 18} Additionally, while an estimated 72% of AIAN young women live in urban areas,¹⁹ research relevant to AEP is largely based on reservation samples.^20–23 The intervention research base for urban AIAN youth and young adults remains thin.^24–26 Urban AIANs tend not to live in concentrated neighborhoods but rather are dispersed across neighborhoods within the urban landscape; thus, common community-based recruitment and intervention strategies are expensive and rarely consider relevant cultural adaptations for urban AIAN people.²⁷ Consequently, urban AIAN people are often not included in research aimed at developing, implementing, and evaluating culturally appropriate interventions. Recent research indicates that mobile health (mHealth) interventions are promising for delivering effective interventions with such hard-to-reach populations,^28–30 especially youth ^{31, 32} and are promising for AIAN youth in particular.^{22, 33, 34}

We respond to this research gap and the promise of mHealth by designing a virtual RCT fully implemented through technology (including recruitment and intervention delivery) to evaluate a CHOICES-based, culturally tailored mobile app called Native WYSE (women, young, strong, empowered) CHOICES (NWC) designed for and with AIAN young women from urban areas across the United States. While online RCTs are not new,^35–37 they are rare for AIANs generally,^{33, 38} and to our knowledge, ours is the first to exclusively focus on the urban AIAN population nationally. Here, we describe the design and methods of the trial.

2. Material and methods

2.1. Community engagement

Using intensive community-based participatory research (CBPR) approaches, one of the co-authors partnered with a tribal community and AEP prevention experts to adapt CHOICES. Based on community guidance, the adaptation focused on all youth, regardless of risk. The result was a substantial shift from a narrowly defined focus on adult women at high risk for an AEP, to youth ages 16–20, regardless of sexual or alcohol experience (i.e. shifting to a universal prevention approach).³⁹ This adaptation included integration and engagement with traditional symbols and stories. For NWC, we used CBPR to adapt this program further, converting this universal in-person AEP prevention program to a mobile app for young AIAN women in urban communities. In tribal settings, the definition of community is relatively straightforward, based on geographic boundaries. Our study, however, involves participants living across many urban and suburban areas. To emulate CBPR principles as closely as possible, we pursued three strategies: (1) we assembled an advisory panel of leaders and researchers at both national and urban AIAN community levels, including urban AIAN young women to guide all phases of the project, (2) we conducted formative interviews (n=29) via phone or videoconference with urban AIAN young women to obtain their input on content, modality, and social media practices and preferences, and (3) we built a social media presence for the project, creating a space of safety and affirmation for urban AIAN young women, and a way to “give back” to the community.

2.1.1. Research review partners

Urban AIAN communities typically lack formal processes for research review and approval common to tribal communities.⁴⁰ Our advisory board built a strong foundation for establishing processes which will ethically, culturally, and methodologically conform to expectations of research with diverse urban AIAN communities.⁴¹ We also convened a Data Safety and Monitoring Board composed of experts in AEP, statistical analyses of RCTs, and urban AIAN population health. We developed a charter outlining policies and review procedures for RCT results. This project was approved by the University of Colorado Multiple Institutional Review Board (18-0574/20-3122), and is registered with ClinicalTrials.gov (#NCT04376346).

2.2. Participants and recruitment

2.2.1. Eligibility

Participants are eligible if they are 16–20 years old, assigned female at birth, live in an urban area (population 50,000+), are not currently pregnant or breast-feeding, can provide a personal email address, and own a smart phone. We exclude participants from Alaska since we were not able to secure approval from the relevant review entity in time for the trial. Because of the potential for fraudulent access to the study for the purposes of collecting participant compensation, we established a two-stage eligibility verification process. Once participants are confirmed as eligible from screener survey responses, staff confirms eligibility based on IP address and checks urban residence. Second, the study staff member confirms the screener survey data with the potential participant via phone call or text. Since fraudsters may not track survey details, they would be unlikely to confirm the data. All verified participants are sent a link to a consent form which includes HIPAA authorization. Our IRB provided a waiver of parental consent for the 16- and 17-year-olds in the study. The electronic consent includes a follow-up quiz on content to assess comprehension and requires an electronic signature. Participants could elect to receive a copy of the signed consent by email or by US Postal Service.

