涂阴肺结核综合诊断数学模型建立及初步评价

Construction of mathematical diagnosis model of sputum smear-negative pulmonary tuberculosis

目的建立涂阴肺结核综合诊断的数学模型,以提高临床诊断的规范性及准确性。方法采集肺结核病与非肺结核病两类病例标本的主要临床观察项目数据,并应用logistic回归模型以及判别分析等统计方法研究。结果皮肤PPD试验、肺部病灶空洞、咯血症状、血WBC计数、血清白蛋白等仍是重要的鉴诊项目;本研究所建立的涂阴肺结核的诊断数学模型,判别准确率达80%以上,个体测得Z值临界值拟定为0,Z值为正值则倾向于涂阴肺结核病诊断,为临床医师在涂阴肺结核病的诊断上提供客观科学的参考工具。

Objective To construct mathematical diagnosis model of sputum smear - negative pulmonary tubercu- losis. Methods The specimens from patients with sputum smear - negative pulmonary tuberculosis or non - pulmonary tuberculosis were collected for analysis. Logistic regression model and diseriminant analysis were applied for analysis. Re- sults Skin PPD test, lung cavity, hemoptysis, WBC count, serum albumin were of diagnoistie significance. Sputum smear - negative pulmonary tuberculosis diagnostic mathematical model was established with accuracy over 80%. Z value over 0 was correlated with sputum smear- negative pulmonary tuberculosis.