In 2019, there were 13.2 HIV diagnoses per 100,000 people in the United States. The Ending the HIV Epidemic (EHE) initiative was announced with the goal of reducing new HIV infections by 90% by 2030 [1]. A key pillar of the EHE is helping more people with HIV achieve and sustain viral suppression to decrease transmission of HIV [2]. Florida, an epicenter of the HIV epidemic in the United States, has high rates of HIV incidence (19.9 per 100,000 people) and prevalence (578.7 per 100,000 people) in both urban and rural settings and in people with a diversity of racial and cultural backgrounds [3-5]. When the Florida Cohort study wave 3 began in 2019, only 70% of the 123,087 people with HIV in Florida had achieved viral suppression [5]. Identifying barriers that prevent people with HIV from sustaining viral suppression and achieving other care continuum steps, as well as facilitators associated with successful HIV outcomes, within Florida is needed to achieve the ambitious targets set by the EHE.

Numerous factors at the individual, interpersonal, community, and policy levels influence HIV viral suppression [6-8], including alcohol use, other drug use, housing status, mental health conditions, HIV stigma, having low income, and the lack of insurance coverage, among others [9-13]. Trends in alcohol use are changing as alcohol use among women has increased, and the COVID-19 pandemic has increased alcohol use in the general population [14,15]. At the same time, the influence of alcohol use on viral suppression may be lessening as antiretroviral therapy (ART) regimens become simpler and more forgiving [10,16,17]. However, alcohol use and ART can both negatively impact liver function. Therefore, monitoring alcohol use and interest in alcohol-related interventions remains important for improving the health of people with HIV [18].

Many of these factors that directly or indirectly impact viral suppression are infrequently recorded in medical records or surveillance data and can only be assessed by directly administering questionnaires to people with HIV. While there are many ongoing cohort studies of HIV, the Florida Cohort study is the only one focused on monitoring HIV-related factors across an entire state. In the previous iteration of the Florida Cohort study (wave 2), we collected data from more than 900 people with HIV between 2014 and 2018, linking data from individual questionnaires, medical records, and statewide HIV surveillance [19]. From these data, our team identified several factors associated with poor HIV viral suppression, including heavy alcohol consumption, other substance use, depression and anxiety symptoms, and suboptimal ART adherence [20-25]. We also examined several dimensions of HIV stigma and found that only health care–associated experiences of stigma and discrimination were significantly associated with poor viral suppression in this sample [26]. Many of these factors, attitudes, and behaviors are not well captured in medical records and may change over time. Moreover, changes in health care policies, HIV medications, alcohol and other substance use, and temporal events (such as pandemics) may correlate with changes in individual behaviors, health conditions, or attitudes and beliefs over time. Therefore, ongoing monitoring is needed to assess their continued impact on HIV care outcomes.

The Florida Cohort study wave 3 focused on factors that are not readily assessed via medical records, including alcohol and other substance use, experiences of stigma and discrimination, and mental health symptoms [27,28]. We sought to capture additional measures related to alcohol consumption and potential interventions that address alcohol use disorder. On the basis of feedback from community members and interests of research investigators, we also included new questions related to HIV stigma and discrimination, barriers to substance use and mental health services, interest in emerging options for health care such as telehealth, and challenges related to ART adherence. We also received supplements to study the impacts of the COVID-19 pandemic and HIV care among gender-diverse people with HIV; therefore, additional questionnaire items were added to capture these issues. Items were also added to support topics related to graduate student research. Thus, the overall goals of the Florida Cohort study wave 3 were to identify barriers to and facilitators of HIV viral suppression and other steps of the HIV care continuum, focusing on alcohol use, and monitor changing attitudes and beliefs that influence HIV care outcomes. This paper describes recruitment, study measures, and baseline characteristics of the study population in accordance with the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) and STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklists (see Multimedia Appendices 1 and 2).

