USING RESPONDENT DRIVEN SAMPLING TO ESTIMATE HOMICIDE AND MOTOR VEHICLE CRASH RISK AMONG ADOLESCENTS AND YOUNG ADULTS IN A LATINO COMMUNITY

Purpose: Young people at highest risk of injury are often excluded when school or household based sampling methods are used. Respondent Driven Sampling (RDS), a variant of chain referral sampling, has been used to recruit probability samples of populations who are hard to reach or recruit into studies. In this study, RDS was used to ensure the inclusion of youth who do not attend school regularly or lack residential stability. RDS analysis estimated the proportion of young people in a Latino community who are at risk for homicide and being in a motor vehicle crash (MVC). Methods: Four initial participants ("seeds") were selected based on desired inclusion characteristics. "Seeds" completed questionnaires and were given coupons to recruit up to three additional participants from their social networks. Seed-recruited participants then completed questionnaires and became recruiters for the next wave. The process continued for 30 days. Questionnaires were self-administered at a local public library in either Spanish or English on a computer tablet using REDCap. A Certificate of Confidentiality was obtained prior to data collection. Adolescents who lived within the defined geographic area; were between the ages of 14-21; and were "seeds" or recruited by prior study participants were eligible to participate in the study. The sampling frame was constructed based on questions that each participant answered about their network size and relationship to their recruiter. To obtain unbiased population estimates, all analyses were weighted to accommodate the RDS design. Weighted means and 95% confidence intervals were calculated using Stata/SE (version 15). Results: With four "seeds" and a maximum of 10 waves, 129 participants were recruited. 88% were Latino, 57% male, 54% 14-18 years old, and 46% 18-21 years old. All variables reached equilibrium by four waves. RDS population estimates (95% CI) indicate that 34.1% (26.4-42.7%) of youth in the area do not regularly attend school and 9.8% (6.1-15.1%) are unstably housed. 48.4% (39.8-57.1%) recently used alcohol, 47.2% (38.5-56.1%) have significant depression symptoms, and 20.7% (13.9-29.8%) binge drink. Based on responses to a Violence Perpetration and Injury Scale, 29.4% (22.1-38.1%) are at risk for homicide. 16.7% (11.4-23.7%) got into a serious physical fight, 12.2% (7.7-18.7%) took part in a group fight, and 9.6% (5.9-15.3%) hurt someone badly enough that they needed bandages or care from a doctor. The most prevalent