Improving the efficiency of randomized trials of behavioral interventions for acute and chronic respiratory failure

这个以病人为导向的研究的中期职业研究者奖的目标是提高博士的能力。 Scott Halpern指导学生，居民，博士后研究员和初级教师开发和测试 行为经济干预，以预防和改善急性和慢性呼吸道感染后的结局 失败患有多种形式的急性和慢性呼吸衰竭的患者，如急性呼吸窘迫 急性呼吸窘迫综合征（ARDS）和慢性阻塞性肺疾病（COPD）的死亡风险较高 并在生活质量和功能方面遭受相当大的短期和长期损害。行为 干预措施，包括那些由行为经济学原理提供信息的干预措施， 这些条件和改善的结果。然而，随机临床试验（RCT）产生的知识， 这种干预由于对患者累积，以患者为中心的结果确定， 以及传统分析确定哪些干预措施最适合哪些患者的能力。 通过利用概念模型，研究基础设施和Halpern博士正在进行的5项研究中收集的数据， 或最近完成的随机对照试验，包括40，000名患者，哈尔彭博士和他的导师将开发和测试方法， 为了克服这三个障碍中的每一个， 和慢性呼吸衰竭首先，他们将进行“轻推”的随机试验，以提高入学率， 保留在真实的RCT中。其次，他们将学习和应用最先进的自然语言处理技术， 机器学习技术，以增强从电子健康中提取以患者为中心的结果的能力 记录，从而确定更有效的方法，在大型，务实的随机对照试验的结果确定。第三、 他们将学习和应用因果推理和异质处理检测的现代方法 为了更好地了解行为干预的整体效果，如果 更多的病人坚持他们，以及哪些干预措施最适合哪些病人。 这些科学目标将在一个结构化的培训计划中完成， 博士Halpern指导以患者为导向的初级研究人员的技能。哈尔彭博士已经成功地 指导了二十多名以病人为导向的研究人员。然而，通过使他能够减少他的临床 和行政承诺，这个奖项将增加他承诺指导的时间，并将提高 通过支持影响力和领导力方面的教学培训以及“导师”， 与更多资深导师会面。该奖项还将使他能够开发一个“初级教师访问 教授计划在肺部和重症监护医学，”从而增加了机会， 其他既定调查员的学员与外地其他领导人建立联系，接受外部反馈 和指导，并获得与不太熟悉的观众交谈时的信心。

The goal of this Midcareer Investigator Award in Patient-Oriented Research is to enhance the ability of Dr. Scott Halpern to mentor students, residents, post-doctoral fellows, and junior faculty in developing and testing behavioral economic interventions to prevent and improve outcomes following, acute and chronic respiratory failure. Patients with many forms of acute and chronic respiratory failure, such as acute respiratory distress syndrome (ARDS) and chronic obstructive pulmonary disease (COPD), respectively, are at high risk of death and suffer considerable short- and long-term impairments in quality of life and functioning. Behavioral interventions, including those informed by behavioral economic principles, hold great promise for preventing these conditions and improving outcomes. Yet the knowledge produced by randomized clinical trials (RCTs) of such interventions is limited due to impediments to patient accrual, patient-centered outcomes ascertainment, and the abilities of traditional analyses to determine which interventions work best for which patients. By leveraging the conceptual models, research infrastructures, and data collected in 5 of Dr. Halpern's ongoing or recently completed RCTs including 40,000 patients, Dr. Halpern and his mentees will develop and test ways to surmount each of these three barriers to producing greater knowledge on how to prevent and treat acute and chronic respiratory failure. First, they will conduct randomized trials of “nudges” to improve enrollment and retention in real RCTs. Second, they will learn and apply state-of-the-art natural language processing and machine learning techniques to enhance the ability to extract patient-centered outcomes from electronic health records, thereby identifying more efficient methods for outcome ascertainment in large, pragmatic RCTs. Third, they will learn and apply modern methods in causal inference and the detection of heterogeneous treatment effects to better understand how effective behavioral interventions are overall, how efficacious they would be if more patients adhered to them, and which interventions work best for which patients. These scientific aims will be completed within the context of a structured training program designed to improve Dr. Halpern's skills in mentoring junior patient-oriented researchers. Dr. Halpern has already successfully mentored more than two dozen patient-oriented researchers. However, by enabling him to reduce his clinical and administrative commitments, this award will increase the time he commits to mentoring, and will improve his mentoring skills by supporting didactic training in influence and leadership and “mentor-the-mentor” meetings with more senior mentors. The award will also enable him to develop a “Junior Faculty Visiting Professor Program in Pulmonary and Critical Care Medicine,” thereby enhancing opportunities for his and other established investigators' mentees to network with other leaders in the field, receive external feedback and mentorship, and gain the confidence that comes with speaking to less familiar audiences.