Factors associated with hospitalization, ICU use and death among vulnerable populations diagnosed with COVID-19

项目摘要。截至2020年4月30日，美国有超过100万人被诊断患有 2019冠状病毒病（COVID-19）。COVID-19患者可能会出现各种症状-而 大多数患者症状轻微，有些需要住院治疗，进入重症监护室（ICU）， 可能会死迄今为止，对与COVID-19严重程度相关的风险因素的了解有限。 首先，已发现老年人出现COVID-19严重症状的风险更高， 更有可能住院或死亡。研究表明，一些潜在的条件，如 高血压、糖尿病或肥胖与COVID-19的严重程度相关。然而，它是未知的， 这些合并症在多大程度上解释了COVID-19严重程度的变化，年龄是否 与COVID-19的严重程度独立相关;以及年龄是否以及如何改变 合并症与COVID-19严重程度之间的关系。第二，据报道， 美国人经历了更高的COVID相关住院率，更有可能死于COVID- 19，与白色美国人相比。然而，目前还不清楚是什么导致了这种种族差异- 无论是由于黑人和白人之间健康状况的差异，还是由于 他们居住的社区的特点，或由于一些其他因素，也是相关的 种族。本研究的目的是确定与严重程度相关的个体风险因素， COVID-19（即住院，ICU使用和死亡），特别是老年人，并了解原因 这可能导致COVID-19严重程度的种族差异。为了实现这些目标，我们将每天- 更新的国家退伍军人事务部（VA）数据，其中包含丰富的退伍军人个人信息 被确诊为COVID-19。截至2020年4月30日，近9,000名退伍军人被诊断患有COVID-19， 大约有500人死亡，因此提供了一个大的研究队列。这项研究有两个具体目的：1） 识别与COVID-19相关住院、ICU使用和 死亡率，了解老年在COVID-19严重程度中的作用，并建立COVID-19的预测模型， 19严重程度的机器学习;和2）检查疾病严重程度的种族差异的原因， 退伍军人诊断为COVID-19：这种差异是否以及如何与个人因素有关， 社区特征，特别是社会经济地位。这项研究是创新的，因为它将是第一个 通过使用国家数据，研究多种风险因素对COVID-19严重程度的作用， 详细的个人层面的信息和机器学习算法;这将是第一次检查 原因，包括社会决定因素的作用，种族差异的COVID-19严重程度。这一拟议 研究意义重大，因为它将有助于识别具有最高风险表型的患者，从而提供见解 疾病预防和资源分配。

Project Summary. As of April 30, 2020, over 1 million individuals in the U.S. have been diagnosed with coronavirus disease 2019 (COVID-19). Patients with COVID-19 may develop various symptoms – while the majority of patients have mild symptoms, some require hospitalization, admissions to intensive care unit (ICU), and may die. To date, there is only limited knowledge on risk factors associated with the severity of COVID-19. First, older adults have been found to have higher risks of developing severe symptoms of COVID-19 and are more likely to be hospitalized or die. Studies have suggested that some underlying conditions, such as hypertension, diabetes, or obesity, are associated with the severity of COVID-19. However, it is unknown to what extent these comorbidities explain the variation in the severity of COVID-19, whether older age is independently associated with the severity of COVID-19; and whether and how older age modifies the relationship between comorbidities and the severity of COVID-19. Second, it has been reported that black Americans experienced a higher rate of COVID-related hospitalization and were more likely to die of COVID- 19, compared to white Americans. However, it is unknown what may contribute to such racial difference – whether it is due to the differences in health conditions between blacks and whites, or due to the characteristics of the community where they reside in, or due to some other factors that are also associated with race. The objective of this study is to identify individual risk factors that are associated with the severity of COVID-19 (i.e. hospitalizations, ICU use and death), especially among older adults, and to understand reasons that may contribute to racial differences in COVID-19 severity. To achieve these goals, we will use the daily- updated national Veterans Affairs (VA) data, which contain rich individual-level information on veterans diagnosed with COVID-19. As of April 30, 2020, almost 9,000 veterans have been diagnosed with COVID-19, and about 500 had died, thus providing a large study cohort. This proposed study has two Specific Aims:1) To identify individual risk factors that are associated with COVID-19 related hospitalizations, ICU use and mortality, to understand the role of older age in COVID-19 severity, and to build a predictive model for COVID- 19 severity by machine learning; and 2) To examine reasons for racial differences in illness severity among veterans diagnosed with COVID-19: whether and how such difference is related to individual factors and community characteristics, especially socio-economic status. This study is innovative because it will be the first study to examine the role of multiple risk factors in the severity of COVID-19 by using national data with detailed individual-level information and machine learning algorithm; and it will be the first to examine the reasons, including the role of social determinants, for racial differences in COVID-19 severity. This proposed research is significant as it will help to identify patients with the highest-risk phenotypes, thus providing insights into disease prevention and resource allocation.