Sexual and gender minority (SGM) persons report twice the rate of suicidal ideation and are at increased risk of suicidal behaviors compared to heterosexual and cisgender individuals. Suicide disparities are particularly high among SGM youth. Among lesbian, gay, and bisexual youth, over 45% reported seriously considering suicide, and nearly 35% of transgender and gender diverse reported past-year suicidal ideation, double the rate in cisgender youth. In addition to sexual orientation and gender identity, other factors (e.g., race, ethnicity, income, substance use) likely influence suicide disparities, but little research has used data-driven approaches examined the ways in which the co-occurrence or “intersection” of these factors impact suicidality among SGM youth. Novel data-mining techniques that organize data into different subgroups by identifying the most critical factors and how they interact to influence the outcome may help reveal subgroups of SGM adolescents and young adults that are at heightened risk for suicidality. Therefore, the aims of the proposed study are to: (1) apply a data-mining technique, known as the conditional inference tree, to existing electronic medical record (EMR) data at Fenway Health, a community health center that caters to LGBTQIA+ persons, to identify subgroups of SGM youth and young adults with intersecting demographic and psychosocial factors associated with increased suicidal ideation and history of suicide attempts; and (2) qualitatively explore factors that are perceived to increase risk (probing differences by developmental stage), barriers to accessing and engaging with behavioral health resources/services, and preferences for an enhanced suicide screening and resource provision intervention to be integrated into primary care, with both patients at high risk for suicidal ideation (identified in Aim 1) and Fenway Health primary care providers. The Aim 1 study sample will be patients who (1) are aged 13-29 years; (2) identify as gay, lesbian, bisexual and/or have a gender identity that differs from their sex assigned at birth; and (3) have completed a Patient Health Questionnaire-9 (PHQ-9) in the past five years, as the final item of the PHQ-9 will the outcome measure for the analyses. In an exploratory analysis, the presence of words that indicate a recent suicide attempt within the open text fields of patients charts will also be used as an outcome measure. Based on our preliminary work, we hypothesize that the intersection of distress, age, discrimination, and income may be associated with increased suicidal ideation. After subgroups with intersecting risk for suicidal ideation are identified, patients with these profiles will be invited to participate in the Aim 2 focus groups (n~32), which will be stratified by age (i.e., 13-17, 18-29). A focus group will also be conducted with Fenway Health providers (n~8). These data will enable us to build an intervention that strengthens automated risk detection via systems-level changes to the EMR, enhances screening when risk has been detected, and offers resources that are developmentally appropriate. System-level risk detection that takes intersectional factors into account will ultimately reduce suicidality in SGM youth and young adults.