2.2.3. Recruitment

We leverage social media platforms, partnerships with other AIAN-serving organizations, and connections at in-person events (as possible with the pandemic) to recruit nationally.

Social media presence:

Prior research efforts have exploited AIAN communities, and social media itself can be a tool of abuse, misuse, and manipulation.⁴² To build trust in our project and a social media presence, we created social media project accounts well before our RCT launch, with a major focus on Instagram, a preferred platform of our age group. Our posts avoid AEP and related content and are designed instead to affirm and celebrate urban AIAN young women. While quantifying the direct impact on recruitment is not possible, our following has grown to over 4,400 in about 2.5 years on Instagram alone, facilitating a broad national context for our multi-pronged recruitment strategy.

Virtual recruitment strategy:

Nationally, 94% of AIAN youth cell phone users go online daily or more.⁴³ Tapping this connection, however, is not straightforward. With an established social media presence, our most proficient strategy relies on boosting Instagram recruitment posts, i.e., paying to amplify a recruitment post across a subset of users based on sex, age, and geography. When a prospective participant clicks on an ad, that individual is sent to a project screening survey and asked questions to determine eligibility. To identify specific areas to place boosted posts, we downloaded race population counts for all urban areas (population 50,000+, according to the 2010 Census), and calculated the proportion of persons identifying as AIAN alone or in combination with another race for each area. Using those areas with >0 proportion AIAN, we created 5 “replicates” of about 70 urban areas each, comprised of a random selection of urban areas, stratified by size. Since early October, 2021, we have rotated boosts weekly across these 5 replicates. This rotation allows us “concentrated” boosts without saturating the audience.

Media partnerships with other AIAN-serving organizations

Our social media toolkit includes links to Facebook, Instagram, or email versions of recruitment material. After confirming with the National Indian Health Service IRB that posting flyers did not constitute engaged research, we met with Directors of Urban Indian Health programs to present our project, and then sent each our toolkit. Similarly, we identified minority-serving colleges and universities and other AIAN-serving organizations in urban areas to ask them to post our material. For each, we make clear that if interested, we will reciprocate and post partner material on our accounts. Finally, we have also actively partnered with several organizations over social media via twitter chats, repostings, and other coordinated theme campaigns. Again, while not necessarily a means to directly recruit, such active partnership events heighten our visibility and help to establish legitimacy.

In-person outreach:

To complement our social media strategy, we also attend powwows, conferences, and community events (subject to pandemic surges), which serves to further establish our legitimacy with AIAN communities and provides opportunities for direct recruitment.

2.3. Intervention

2.3.1. Intervention description

The Native WYSE CHOICES mobile app supporting AEP prevention for urban AIAN young women includes 6 sessions, covering alcohol use, contraceptive information, AEP and FASD, risk assessment, goal-setting, and sustaining choices. Each session contains various methods of engagement, such as quizzes, alcohol and sexual risk assessment vignettes, decisional balance and readiness to change exercises, and strategizing options for support and resilience. Throughout, Auntie Amanda, a social media Influencer popular through a partner program, We R Native,⁴⁴ narrates the progress through ”touchstone” video clips. Through these clips, she provides a consistent presence across app sessions, acknowledging the challenges of young adulthood and providing encouragement for participants. (See Table 1 summarizing app content.) We hypothesized three key areas supporting behavior change for this population: (1) Knowledge acquisition; (2) motivating behavior change through risk assessment and goal setting; (3) infusing content with a culture-centered narrative. The first two of these align with motivational interviewing (MI),⁴⁵ a client-centric method of change through reflection and goal-setting, and a core element of the in-person CHOICES program. For knowledge acquisition, we use the Medicine Wheel as a learning heuristic, woven through the three informational sessions by way of short animations. For those unfamiliar with the Medicine Wheel’s structure and symbolism, the narrated animations invite participants to think of how their choices about alcohol, contraception, and condom use affect their physical, emotional, relational, and educational/occupational health. Building on the informational sessions and using a variety of engagement tools in the app, we approximated MI by providing participants with opportunities to make individualized choices, for example by creating goals, or through decisional balance exercises, assessing the positive and negative consequences of a given behavior. Throughout, participants receive affirming, non-judgmental feedback that emphasizes personal choice. The third area, centering culture, was a theme that emerged strongly from formative interviews, aligning closely with extensive prior research.^{26, 46, 47} Each session takes about 30 minutes to complete, for a total engagement time of about 3 hours.

