This mixed method, explanatory sequential study was conducted among individuals who identified as members of the LGBTQ+ community between January 2021 and September 2021 in Atlanta, Georgia.

This study was conducted according to the guidelines 2018 Federal Regulations 21 CFR 56.110(b) and 45 CFR 46.110(b) (for expedited review) and was approved under categories _6, _7 per 63 FR 60364. All procedures involving research study participants were approved by the Mercer University Institutional Review Board on April 2, 2020. Written informed consent was obtained from all participants.

On the opening page of the survey, participants were presented with information about the study’s objectives and scope, followed by an informed consent statement. Participants were free to withdraw at any point without consequence. In addition, participation was anonymous, and no personally identifiable information was collected or used. Informed consent was obtained from all participants involved in the study, in compliance with the H2004105 Mercer University Institutional Review Board protocol. Participants were compensated for their participation via a US $40 Amazon gift card for completion of both the intake survey and participant interview.

Participants were recruited using convenience sampling to recruit participants for the study. Using statewide organizational connections, community partnerships, web-based or email advertising, and word-of-mouth recruitment, 60 participants were recruited. Eligible participants were those aged 18 years or older, currently residing in Atlanta, Georgia, capable of providing informed consent before starting the survey, and identifying as members of the LGBTQ+ community.

After enrolling in the study via SuperSaaS software (SuperSaaS B.V.), survey participants were emailed the link to the Qualtrics survey with the informed consent and intake survey to complete prior to their interview participation. Participant information was deidentified using a participant identification number assigned to all registrants via the SuperSaaS software [48]. Reminder emails were sent 1 week prior to the session, with follow-up emails sent 3 days before the interview session. Follow-up emails were sent to nonrespondents to encourage participation. If no response was received by the morning of the final day of data collection, those participants were removed from the study, resulting in a cohort of 47.

Demographic data were obtained from the 15 initial intake questions distributed after participants had signed up for an interview session. The intake survey was distributed via Qualtrics, took approximately 5‐10 minutes to complete, and was developed with input from a subject matter expert specializing in LGBTQ populations. Scale items were adapted from a previous HPV study assessing 3 key constructs: risk perception (3 items), self-efficacy (3 items), and behavioral intention (4 items) using a 7-point Likert scale(1=strongly disagree to 7=strongly agree) [49]. The total score was calculated from the sum of the response items, with higher scores indicating a greater perceived risk of HPV. An additional LGBTQ-related demographic question was added, but was not revalidated or pretested.

Due to special precautions associated with the COVID-19 pandemic, all interviews were conducted via Zoom and lasted about 45 minutes. Three student researchers previously trained in qualitative data collection, ethics, and compliance were enlisted to facilitate the interviews. Student researchers verified that the informed consent and intake survey were completed prior to the start of the interview, and using the intake survey, we were able to capture demographic data from our sample. The interview protocol consisted of open-ended questions adapted from a similar study surrounding the initiation of pre-exposure prophylaxis amongst Black men who have sex with men [50]. Two versions of the interview protocol were developed based on vaccination status. If participants indicated that they were vaccinated, their interview focused on the facilitators/motivating factors that encouraged them to do so. On the other hand, if indicated that they were not vaccinated, interview questions focused on the barriers to completing the vaccine series. Following the completion of the interviews, student researchers accessed the Zoom transcript files and cleaned them for data analysis.

Descriptive statistics were computed for all variables using StataSE (version 14; StataCorp LLC). The total number of responses and percentages were reported for categorical variables; mean and standard deviation were reported for continuous variables.

Using NVivo software (QSR International), 2 researchers trained in qualitative data analysis and NVivo software technology coded the 47 interviews. Using the constant comparative method, a codebook was developed for thematic coding and analysis by the research team [51]. Additionally, to reduce bias and to establish interrater reliability, the researchers developed a schedule for coding and a coding scheme. Each week, the researchers would code 10% of the transcripts independently and then merge coding findings into one dataset. The research team would then meet to run a coding comparison query in NVivo. Any data set sample that reported an agreement less than 90% was discussed to ensure that each individual researcher understood the codes and associated definitions/applications. The data set was coded again by the researcher individually, and the process was replicated until 90%‐100% agreement was achieved. A coding agreement of 90%‐100% means that minimal to no disagreements were reported using kappa coefficient in NVivo and consensus was established [52]. After coding was completed, the data were organized using NVivo software to explore emerging patterns and themes.

