新型冠状病毒肺炎临床诊断评分体系的建立及诊断效能评估

Establishment and preliminary evaluation of clinical diagnostic scoring system for COVID-19

目的建立新型冠状病毒肺炎(COVID-19)临床诊断评分体系,并对其诊断价值进行初步评估。方法收集16例COVID-19确诊患者和62例COVID-19疑似患者的年龄、性别、基础疾病、临床症状、流行病学史、合并症、CD4+T细胞数、CD8+T细胞数、白细胞、淋巴细胞绝对计数、血小板数、入院CT影像学表现、发病时间、肝肾功能、肌酶等资料,选择COVID-19确诊患者与疑似患者间有统计学差异的变量进行Logistic回归分析,筛选出对COVID-19诊断有独立影响的变量,根据各变量的β回归系数设立相应分值,建立COVID-19的临床诊断评分体系。用ROC评价COVID-19临床诊断评分的诊断价值。结果成功建立了由临床症状、流行病学史、肺部CT评分、淋巴细胞计数小于1.0×109/L组成的COVID-19临床诊断评分体系:临床症状(1分)、流行病学史(70分)、肺部CT评分(6分)、淋巴细胞计数小于1.0×109/L(2分)。当COVID-19临床诊断评分ROC下面积为0.889(95%CI:0.816~0.960,P<0.05)时,临床诊断评分诊断COVID-19的最佳诊断界值为81分,灵敏度、特异度分别为100%、77.4%。结论成功建立了由4个变量(临床症状、流行病学史、肺部CT评分、淋巴细胞计数小于1.0×109/L)组成的COVID-19临床诊断评分体系,此评分体系对COVID-19有较好的诊断效能。

Objective To develop a clinical diagnostic scoring system for diagnosis of coronavirus disease 2019(COVID-19)and to testify the application value.Methods Data of 16 confirmed COVID-19 patients and 62 suspected patients with COVID-19 were collected,including age,sex,basic diseases,clinical symptoms,epidemiological history,complications,number of CD4+T cells,number of CD8+T cells,leukocytes,lymphocyte absolute count,platelet number,CT imaging manifestations on admission,onset time,liver and kidney functions,muscle enzyme,etc.Logistic regression analysis was used to select the variables with statistical difference between the confirmed patients and suspected patients of COVID-19,and the variables with independent influence on the diagnosis of COVID-19 were selected.The corresponding score was set up according to theβregression coefficient of each variable,and the clinical diagnosis scoring system of COVID-19 was established.The receiver operating characteristic(ROC)curve was used to evaluate the diagnostic value of the COVID-19 clinical diagnosis score.Results The COVID-19 clinical clinical diagnostic scoring system consisting of clinical symptoms,epidemiological history,lung CT score,and lymphocyte count less than 1.0×109/L,was successfully established:clinical symptoms(1 point),epidemiological history(70 points),lung CT score(6 points),and lymphocyte count less than 1.0×109/L(2 points).When the area under ROC of COVID-19 was 0.889(95%CI:0.816-0.960,P<0.05),the best diagnostic threshold of COVID-19 was 81,with sensitivity and specificity of 100%and 77.4%,respectively.Conclusion The COVID-19 clinical diagnostic scoring system is successfully established,which was composed of 4 variables(clinical symptoms,epidemiological history,lung CT Score,and lymphocyte count less than 1.0×109/L),and has good diagnostic efficiency.