COVID-19 Data for the first wave

We collected county-level cumulative COVID-19 confirmed cases and death from Mar 25 to Nov 12, 2020, across the contiguous United States from <i>USAFacts</i> (usafacts.org). We considered Mar 25 to Jun 3 as the “1^st wave”, Jun 4 to Sep 2 as the “2^nd wave”, and Sep 3 to Nov 12 as the “3^rd wave” of COVID-19. For the 2^nd and 3^rd waves, we analyzed the targeted counties in the sunbelt region (including AL, AZ, AR, CA, FL, GA, KS, LA, MS, NV, NM, NC, OK, SC, TX, TN, and UT states) and great plains region (including IA, IL, IN, KS, MI, MO, MN, ND, NE, OH, SD, and WI states), respectively. MIR, as a proxy for survival rate, is calculated by dividing the number of confirmed deaths in each county by the confirmed cases in the same county at the same time-period multiplied by 100. MIR ranges from 0%-100%, 100% indicating the worst situation where all confirmed cases have died. Thirty-eight potential risk factors (covariates), including county-level MR of comorbidities &amp; disorders, demographics &amp; social factors, and environmental factors, were retrieved from the <i>University of Washington Global Health Data Exchange</i> (http://ghdx.healthdata.org/us-data). Comorbidities and disorders include CVD, cardiomyopathy and myocarditis and myocarditis, hypertensive heart disease, peripheral vascular disease, atrial fibrillation, cerebrovascular disease, diabetes, hepatitis, HIV/AIDS, tuberculosis (TB), lower respiratory infection, interstitial lung disease and pulmonary sarcoidosis, asthma, COPD, ischemia, mesothelioma, tracheal cancer, leukemia, pancreatic cancer, rheumatic disease, drug use disorder, and alcohol use disorder. Demographics &amp; social factors include age, female African American%, female white American%, male African American%, male white American%, Asian%, smokers%, unemployed%, income rate, food insecurity, fair/poor health, and uninsured%. Environmental factors include county population density, air quality index (AQI), temperature, and PM. A descriptive table, including all potential risk factors, is provided in Table S1).<br>

我们从USAFacts（usafacts.org）获取了2020年3月25日至11月12日期间美国本土县级累计新型冠状病毒肺炎（COVID-19）确诊病例与死亡病例数据。我们将2020年3月25日至6月3日划定为新冠疫情“第一波”，6月4日至9月2日为“第二波”，9月3日至11月12日为“第三波”。针对第二波与第三波疫情，我们分别对阳光地带（包含AL、AZ、AR、CA、FL、GA、KS、LA、MS、NV、NM、NC、OK、SC、TX、TN、UT共17个州）以及大平原地区（包含IA、IL、IN、KS、MI、MO、MN、ND、NE、OH、SD、WI共12个州）的目标县域开展分析。

我们以感染死亡比（Mortality-to-Infection Ratio, MIR）作为生存率的代理指标，其计算公式为：同一县域同一时间段内的确诊死亡数除以同期确诊病例数，再乘以100。MIR的取值范围为0%~100%，其中100%代表所有确诊病例均死亡的最严重场景。

我们从华盛顿大学全球健康数据交换中心（University of Washington Global Health Data Exchange, GHDx，http://ghdx.healthdata.org/us-data）获取了38项潜在风险因素（协变量）数据，涵盖县级共病与疾病死亡率、人口统计学与社会因素，以及环境因素。其中，共病与疾病类别包括心血管疾病（CVD）、心肌病与心肌炎、高血压性心脏病、外周血管疾病、心房颤动、脑血管疾病、糖尿病、肝炎、人类免疫缺陷病毒/获得性免疫缺陷综合征（HIV/AIDS）、结核病（TB）、下呼吸道感染、间质性肺疾病与肺结节病、哮喘、慢性阻塞性肺疾病（COPD）、缺血性疾病、间皮瘤、气管癌、白血病、胰腺癌、风湿性疾病、药物使用障碍及酒精使用障碍；人口统计学与社会因素类别包括年龄、非裔美国女性占比、白人美国女性占比、非裔美国男性占比、白人美国男性占比、亚裔占比、吸烟者占比、失业人口占比、收入率、粮食不安全率、健康状况一般/较差占比、无医保占比；环境因素类别包括县域人口密度、空气质量指数（Air Quality Index, AQI）、气温与细颗粒物（PM）。

包含所有潜在风险因素的描述性统计表详见附表S1（Table S1）。