The diverse racial, cultural, and professional backgrounds of the research team included Black Caribbean, Black US American, White US American, and Asian American. This diversity allowed for study participants to authentically express their thoughts and feel comfortable sharing their lived experiences. Four of the authors hold primary backgrounds in public health, with experiences in health equity, community-centered research, and systems-level changes that shaped the framing of research questions, participant response, and theme prioritization during the qualitative analysis, which enriched the study depth. The secondary author is a foreign-trained medical physician whose clinical experiences in low-resource settings shaped a unique understanding of the structural inequities within health care and the importance of the patient-physician relationship, allowing for a deeper contextual understanding of participant responses.

Furthermore, each author’s unique personal experiences shaped how we approached the research process, and we acknowledge that our diverse professional and cultural contexts may have introduced potential biases within the analysis. To avoid speaking for the data, the team used reflexivity and iterative discussions. This included note-taking, which allowed the team to document and bring to light preconceptions while simultaneously ensuring that our interpretations were grounded in participants’ voices. The team sought to ensure that our research findings were both ethically grounded and accurately interpreted through collaborative discussion and critical reflection.

Although initially 60 participants signed up to participate, only 47 total interviews were conducted, resulting in a 78.33% completion rate. Demographic characteristics observed within our cohort were as follows: 51.1% (24/47) of the cohort were 27 years and older, and 48.9% (23/47) of the participants were White (Table 1). The breakdown concerning gender identity were as follows: 44.7% (21/47) of the participants identified as a man, 36.2% (28/47) of the participants identified as a woman, 6.4% (3/47) of the participants were nonbinary, and 4.3% (2/47) of the participants were gender fluid. A total of 29.8% (14/47) of the participants of our sample identified as gay, 19.1% (9/47) of the participants were bisexual, 14.9% (7/47) of the participants were lesbian, 12.8% (6/47) of the participants were pansexual, and 8.5% (4/47) of the participants were queer. In addition, 53.2% (25/47) of the participants indicated that they were single, and 19.1% (9/47) of the participants indicated that they were in a monogamous relationship. Around 74.5% (35/47) of the participants of our sample had been sexually active in the last 6 months, and 14.9% (7/47) of the participants engaged in unprotected sex of some kind. Thirty-six out of 47 (54.4%) participants indicated that their total lifetime partners were between 5 and 29 (54.4%) and 29.8% (14/47) of the participants never had a sexually transmitted disease. Forty-three out of 47 (91.5%) participants had not been diagnosed with HPV, and 28/33 (84.9%) participants tested indicated that they were HIV negative. Additionally, 22/47 participants (46.8%) indicated that they initiated the HPV vaccination.

Of our sample, 68.2% (15/47) of the participants indicated that they had completed the HPV vaccination series (3 doses received) and most 46.8% (22/47) of the participants had health insurance through an employer followed by 19.1% (9/47) of the participants having private insurance and 17.0% (8/47) of the participants being public/government sponsored (Table 1).

Risk for getting HPV (how susceptible they were to contracting the virus) was assessed using 5 statements on a 1 to 7 Likert scale (Table 2). When provided with the statement “I am likely to get HPV,” 27.66% (13/47) of our participants responded not sure, followed by 23.40% (11/47) of the participants who strongly disagreed. Overall, 29.79% (12/47) of our study sample strongly disagreed with the statement “I am at high risk for getting HPV” and 27.66% (13/47) of the participants indicated that they were not sure when presented with the statement “HPV would be a serious threat to the quality of my life.” A total of 29.79% (14/47) of the participants responded not sure to the statement “HPV would be a severe threat to my health and 27.66% (13/47) of the participants strongly agreed that “HPV would be a severe threat to my sex life.” Cumulatively, our sample responded with an average score of 18.73 out of a possible 35 for all 5 statements used to measure perceived risk (Table 3).