Table 1.

Native WYSE CHOICES mobile app summary content

Session	Summary Content
Welcome & Orientation Session	• Auntie Touchpoint Video: Welcome and overview of the Native WYSE CHOICES app. • Motivational Interviewing (MI) Components: Decisional balance exercise using interactive questions and answers to assess interest in learning about alcohol, birth control, protecting the health of future generations; participant receives encouraging feedback on all responses. • Native WYSE CHOICES Medicine Wheel Video (Orientation): Overview, 4 aspects of health of Medicine Wheel.
Session 1: Learning the Facts about Alcohol	• Auntie Touchpoint Video: Provides overview of Session 1, stresses importance of learning about alcohol’s risks to make informed choices. • Native WYSE CHOICES Medicine Wheel Video (Alcohol): Shows how choices about alcohol can affect 4 aspects of health depicted in Medicine Wheel. • Knowledge Development: Interactive images on alcohol types/alcohol volume; multiple choice quiz with feedback for each question. • MI Components: Decisional balance exercise using questions and answer lists to prompt reflection on reasons and situations when participant would choose/not choose to drink, feelings associated with considering reason/situations when participant would drink/not drink, and strategies for making decisions about alcohol.
Session 2: Learning the Facts about Birth Control and Condoms	• Auntie Touchpoint Videos: (Video 1) Provides overview of Session 2, emphasizes importance of learning about birth control and condoms to make informed choices; (Video 2) reviews Medicine Wheel video, stresses importance of making birth control/condom use choices that are right for participant. • Native WYSE CHOICES Medicine Wheel Video (Birth Control): Shows how choices about birth control and condom use can affect 4 aspects of health in the Medicine Wheel. • Knowledge Development: Interactive chart to learn about different types of birth control and condoms. Stresses personal choice, talking to healthcare provider, and talking to a healthcare provider to learn more; multiple choice and true/false quiz with feedback for each question. • MI Components: Decisional balance exercise using questions and answer lists and vignettes to prompt reflection on reasons to use/not use birth control, feelings associated with considering reasons to use/not use birth control, and strategies for deciding to use/not use birth control.
Session 3: Learning the Facts about AEP and FASD	• Auntie Touchpoint Videos: (Video 1) Reviews participant progress and next steps in app; (Video 2) recaps session, reiterates what constitutes AEP risk and prevention, draws on cultural wisdom of considering impact of one’s choices on future generations. • Knowledge Development: Engaging images and text to build understanding about risks of drinking during pregnancy for AEP and FASD; quiz to test understanding about AEP risk; interactive images and text with direct statements about how to prevent AEP.
Session 4: AEP Risk Assessment	• Auntie Touchpoint Videos: (Video 1) Gives encouragement about progress in app, previews risk assessment in current session, stresses privacy of answers and support for making own choices; (Video 2) shares appreciation for honesty and provides non-judgmental support for using feedback to make informed choices about drinking and birth control. • MI Components: Risk assessment for binge/heavy drinking, unplanned pregnancy, and sexually transmitted infection; yes/no questions with feedback about risk on all responses.
Session 5: Goal Setting	• Auntie Touchpoint Videos: (Video 1) Reflects on last session, provides non-judgmental support for making choices aligned with goals in life; (Video 2) reinforces connection between choices and goals, expresses support for taking time to consider choices and goals. • MI Components: Multiple choice options to set goals to reduce/avoid risky drinking and reduce unplanned pregnancy and sexually transmitted infection risk, rulers to assess importance, certainty, and readiness to pursue goals, and prompts to consider concrete next steps to pursue goals.
Session 6: Staying on Track	• Auntie Touchpoint Videos: (Video 1) Compliments participant for completing sessions, reflects on accomplishments; (Video 2) expresses appreciation for participant’s work on topics, points to resources to help stay on track with goals. • MI Components: Questions with multiple prompts to consider reasons/motivations for selecting goals, what to do to stay on track with goals, what can get in the way of goals, and how to know goals are met.