The Florida Cohort study is a prospective cohort study that obtained baseline questionnaire data from study participants and linked questionnaire data to statewide HIV surveillance data, including HIV viral load, which will be tracked for at least 5 years after enrollment. The baseline questionnaire included several core measures. Participants had the option to complete additional questionnaires at baseline and follow-up questionnaires at 12 months based on the timing of enrollment and staff resources available at the different recruitment sites. Participants were eligible if they were aged 18 years or above, were living with HIV, and received most of their HIV-related care in Florida. All recruitment settings provided the study in English, whereas some sites offered the study in Spanish, and 1 location offered the study in Haitian Creole. We sought to enroll a sample of people with HIV who are similar to the overall population of people with HIV in Florida in terms of age, gender, race, and ethnicity, including Haitian Creole and Spanish-speaking individuals [5]. The enrollment target was 1000 people with HIV, which included at least 200 (20%) people who met the criteria for at-risk drinking (assessed via the Alcohol Use Disorders Identification Test–Consumption), 50 (5%) people who spoke Haitian Creole, and 150 (15%) people living and receiving care in rural areas [29]. These populations were underrepresented in the previous versions of the Florida Cohort study but are key populations for addressing the HIV epidemic in Florida.

Recruitment began in October 2020 and continued through December 2023 and used a convenience sampling method. The study recruited participants from 9 counties in Florida (major cities noted in parentheses): Alachua (Gainesville), Columbia, Marion (Ocala), Duval (Jacksonville), Hillsborough (Tampa), Brevard (Melbourne), Palm Beach, Polk, and Miami-Dade (Miami; Figure 1).

Most (828/836, 99%) participants were recruited from clinical care or community settings, including public health, academic, and community-based HIV clinics, and case management agencies. In Alachua, Columbia, Marion, Hillsborough, and Polk counties (Figure 1), recruitment occurred primarily within the Florida Department of Health (FDOH) clinics, where potential participants either provided a phone number to be contacted by a research team member or met with a study staff member who was present at the clinic. Recruitment sites in north-central Florida also included an academic infectious disease clinic in Alachua County and mailing flyers to Ryan White HIV/AIDS Program clients in area 3/13. In Duval County, recruitment focused on an HIV case management agency. In Miami-Dade, Brevard, and Palm Beach counties, recruitment focused on community-based Ryan White HIV/AIDS Program–funded clinics. Recruitment strategies in Miami-Dade also included displaying advertisements on buses. Spanish versions of the study were available in Polk and Miami-Dade counties, and the Creole version was offered in Palm Beach County. Initial enrollment procedures were modified to support remote enrollment after the start of the COVID-19 pandemic in 2020, including procedures to confirm eligibility, obtain consent, and enroll people via telephone or videoconferencing. By the end of 2020, most recruitment sites were able to enroll participants in person.

Additional study funding was obtained to provide a more representative sample of gender-diverse participants. To increase enrollment of gender-diverse participants, we provided recruitment information to medical and support agencies supporting transgender and gender nonconforming individuals in Alachua and Miami-Dade counties and worked with a private recruitment resource (TrialFacts) to publish recruitment advertisements on social media in the fall of 2023. Individuals who visited Facebook and Instagram within Florida were shown an advertisement for the study and could voluntarily click on the advertisement to be redirected to a prescreening form managed by TrialFacts. Those who indicated that they were living with HIV shared their contact information with the research team for routine screening and consent procedures, if eligible. Ancillary study protocols enrolled gender-diverse people without HIV, and details of these procedures will be described elsewhere.

All participants completed a brief screening form to confirm that they met the age and HIV status criteria and estimate the number of surveys available to them as a part of the study. These screenings could be performed in person or over the phone. Once eligibility was confirmed, they were scheduled for an appointment to obtain consent. The consent form was accessible through a unique link that was emailed to them, or they could review and sign a paper copy of the consent form. Paper consent forms could be mailed and reviewed over the phone or could be provided in person if the potential participant and recruiter were meeting face-to-face. Those who were emailed the consent form could proceed to the questionnaires through that same link after providing consent.

