Design, Analysis, and Optimization of Equitable and Value-based Baseline Testing Policies for Sports-Related Concussion

项目摘要/摘要 该项目的目标是设计和优化一种公平的、基于价值的基线测试方法 综合机器学习和系统科学的方法，对运动相关的脑震荡进行识别。脑震荡， 创伤性脑损伤是最常见的类型之一，每年困扰着超过360万人，是一种 重大公共卫生问题。及时有效的脑震荡治疗被认为是 减轻伤害的短期和长期后果。基线测试是一种广泛使用的实践 这为医生和运动教练提供了一个参考点，表明一个人的“正常”表现 跨越了几个脑震荡特定的功能区。基线测试也是一个资源密集型过程， 需要在耗时的多维脑震荡管理方面具有特定的专业知识 评估小组；尽管如此，基线测试被认为是伤害管理过程的关键 适用于脑震荡风险较高的人，包括学生运动员和军事人员。尽管广泛存在 在使用基线测试时，缺乏关于谁应该优先进行基线测试的循证指导 在资源有限的环境中进行测试。 多学科研究团队的目标是通过综合机器学习和 系统科学方法和来自震荡评估、研究和教育联盟的数据- 关于运动相关脑震荡的最大的多站点数据集之一。具体地说，该项目旨在 首先使用可解释的机器学习方法和统计建模来估计诊断效用 在不同类型的学生运动员队列中进行基线测试。接下来，该项目旨在设计一个决策- 可以优化分配基线测试的分析模型。该模型将考虑：(1)个性化 评估基线测试的诊断效用，(2)运动相关脑震荡的个体化风险，(3) 特定机构的资源限制，以及(4)基线测试分配中的公平考虑。这 研究可以改变临床医生、运动训练师和其他训练有素的医务人员接近基线的方式 测试和脑震荡诊断对那些可能在现有的发展中代表不足的人 临床指南，导致更及时和准确的脑震荡诊断。此外，节省的资源 通过有效地分配基线，可以将测试重新分配到其他有价值的任务上 专业医务人员，包括脑震荡护理连续体的其他任务，中暑疾病 预防，新冠肺炎筛查。

Project Summary/Abstract This goal of this project is to design and optimize an equitable and value-based approach to baseline testing for sports-related concussion by synthesizing machine learning and systems science methods. Concussion, one of the most common types of traumatic brain injury, afflicts upwards of 3.6 million people annually and is a major public health issue. Timely and effective concussion management is considered a major factor in mitigating both short-term and long-term consequences of the injury. Baseline testing is a widely used practice that provides physicians and athletic trainers a reference point indicating someone’s “normal” performance across several concussion-specific functional domains. Baseline testing is also a resource-intensive process, requiring specific expertise in the time-consuming administration of a multi-dimensional concussion assessment battery; nevertheless, baseline testing is considered essential to the injury management process for those at elevated risk of concussion, including student-athletes and military personnel. Despite widespread use of baseline testing, there is a lack of evidence-based guidance on who should be prioritized for baseline testing in resource-limited environments. The multidisciplinary research team aims to address this knowledge gap by synthesizing machine learning and systems science methods with data from the Concussion Assessment, Research, and Education Consortium – one of the largest multi-site datasets available on sports-related concussion. Specifically, the project aims to first use interpretable machine learning methods and statistical modeling to estimate the diagnostic utility of baseline testing in a heterogeneous cohort of student-athletes. Next, the project aims to design a decision- analytic model that can optimally allocate baseline tests. This model will take into account: (1) personalized estimates for the diagnostic utility of baseline tests, (2) individualized risk for sport-related concussion, (3) resource constraints at a given institution, and (4) equity considerations in the allocation of baseline tests. This research can transform how clinicians, athletic trainers, and other trained medical staff approach baseline testing and concussion diagnosis for those who may be under-represented in the development of existing clinical guidelines, leading to more timely and accurate diagnosis of concussion. Moreover, resources saved through an efficient allocation of baseline tests can be reallocated to other valuable tasks performed by specialized medical personnel, including other tasks along the concussion care continuum, heat illnesses prevention, and COVID-19 screening.