2.3.2. Comparison arm program

In research with AIAN communities, using “treatment as usual” or “no intervention” for the comparison group is unacceptable. Communities feel strongly that all participants should benefit directly from a program when they engage in research.⁴² As a part of our formative work, AIAN young women provided insights through in-depth interviews on their preferred topics for the comparison arm in the mobile app. Based on that input, we created a 5-session mobile app featuring financial literacy, resume writing and professional development. Auntie Amanda also narrates this content. Each session is estimated to take about 30 minutes, for a total of 2.5 hours of engagement. We ask participants in both intervention and comparison arms to complete all sessions within 1 month.

2.3.3. Randomization

Once recruited and consented, eligible participants complete the baseline survey, and are asked to download the app, available on Apple Store and Google Play. Once opened, the app is programmed to randomly assign participants to either the intervention or the comparison arm. Both reside in the one app, which is programmed to deliver appropriate content to users based on their assigned arm. Participants are paid $10 in an electronic gift card for registering and downloading the app, and another $10 for completing all modules. Since we were able to assign treatment arms within one app, we have the added advantage of monitoring app engagement for both arms.

2.4. Study design

2.4.1. Overall research design

Enrolled participants are asked to complete sessions of the app arm to which they are randomized. Following intention-to-treat principles, they are then provided 1-, 6-, and 12-month follow-up surveys, regardless of level or extent of engagement with the app. Participants who miss a follow-up survey are still invited to complete subsequent follow-up surveys. Diagram 1 shows participant flow through the project.

Diagram 1.

Native WYSE CHOICES Participant Flow

2.4.2. Retention

Virtual RCTs have limited direct engagement with participants. This lack of personal connection could be deleterious to response rates at follow-up survey points. In addition, contact information for youth in this age group can change quickly. To counter this possibility, we designed several methods for incentivizing participation. Eligible participants are paid $25, $35, $40, and $50 respectively for the baseline, 1-, 6-, and 12-month surveys they were invited to complete. In between surveys, about every two months, we send an email to participants asking for verification of contact information. Upon verification, they receive an additional $10 e-gift card as a thank you.

2.5. Measures

With the youthful age of participants, and likelihood that they would be answering online surveys using a mobile phone with a small screen area, we limited the number of questions so that the survey could be completed within about 30 minutes. Data collection began in August, 2021.

2.5.1. Alcohol use

We assess alcohol use using questions from the Youth Risk Behavior Survey,⁵ including ever used alcohol at least once (not including sips for ceremonial or religious purposes), age of first drink, and number of days drinking in the past 30 days. We assess risky drinking by asking, in the last 30 days: the number of days a participant had 4 or more drinks; the number of drinks on one occasion; and whether a participant consumed 8 or more drinks within a 7-day period. All quantity questions were aided by a drink chart. We also assess if a participant had alcohol or substance use involved sex in the past 30 days.

2.5.2. Effective contraception

We assess use of effective contraception by asking participants what methods of contraception they used when they had vaginal sex in the last 30 days.⁴⁸ Condom use is also assessed for consistent use.²²

2.5.3. Abstinence

This intervention program is universal. It does not exclude participants based on risk status and instead is designed to support young women in their choices to remain abstinent (either sexual or alcohol use), or to return to abstinence after the program. Thus, we also measure the effect of the program on these states (continued abstinence or return to abstinence).