The study team reviewed the questionnaires developed for the Florida Cohort study wave 2 and identified areas that were underexplored or not explored in previous iterations of the study [19]. Members of the research team then reviewed the relevant literature on the areas identified for further exploration in wave 3 to determine measures to assess these domains. An updated version of the questionnaire was then circulated to the wider research team to identify other areas of exploration, and researchers could suggest measures for inclusion. The initial version of the questionnaire was then reviewed by community consultants and revised accordingly. The surveys were then built in Qualtrics (Qualtrics International Inc) and tested by the research team to verify that the wording and skip logic were correct. Subsequently, the questionnaire was piloted among people with HIV enrolled in another Southern HIV and Alcohol Research Consortium (SHARC) study and revised based on participant feedback. Data from the pilot study were not included in the final dataset. Subsequent questionnaires were reviewed by the research team and community consultants before being added to the study.

Participants were encouraged to complete the study questionnaires online using a Qualtrics-based survey that could be completed either at the recruitment location or at home. The Qualtrics survey showed 1 or 2 questions per page, and the number of pages varied based on participants’ responses. They were allowed to skip questions after dismissing a prompt to go back and answer any blank questions on the page. Participants could also return to the survey and change previous responses. Cookies were enabled so that participants could return to the same page that they were last on if they used the same device and browser.

Participants also had the option to complete questionnaires on paper or have the research assistant read the questions and enter the answers in Qualtrics. The Haitian Creole version of the survey was administered orally by a research assistant who was fluent in Haitian Creole. The order of questions was the same regardless of the administration method.

Alcohol use in the preceding 12 months was assessed using the Alcohol Use Disorders Identification Test–Consumption, a 3-item measure that includes drinking quantity and frequency and heavy drinking episodes. At-risk drinking was defined as a score of 3 or greater for women and 4 or greater for men. Low-risk drinking was defined as any alcohol use below these cutoff scores [29].

Rurality was determined by linking rural-urban commuting area codes to participants’ zip code data [49]. We grouped the codes into 2 categories, so that all rural-related codes were grouped together [50].

All study recruiters received training before the enrollment of participants. In addition to training required by the institutional review boards (IRBs), training included sessions on the enrollment process; a review of the questionnaires; an overview of the use of REDCap (Research Electronic Data Capture; Vanderbilt University) for data collection and storage; secure communication using Microsoft Teams; and a discussion of ethical issues relevant to the study, including the informed consent process. Before actual enrollment, each recruiter was observed conducting a mock enrollment, and additional training was provided to certain staff who might conduct interviewer-administered versions of the questionnaire involving reading the questions out loud to participants.

After obtaining participant consent, we securely linked survey participants to statewide FDOH HIV surveillance data under a formal data use agreement with FDOH. The linkage procedures have been previously described in detail [19]. The surveillance data include dates and results from laboratory tests for CD4 and HIV viral load tests from the Enhanced HIV/AIDS Reporting System (eHARS) managed by FDOH. Data are also linked to the National Death Index, which allows the team to examine survival over time. These results can be used to calculate aspects of the HIV care continuum, including engagement in care and HIV viral suppression. For the subset of participants who were recruited from FDOH clinics, the team also received data from electronic medical records managed by the FDOH, including information on medication prescriptions and medical condition diagnoses.

The survey follow-up ended with administration of the 12-month follow-up questionnaire for participants at the sites where survey follow-ups were offered. Study staff attempted to contact participants who were enrolled at a site where the 12-month follow-up questionnaire was offered at 6 months to update their contact information as needed. For the 12-month follow-up questionnaire, up to 6 contact attempts were made using at least 2 forms of contact given at baseline before the participant was considered lost to follow-up.

All Florida Cohort study wave 3 participants will be followed in eHARS and through FDOH medical records through 2028, allowing for longitudinal data collection for these HIV-related outcomes. FDOH and eHARS records will be updated yearly.

All study data are maintained on secure servers at the University of Florida (UF). All enrollment, consent, and visit log data were entered into REDCap. Consent forms completed on paper were scanned into REDCap. Deidentified paper questionnaires were entered into Qualtrics by research study staff using a double data entry procedure to minimize data entry errors. Participant data were excluded from the final dataset if participants had previously completed the same study questionnaire, when questionnaires had substantial missing information, or when data entry seemed highly infeasible (eg, completing an entire questionnaire in less than 5 minutes when most other participants took 20 or more minutes to complete it; 8 participants). Data managers also reviewed the data to identify likely duplicate records. This included participants who gave different names but the same phone number or who had the same first and last name but with varying middle names or initials. IP addresses were not collected and could not be used to identify duplicates.

