Suicide is the second leading cause of adolescent mortality in the US. Rates of adolescent suicide have increased in recent years, along with the advent and growth of social media use (SMU) in this population. SMU has been linked with sleep disruption, depressed mood, and suicidal ideation (SI), all of which are precursors of suicidal behavior. Existing research on the impact of SMU on sleep, mood, and suicidality is limited by cross- sectional designs and self-reported measures of SMU and sleep. To advance the field, prospective designs are needed that include objective measures and intensive monitoring of high-risk samples to rigorously examine the temporal relationships between SMU and suicide risk, defined as depressed mood and SI. This K01 proposal uses precisely such a design to test a conceptual model in which SMU predicts sleep disruption, which, in turn, contributes to depressed mood and SI among adolescents. The proposed study will harness smartphone technology to assess SMU using real-time passive data capture, actigraphy, and ecological momentary assessment (EMA) in high-risk adolescents. These methods, combined with in-person study visits, will examine the temporal and unique within-person associations between SMU and sleep disruption (Aim 1), SMU and suicide risk (Aim 2), and test whether sleep disruption mediates the relationship between SMU and suicide risk (Aim 3). Data-driven approaches of supervised machine learning will leverage the high-dimensional, intensively- monitored data to identify key SMU features predictive of suicide risk in adolescents (Exploratory Aim). The candidate will build on her strong foundation in adolescent depression and sleep research by acquiring new conceptual and methodological training in: 1) adolescent suicide risk, 2) technology and mental health, including social media use and mobile technology for real-time assessment; and 3) advanced computational skills. The candidate has assembled an interdisciplinary mentorship team to achieve her training goals within the exceptional environment of the Department of Psychiatry at the University of Pittsburgh. Mentors (Brent and Moreno) are experts in adolescent suicide and social media, respectively, and both have had success in mentoring early career scientists. The candidate?s consultants have extensive expertise in adolescent development and EMA (Silk), adolescent sleep and actigraphy (Franzen), smartphone sensing (Ferreira) applied to clinical health (Low), and advanced statistical and machine learning approaches (Wallace). The proposed training and research will inform future R01 studies that use these innovative methods to identify and modify clinically-actionable risk factors, including SMU and sleep disruption, to attenuate near-term risk for adolescent suicide. This program of research has the potential to yield high-impact results that inform the development of suicide prevention programs that are personalized, scalable, and delivered in real time. The proposed research along with this training plan will uniquely position the candidate to be an independent investigator and leading scholar in the important public health problem of adolescent suicide.
Adolescent suicide and suicidality have increased in recent years, which parallels the advent and rapid growth of social media use in adolescents? daily lives. The proposed study will be the first in a broader program of research that leverages smartphone technology to examine how social media use impacts risk for the development and worsening of suicidal ideation and behavior among adolescents. This research program aims to identify modifiable risk factors, including social media use and sleep disruption, for adolescent suicide risk, and ultimately, contribute to the development of interventions that are personalized, scalable, and delivered in real time to prevent suicide.