2.5.4. Knowledge and self-efficacy measures

We conceptualize knowledge and self-efficacy as mediators of behavioral outcomes. These are assessed by survey items on AEP knowledge,⁴⁹ effective contraceptive knowledge,⁵⁰ alcohol self-efficacy,⁵¹ contraceptive self-efficacy,⁵² and sexual practices self-efficacy. ⁵³

2.5.5. Other measures

Other measures include age, educational status, gender and sexual identity, marital status, employment status, household composition, and family connectedness. Because household economic status is difficult for many young people to report, we ask about governmental assistance for the household, for example, food stamps, free/subsidized school lunches.⁵⁴ Finally, we include measures of enculturation⁵⁵ to tap participants’ cultural connectedness.

2.5.6. COVID measures

In July, 2021, we were awarded supplemental funding to assess AEP in the context of the pandemic in both quantitative and qualitative terms. We added questions about experiences with COVID, including disease, risk exposure, and safety; education and socioeconomic impacts; vaccine status and attitudes; and cultural and historical impacts. While many of these measures were based on current questions found in Native-focused surveys or the PhenX Toolkit,^{56, 57} the pandemic has evolved rapidly and few questions relevant for this population and the rapidly changing epidemiological and social context exist – many measures were developed by the project team and tested in a pilot survey, fielded May-August, 2021. Qualitative data collection is ongoing.

2.6. Data analysis

2.6.1. Baseline equivalency.

Prior to other analyses, we will confirm baseline equivalency of intervention and comparison group participants through comparisons (using Chi-square and t-tests) of demographic characteristics and pre-intervention measures of self-efficacy, alcohol use, and sexual risk. We will monitor attrition-related bias throughout the follow-up periods by evaluating the cross-wave comparability of participation levels, demographic characteristics, and behavioral risk profiles in both study groups. Any non-equivalency will be handled by appropriate methods (e.g., propensity scoring).⁵⁸

2.6.2. Data distributions and missingness.

Distributions of all outcome measures will be inspected for patterns of missing data. When normality assumptions are not met for continuous or ordinal measures (e.g., alcohol frequency/quantity or self-efficacy), we will use transformations to achieve normality or create collapsed or dichotomized measures where appropriate. Since data collection will be computer-administered, we anticipate that missing data will be minimal. If initial analyses provide support for at least a missing-at-random assumption, we will analyze data using Mplus⁵⁹ which provides Full Information Maximum Likelihood methods, where all available data are used in the estimation of model parameters.

2.6.3. Analyses.

All intervention effectiveness analyses will be based on an intention-to-treat approach. For continuous outcomes (frequency/quantity indices of alcohol) and mediational outcomes (alcohol and contraceptive self-efficacy), we will estimate a series of latent growth curve models (LGCM) to assess the effect of intervention exposure on the trajectories of those outcomes over the study period.⁶⁰ The “normative” trajectory of an outcome is estimated for the comparison group before estimating the same trajectory among those exposed to the intervention. The third step estimates the degree to which the normative growth trajectory is altered by intervention exposure. On the final step, the interaction between intervention exposure and pre-intervention status on the outcome is tested. This permits an assessment of the differential effectiveness of the intervention as a function of baseline levels of the outcome variable.

For dichotomous outcomes, Latent Markov Models (LMM) will be used to evaluate transitions between behavioral states (e.g., sexually active to abstinent) throughout follow-up.^{61, 62} Intervention status is modeled as a time-invariant covariate; its effect on the transition probabilities is estimated freely. Effectiveness is indicated if those in the intervention group have significantly lower probabilities than comparison group participants of transitioning out of the abstinence state and significantly higher probabilities of transitioning out of the sexually active state. LMM may also reveal that early sexual experience among comparison group participants predicts sexual activity across the course of the study but that this relationship is interrupted among intervention participants. Other moderators of intervention effectiveness will also be explored (e.g., enculturation, baseline risk status).

Both sets of analyses will allow us to examine patterns of intra-individual change in AEP risk factors over time and estimate the impact of intervention exposure on those temporal trends. By testing interaction terms involving baseline risk levels and study arm, we will also be able to assess the differential effectiveness of intervention exposure among participants with varying levels of baseline AEP risk, including those with no or small baseline risk.