Deidentified datasets that do not contain FDOH information are available to investigators who submit a research plan through the SHARC Concepts System [51]. This requires researchers to submit the study rationale, aims and hypotheses, and a detailed research plan, which are reviewed by a team of researchers involved with SHARC and the Florida Cohort study wave 3. The reviewers provide feedback and can request changes and clarifications before approving the research plan and sharing the data. Data that include information from FDOH surveillance data or medical records will require separate data use agreements and approvals from FDOH.

The Florida Cohort study wave 3 allows for both longitudinal and cross-sectional analyses. Cross-sectional analyses include estimating the prevalence of at-risk alcohol use, identifying common motivations for alcohol use and negative outcomes related to alcohol use, and assessing interest in different alcohol-related interventions and treatments. We will also estimate the prevalence of other substance use, unstable housing, symptoms of anxiety and depression, and interactions between these conditions and alcohol use. For the longitudinal analyses, we will assess how alcohol use influences engagement in HIV care and viral suppression over time and how the COVID-19 pandemic impacted alcohol use among people with HIV. Additional analyses will be carried out to meet the goals of the junior faculty and student projects. Examples of these analyses include assessing the impact of psychosocial distress on ART adherence, assessing interest in mobile phone–based mental health interventions, and assessing the impacts of interpersonal violence on HIV care engagement.

The IRB at the UF served as the IRB of record for this multisite study and approved all study activities before implementation (201801680). The IRBs at the FDOH, Florida International University, and the University of South Florida ceded to the UF. All participants provided written informed consent before participation, and individuals had the opportunity to opt out of the study. Identifiable participant data were maintained separately from research data on secure servers at the UF, and limited datasets were prepared by the data management team to ensure participant privacy and confidentiality. Data from the FDOH, including eHARS data, were only accessible through ResVault. Participants received incentive payments via gift cards or cash, depending on the study setting. The amount of compensation varied based on the number of questionnaires completed and the length of those questionnaires (ranging from US $10 to US $25 per questionnaire). Total compensation ranged between US $25 and US $120.

The overall goals of the Florida Cohort study wave 3 were to identify barriers to and facilitators of HIV viral suppression and other steps of the HIV care continuum and monitor changing attitudes and beliefs that influence HIV care outcomes and beliefs among a diverse group of people with HIV in Florida. We also sought to learn more about alcohol consumption patterns and potential treatment interventions, understand the experiences of people with HIV related to the COVID-19 pandemic, and improve our understanding of the experiences of gender diverse individuals regarding HIV care and prevention. The Florida Cohort study also offered opportunities for ancillary data collection for other projects and will provide access to deidentified data for additional analyses.

The sample enrolled in wave 3 was similar to the overall population of people with HIV living in Florida [5]. There were a few exceptions to this. First, cisgender women and transgender people were overrepresented in the Florida Cohort study sample compared to their distribution statewide. Second, Hispanic people with HIV were underrepresented in wave 3. Although almost all questionnaires were available in Spanish, enrollment and data collection in Spanish were limited to 2 recruitment sites (Polk and Miami-Dade counties) that had a Spanish-speaking research assistant. Third, enrollment was higher in north and central Florida, which have lower overall prevalence of Hispanic people, than in south Florida, which has higher rates and total numbers of people with HIV [52]. Finally, the Florida Cohort study did not enroll people with HIV aged below 18 years. We did not enroll people with dementia or living in nursing homes. Therefore, adolescents and persons aged 70 or older were underrepresented.

The Florida Cohort study wave 3 includes participants recruited from a network of clinics and organizations across the state. Unlike the previous wave, which focused largely on recruitment from FDOH clinics, wave 3 included a wider range of community clinics, case management agencies, and recruitment from the public. The Florida Cohort study wave 3 also included several counties that are listed as priority jurisdictions in the EHE initiative [3]. This wave also made a concerted effort to recruit key populations and surpassed the previous wave in recruitment of those who met the criteria for at-risk drinking, individuals who identified as transgender or gender nonconforming, and participants from rural areas, although the target of 150 was not met. Furthermore, we successfully enrolled a subset of participants who spoke Haitian Creole who had not been included in previous waves of the Florida Cohort study but will be helpful in understanding and addressing HIV disparities. People with HIV of Haitian descent are more likely to be diagnosed at a later stage and not achieve key steps in the HIV care continuum [53-55].

