Lessons learned using FB to recruit LGBT adults across eastern Africa into online sexual health focus groups

Lessons learned using FB to recruit LGBT adults across eastern Africa into online sexual health focus groups

• 1.
  Lessons learned usingFB to recruit LGBT adults across eastern Africa into online sexual health focus groups MICHELE YBARRA1 MPH PhD, EMILIE CHEN1 BA, TONYA PRESCOTT1 BA Center for Innovative Public Health Research, San Clemente, CA, USA IN T RODUC T ION Because of cultural and legal discrimination enacted towards lesbian, gay, bisexual, transgender (LGBT) and other sexual and gender minority people in eastern Africa, little is known about the sexual decision making and HIV risk of LGBT people. Innovative methods are needed to safely engage LGBT Africans in sexual health research. ME T HODS Between December 2016 and April 2017, we explored various online data collection methodologies to determine the appropriate format for focus groups (FGs) to gather feedback from LGBT adults about their lived experiences. Participants were recruited via Facebook (FB) ads. Ads were targeted to adults 18 years and older; living in Kenya, Rwanda, South Sudan, Tanzania, or Uganda; who had selected English language preferences; and indicated attraction to others of the same sex (e.g., men “interested in men” or “interested in men and women”, but not “interested in women”). People who clicked on the ad were linked to the study website, which explained the FG activity and provided an eligibility screener. Eligibility criteria for participating in the FGs for men interested in men or women interested in women included: being 18 years or older; having the same gender identity as their biological sex; living in Kenya, Rwanda, South Sudan, Tanzania, or Uganda; and one of the following: attraction to others of the same sex; sexual minority identity (e.g., gay, lesbian, bisexual, queer); or past 12- month sexual behavior with someone of the same sex. Eligibility criteria for participating in the transgender FG included the same age and country of residence criteria as above and a different gender identity than their biological sex. Attraction to others of the same sex, sexual minority identity, and past 12-month sexual behavior criteria did not affect eligibility for the transgender FGs. Those who submitted eligible answers were then forwarded to a consent form. Once consented, participants were linked to the online FG activity. RES ULTS CON CLUS ION Although online data collection holds promise in reaching hidden populations, piloting work is necessary to determine the most effective way to ensure privacy and promote participation. The project described is supported by Award Number R21 MH109583 from the National Institute of Mental Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Mental Health. FB ad campaigns for each pilot ran for 29 hours. The first FG format piloted was an online bulletin board, with ads targeting sexual minority men. 176,480 people from across eastern Africa were reached. Close to 12,000 clicked on the ad. Most clicks (59%) were from people in Tanzania; Kenya (22%) and Uganda (13%) were the next most common sources of clicks. Thirty-three people completed a screener and were eligible, and 22 consented to take part in the first pilot FG. Of those who consented, 5 people posted at least one time in the online FG bulletin board. Only 4 of the 16 total threads received at least one post, and none of these four threads discussed sexual health topics. To invigorate response rates for the second ad campaign, we shortened the FG script from 16 to 6 threads and displayed them all on one page rather than across three separate pages. We then ran the second campaign, again targeted to sexual minority men. A similar number of eligible people were identified (n=35), while more consented (n=34) compared to the first pilot. However, only four people made at least one post in the bulletin board. After three more ad campaigns, we had only received posts from 3 more participants. Next, we reduced the navigation between questions and the total number of questions. Because none of the participants who responded logged back in to post at a later time in the bulletin board format, we transitioned to a survey format: The two most important questions were asked of everyone, and the proceeding questions were rotated so that they would have an equal chance of being presented to participants. We also expedited the website flow so that eligible candidates identified by the online screener linked directly into the survey, without needing to create a username as they did to access the bulletin board. The sixth pilot resulted in 52 eligible candidates, 47 of who were males who indicated same sex attraction or a non- heterosexual sexual identity, and 5 of who were transgender. Forty-one of those eligible consented. The first question in the survey received 13 responses, and the second question received 16 responses. Using this format, we then ran a continuous FB ad campaign from April to June, resulting in 417,202 FB clicks. Defining ‘participation’ as at least a coherent sentence to at least 2 of the survey questions, 96 men with same sex attraction or a non-heterosexual sexual identity, 85 women with same sex attraction or a non- heterosexual sexual identity, and 31 transgender adults participated.