2.6.4. Power and sample size calculation.

We used the Muthén and Curran approach to estimate the sample size needed to detect an intervention effect in the LGCM analyses.⁶⁰ Effect sizes were based on prior CHOICES-related studies and meta-analytic reviews of nearly 150 studies related to motivational interviewing on health behavior outcomes (including alcohol and sexual activity).^{63, 64} Our power analyses yielded a total final sample size of approximately 525 in order to detect a .20 intervention effect with 80% power at an alpha level of .05. Retention estimates for studies using electronic interventions with adolescents varied widely, from 43–97%⁶⁵ as do those for electronically administered brief screening interventions for alcohol use (50–99%).⁶⁶ Although we have presented a strong retention plan, we based our estimates on a conservative rate of 75% of baseline retained at 12-month posttest. Thus, we originally projected a recruitment of 700 participants at baseline to provide sufficient power for the proposed intervention analyses. Subsequent analysis based on actual project data indicated that collecting baseline sample beyond about 425 would not sufficiently alter our power to detect effects.

3. Results and discussion

The preceding describes a virtual randomized trial to test the effectiveness of a mobile alcohol-exposed pregnancy prevention app, designed for and with urban AIAN young women. The virtues of a virtual RCT became apparent with the arrival of the pandemic. Projects designed to work in-person with communities were scrambling to find virtual connection as communities, especially AIAN communities who had been hit particularly hard with COVID-19, shutdown or severely limited any in-person activities.⁶⁷ While we continue to adapt to a post-pandemic world, elements of hybrid research will likely continue. This project provides elements of guidance for best practices in planning virtual elements. Our project has been designed to build a virtual community prior to recruitment; verify participant identity to uphold the integrity of data responses from this population; and distribute recruitment advertising rotationally across the country both to avoid saturation and to reach those in small urban communities.

Our project is well-positioned to extend AEP prevention work ongoing in reservation communities to youth and young adult AIAN women – the majority – living in urban areas. This work leverages the interest and high levels of engagement of AIAN youth with technology to reach those who otherwise may have few culturally tailored resources. The reach of social media as a recruitment tool means that even those who reside in areas with limited cultural engagement can feel connected, an experience reinforced through the project social media community. In our study, we use social media to recruit, but prior to doing so, we have used it to build a presence and sense of community with our intended population. The distrust of many AIANs of research because of past abuses likely continues in virtual research, perhaps even more so given limited researcher-participant interaction. Building a social media presence has contributed to a sustained engagement with urban AIAN young women and those who support them as evidenced by a growing number of followers and content sharing. It has provided a way to deliver affirming messages and relevant resources, thus also providing a way to give back to the community. We believe this effort, in turn, facilitates recruitment.

Virtual recruitment also invites fraudulent participation. Fraudulent participation compromises the outcomes of the research and dilutes true voices of the population of focus. Our project sets a high standard for participant verification thus honoring those sharing their experiences through the project survey. Finally, our virtual recruitment advertising method is tailored to an urban AIAN population, based on census data and rotating replicates. The replicates, each including a range of small and large urban communities, serve to include urban AIAN young women who may be isolated from cultural connection. These design elements, set within a context of intentional and meaningful urban AIAN oversight and advisory guidance, provide a strong blueprint for virtual RCTs.

The many strengths of virtual RCTs with AIAN populations need to be assessed against their disadvantages. As with other virtual RCT’s we had to design manual processes for verifying potential participant identity since we experienced fraudulent hits on the screening surveys.⁶⁸ The many pieces of data collection and retention required an extensive database of tracking indicators. And of course, the inability to connect personally with participants and with the communities where they live is a challenge to an engaged research process valued by AIAN communities and researchers alike.

4. Conclusion

The study described in this article provides an outline of a unique and innovative approach to research with AIAN communities. With the advent of the COVID pandemic, when so much research was forced to move to remote administration, many have learned about the challenges and advantages of virtual engagement. Hybrid research models are likely to arise quickly as we learn about the proclivities of post-pandemic research. Our description here provides intentional planning and design for virtual research and thus may contribute needed guidance for those pursuing this type of research with AIAN individuals or other hard-to-reach communities in the future.