A major challenge to enrollment, especially during the first year of the study, was the COVID-19 pandemic and changing mitigation strategies implemented across the state at different times. This study was originally slated to start in March 2020, but the pandemic delayed the initiation of the study until October 2020. This study relied heavily on recruitment through clinics, which shifted to telehealth during the pandemic. Moreover, the management and staff at clinics that maintained in-person visits were understandably hesitant to allow recruiters into clinical spaces. In response, the team shifted to remote recruitment, which was generally less successful than in-person recruitment. Relying on remote recruitment methods where clinical staff served as intermediaries led to lower-than-expected enrollment in the first year of the study for all active sites, and this pattern continued in Miami-Dade County, where the COVID-19 restrictions lasted the longest.

Many lessons were learned during the implementation of this study. To reach a wider range of people with HIV and help the study better integrate into the clinic flow in a variety of settings, this study offered participants many ways to enroll, complete the surveys, and receive compensation. Although this worked well for many participants and sites, it also greatly complicated research staff training and the administration of the study. In addition, some of these methodologies worked better than others. For example, sites that offered consent forms, questionnaires, and payments via US mail frequently encountered problems with items being lost or participants returning either the consent forms or the questionnaires but not both. Once clinics began seeing patients in person again and allowed recruiters back into the clinic, the study team found that obtaining consent from participants in person and following up with a link to the questionnaires worked the best for most participants.

Follow-up rates for the 12-month questionnaire were lower than desired, although follow-up questionnaire data were not needed to answer the key study research questions. To improve follow-up rates, we would benefit from more consistent staffing and strategies to try to keep people more engaged in the study. We had limited face-to-face connections with people, and that could have also impacted follow-up rates. We successfully integrated new questionnaire items and collected information that individual researchers and graduate students could use for their research. At least 3 faculty members are using the Florida Cohort study data to support their research funded by National Institutes of Health individual career development awards, and at least 5 graduate students in epidemiology and nursing are using the data as part of their doctoral dissertations. This study offered opportunities for the research team to engage with clinic staff and HIV service team members across the state. By working with multiple clinic settings and locations, the investigators could learn about issues that were emerging in the community and that could be added to our questionnaires. Community members also helped us update our stigma questionnaire.

In the future, coded data without direct identifiers from this study may be linked to other datasets using a hash ID number, created by a formula that uses direct identifiers but cannot readily identify an individual. Subsequently, all ID numbers, dates, and codes can be removed to prepare a completely deidentified dataset. With IRB review and approval, this procedure has been tested linking data from the Florida Cohort study questionnaires to clinical, dispensing and claims data from the OneFlorida+ Data Trust, which includes medical information for more than 20 million patients on both public and private insurance [56], and the Substance Abuse Mental Health Information System, which includes information on substance use and mental health treatment.

Our team used different recruitment strategies across different recruitment sites; therefore, it is not appropriate to compare any health behaviors or outcomes specifically by recruitment setting. Because we used a convenience sampling method, and non-English questionnaires were not available at all sites, any comparisons across race and ethnicity will need to be done with caution. In addition, several items, including ART preferences and COVID-19 vaccinations, were updated over time, limiting the ability to assess these constructs longitudinally. However, key measures such as symptoms of depression and anxiety and alcohol use remained the same.

In summary, the Florida Cohort study wave 3 enrolled a diverse group of people with HIV from several sites around the state, including both EHE and non-EHE counties. The data can be used to help address key barriers to steps in the HIV care continuum and achieve the goals of the EHE initiative within the state. The Florida Cohort study wave 4 began enrollment in 2024 and will include many of the core measures from wave 3 with additional items collecting formative data to inform alcohol and mental health interventions and enrolling people with HIV from new recruitment locations.