Approximating Mechanisms of Suicide Risk to Innovate Interventions for Mid-to-Late-Life Veterans

目前的两年职业发展奖(CDA-1)提案旨在为迈克尔博士 鲁德曼，退伍军人事务部精神病学家，拥有先进研究方法的基础知识，在退伍军人事务部工作 作为精神病学家进行研究，他将推动自杀预防和 干预策略-确定和定位自杀的原因机制 干预。鲁德曼博士将通过追求和完成培训活动来实现这一目标， 来自专家指导团队的支持，并完成了一个旨在弥合差距的研究项目 因果推理研究方法和临床知识之间的关系，为自杀干预提供信息。CDA-1 该项目将最终生成关于自杀风险因果因素的试点信息，并根据 鲁德曼博士提交CDA-2申请的专家启发式。 鲁德曼博士提出的CDA-1项目支持退伍军人管理局的首要临床任务--防止自杀。 此外，这项建议与退伍军人管理局的任务直接一致，即优先进行有助于开发的研究 有针对性的自杀干预，找出为什么某些退伍军人有自杀的风险。尽管退伍军人事务部很强大 预测分析对有自杀风险的退伍军人进行分层，导致自杀的机制很差 明白了。这种知识的缺乏阻碍了有针对性和有效的干预措施的创新 可用。国家自杀研究议程敦促调查人员利用现有数据并确定 可能定义或开发有效自杀干预措施的潜在因果目标。然而，很少有大规模的 现有的数据库将有能力协调和链接到正确的广度和深度 成功识别自杀等罕见(但意义深远)结果的因果目标的信息。因此，这一点 CDA-1拟议的研究利用了艾米·拜尔斯博士(CDA-1的主要导师)的CSR&D功勋奖 项目(CX001119)，该项目独特地形成了一个由500万名50岁及以上的退伍军人组成的纵向队列 包括目前近12,000例自杀死亡人数和人口统计数据，住院病人，门诊病人， 药物、实验室和发病率。CDA-1项目将在拜尔斯博士对晚年自杀的研究基础上进行扩展 风险寻找预后因素并填补该领域的重大空白，因果推断，补充 弗吉尼亚大学自杀风险研究中的预测性分析。此外，关注中老年退伍军人是理想的 因为它在这个研究不足和高度脆弱的群体中提供了有针对性的信息，他们拥有 自杀死亡人数最多(约占所有退伍军人自杀死亡人数的70%)，占 退伍军人群体。 从大量的次要数据中发现候选数据需要一种可以提取因果关系的方法 有效地提供信息，同时对临床实用的可能性进行优先排序。数据驱动的因果方法具有 有可能做到这一点，但前提是这些方法与现有的临床和其他科学方法紧密联系在一起 知识。因此，我们提出了一种方法，其中我们：首先(目标1)利用因果发现技术来 确定50岁及以上退伍军人自杀的初步原因；然后，第二(目标 2)制定一项方案，以获取有关自杀潜在机制的临床专业知识，这将提供 CDA-2应用程序的必要试点信息。这种两阶段的方法确保了专家知识 与计算分析相结合，最大限度地提高预防自杀的临床效用的可能性。对这件事 最后，这个为期2年的CDA-1的目标和培训将为Michael Ruderman博士提交CDA-2做好准备 申请--为作为计算精神病学家的独立研究项目扫清道路，搭桥 实施可操作的改变的方法，降低退伍军人的自杀风险，并赋予他们的提供者权力。

The current 2-year career development award (CDA-1) proposal is designed to prepare Dr. Michael Ruderman, a VA Psychiatrist with foundational knowledge of advanced research methods, for a career in VA conducting research as a psychiatrist scientist who will advance innovation in suicide prevention and intervention strategies—identifying and targeting causal mechanisms for suicide that are amenable for intervention. Dr. Ruderman will accomplish this goal through the pursuit and completion of training activities, support from an expert mentoring team, and completion of a research project aimed at bridging the gap between causal inference research methods and clinical knowledge to inform suicide interventions. The CDA-1 project will ultimately generate pilot information about causal factors for suicide risk with refinement based on expert elicitation for Dr. Ruderman’s submission of a CDA-2 application. Dr. Ruderman’s proposed CDA-1 project supports VA’s top clinical priority—Preventing suicide. Moreover, this proposal is directly aligned with VA’s mandate to prioritize research that can help develop targeted suicide interventions by finding out why certain Veterans are at risk of suicide. Despite VA’s strong predictive analytics for stratifying Veterans at risk for suicide, the mechanisms leading to suicide are poorly understood. This lack of knowledge has impeded the innovation of targeted and effective interventions available. National suicide research agendas urge investigators to leverage existing data and determine potential causal targets that could define or develop effective suicide interventions. However, few large-scale databases exist that would have the capacity to harmonize and link to the right breadth and depth of information to successfully detect causal targets for a rare (yet, profound) outcome like suicide. Thus, this CDA-1 proposed research leverages Dr. Amy Byers’ (primary mentor on the CDA-1) CSR&D Merit award project (CX001119), which uniquely formed a longitudinal cohort of 5 million Veterans aged 50 years and older including, currently, nearly 12,000 suicide deaths and data on demographics, inpatient, outpatient, medications, labs, and morbidity. The CDA-1 project will expand upon Dr. Byers’ research on late-life suicide risk looking at prognostic factors and fill a significant gap in the field, causal inference, complementing predictive analytics in suicide risk research at VA. Furthermore, focusing on mid-to-late-life Veterans is ideal because it supplies targeted information in this understudied and highly vulnerable group, who have the highest number of lives lost to suicide (~70% of all Veteran suicide deaths), as well as make up over 70% of the Veteran population. Discovering candidates from large secondary data requires an approach that can extract causal information efficiently while also prioritizing likelihood of clinical utility. Data-driven causal methods have the potential to do this, but only if such methods are tightly linked with existing clinical and other scientific knowledge. Therefore, we propose an approach where we: first (Aim 1) utilize causal discovery techniques to identify preliminary causal candidates for suicide in Veterans aged 50 years and older; and, then, second (Aim 2) develop a protocol to elicit clinical expertise on potential mechanisms of suicide, which will provide necessary pilot information for a CDA-2 application. Such a biphasic approach ensures expert knowledge is integrated with computational analysis to maximize likelihood of clinical utility for suicide prevention. To this end, the aims and training of this 2-year CDA-1 will prepare Dr. Michael Ruderman to submit a CDA-2 application—clearing a path toward an independent research program as a computational psychiatrist, bridging methods to institute actionable change, reducing suicide risk for Veterans, and empowering their providers.