Computerized Adaptive Testing of Suicidality in LGBT Young Adults

Despite strong indications of elevated risk of suicidal behavior in lesbian, gay, bisexual, and transgender people, limited attention has been given to research, interventions or suicide prevention programs targeting these populations. Recent studies using Denmark's extensive registries of vital statistics and other sociodemographic data noted that same-sex registered domestic partners were 3–4 times more likely than heterosexual married persons to die by suicide. Population-based surveys of U.S. adolescents that have included questions about sexual orientation have consistently found rates of reported suicide attempts to be 2-7 times higher in high school students who identify as LGB. However, standard measures of psychological distress and suicidal ideation have poor predictive power for Major Depression and suicidal behavior among LGBT youth, suggesting the need to refine and validate mental health screeners widely used in the general population before they are adopted with LGBT youth. Recent advances in computerized adaptive testing (CAT) based on multidimensional item response theory (MIRT) have revolutionized patient reported outcomes (PRO). Using this approach, we have created multidimensional CATs for depression, anxiety, hypomania/mania spectrum and suicidality that increase precision and decrease patient and clinician burden relative to traditional mental health testing and clinical interviews. MIRT-based CAT is highly relevant for assessing suicide risk because of the inherent multi-dimensionality associated with suicide. Such an approach allows us for the first time to bridge the gap between depression, anxiety, and suicidality in a psychometrically rigorous way and to identify the particular symptoms of depression and anxiety that are the precedents of suicidal ideation and behavior. To this end, we have developed and validated the CAT-Suicide Scale (CAT-SS) using the data from our original NIMH funded study (R01MH66302) and validated (against clinical interviews) the measure in emergency departments at the University of Chicago and University of Massachusetts. There is no guarantee, however, that all of the items that provide excellent discrimination of high and low suicide risk in adult ED and psychiatric clinic settings will also do so for LGBT youth. This is particularly true because the CAT-SS assesses multiple sub-domains of suicide risk including suicidality, depression and anxiety. The proposed study will answer this question and lead to a modification of the CAT-SS for LGBT youth as required. This study will also add to the existing literature on the CAT-SS by validating the scale as a predictor of suicide and suicidal behavior in the following 6 months, in a high risk population.