Suicide is a major public health problem for youth, yet there is a gap in the research literature on identification of near-term predictors that can improve clinicians' ability to manage risk, prevent suicidal behaviors, and improve treatment outcomes for suicidal youth. The long-term objectives of the current line of research are to identify dynamic near-term predictors of youth suicidal behaviors that are useful to guide real-time clinician decisions to address the aspirational goals of the 2014 Prioritized Research Agenda for Suicide Research. The current study will use data collected during Dialectical Behavior Therapy for Adolescents (DBT-A) to develop models of dynamic near-time predictors of youth suicide. Specifically, the data from diary cards completed by DBT-A clients, medical records, and a DBT-A outcome evaluation database will be combined, and intensive individual level time-series modeling will be used to identify patterns of change useful for predicting youth suicidal behaviors and improving DBT-A. While DBT-A is an evidence-based psychotherapy that is effective at decreasing youth suicidal and self-harming behaviors, data suggests that DBT-A is less effective at eliminating other risk factors for suicide (e.g., depression) and that many youth continue to seek additional mental health treatment after completing DBT-A. Additionally, some youth dropout of DBT-A meaning they may not be getting the care they need despite being at high risk for suicide and poor long-term psychosocial functioning. Thus, there is room for improvement. Identifying patterns by analyzing the diary card data, completion of which is standard in DBT-A, is likely to be useful as DBT-A clinicians already use this data to inform clinical decisions. Thus, this approach has the potential to improve how DBT-A clinicians manage risk as well as make ongoing treatment decisions to improve client outcomes. Despite this potential, diary card data collected over the course of DBT-A has rarely been analyzed. Given the goal of clinical utility and potential next steps (changing self-monitoring procedures, leveraging digital technology, providing data to guide ongoing treatment decisions), under Aim 1 we will establish an advisory board to obtain input from multiple stakeholders on this innovative research approach. The board will meet regularly and include individuals with expertise in suicide research and DBT as well as those with lived experience, suicide survivors and former DBT clients. Aims 2 and 3 of the current study are to identify models of suicidal ideation, emotions, self-harm, and personalized target behaviors that predict suicide attempts and treatment trajectories (e.g., dropout, length of treatment) using intensive time-series modeling. The hypotheses relevant for aims 2 and 3 are based on the theory of emotion regulation that guides DBT-A and research suggesting that the variability, instability, and inertia of suicidal ideation and emotions are particularly strong predictors of suicidal behaviors. The iterative modeling process will facilitate stakeholder input throughout, which will improve the interpretation of results, support the utility of the models for preventing suicide and improving treatment outcomes, and shape next steps in research that will be acceptable and impactful.