Participant interviews revealed that “Real Outcomes,” and “Accurate Representation” were the main components that should be considered for inclusion in an HPV documentary tailored for sexual and gender minority populations. Interviewees also revealed insights into their unique “Health Care Experiences” and the “Importance of using social media to disseminate information.”

Our goal within this study was to identify what characteristics are essential to include in the development of HPV-related vaccination intervention for an LGBTQ+ audience and to identify dissemination methods that would ensure virality. This study has highlighted 2 key items to include in digital stories: real outcomes and representation.

Past research surrounding patient narratives has shown them to be a useful tool to gather information, communication, engagement, persuasion, and health behavior change [54]. DST expands upon the utility of patient narratives through its ability to focus on capturing the experiences of populations to share findings in an engaging manner through digital media [34]. However, the use of DST has some trade-offs with health care professionals arguing that the quality or accuracy of media-based DST is a major concern due to their lack of medical expertise [54]. Thus, the creation of any DST method for any behavioral intervention should allow for the opportunity to collaborate with health care professionals to ensure the provision of reliable, evidence-based medical guidance.

Results from our formative inquiry support the well-reported health disparities that the LGBTQ+ population frequently experiences, such as stigma and discrimination from health care providers [43-46]. Stigma and discrimination in health care settings undermine effective communication, which is essential for delivering quality care and fostering self-confidence. When comparing the communication and care quality between races, researchers found that Hispanic and Asian survivors reported poorer communication than White cancer providers, with Asians also reporting poorer care quality [55].

There is a pattern of discrimination, stigma, and lack of awareness among health care providers that has impeded LGBTQ+ populations from accessing health care services and has had a major impact on their health [41] resulting in less sexually transmitted disease screenings, inappropriate contraceptive counseling, less contraceptive use, and misinformation on sexual education and reproductive promotion. For example, studies have shown that lesbian women have lower use of sexual and reproductive health services, leading to significantly lower rates of pap smear testing [56]. Additionally, the Center for Disease Control has reported that 71% of new HIV infections seen were among gay and bisexual men in the United States in 2022 [57]. The use of DST within this population would allow for increased engagement and connections with others (ie, patients, advocacy groups, caregivers, health care professionals, and policy makers) that share similar lived experiences. Furthermore, through sharing their unique stories via this method, they can alter preconceived notions about their susceptibility/risk, attack misinformation concerning vaccinations, and reduce the health care disparities seen, thus improving the overall health of the LGBTQ+ population.

Several limitations exist within the study. First, our sampling methodology could be a limitation [58]. We used voluntary convenience sampling to recruit study participants [59]. It is argued that this type of nonprobability sampling is not truly representative of the entire population, as the sample is likely to overrepresent or underrepresent certain groups and mainly targets persons with stronger opinions. Furthermore, this method has a low external validity and a lack of transferability to other populations [60]. Future research should consider partnering with community organizations during the recruitment period, thus diversifying participant groups and perspectives. Second, our cohort was concentrated within the Atlanta area. Atlanta has a unique demographic, cultural, and public health characteristic, which includes a greater exposure to public health campaigns that may influence overall health behaviors and beliefs. DST interventions conducted in other states/ countries may encounter alternative barriers or facilitators to HPV vaccination and thus may discover that other elements may need to be included.

A more in-depth analysis of the resulting themes could have yielded a deeper understanding of participant viewpoints and their implications. While we used a mixed methods research design, we failed to triangulate the data collection due to the practical constraints related to participant recruitment, such as patient and researcher availability and stringent COVID-19 protocols. As a result, the conclusions and insights from the participants were based on a single approach, thus limiting their validity. Additionally, as we did not collect or analyze data regarding differences in findings between vaccination status and other characteristics, which limited our ability to explore how the perceived risk of HPV may vary across subgroups. Future research should consider analyzing this area in order to not only provide a deeper understanding of the effect that individual factors may have on perceived risk but also provide insight into how to tailor public health interventions.

The LGBTQ+ community is often misrepresented, underrepresented, or not at all represented in media, nor do they often feel safe and comfortable enough to share their experiences with others outside of their community. The use of DST methods involving real outcomes and accurate representation would bridge the gap created by stigmatization in order to improve HPV vaccination adherence and reduce health care disparities within this population.