Diagnostic efficacy of clinical decision tree and integral model in tuberculous pleurisy

Objective To explore the diagnostic efficacy of clinical decision tree and integral model in tuberculous pleurisy. Methods A prospective randomized controlled study was conducted in patients with pleural effusion admitted to Wuhan Pulmonary Hospital from January 1 to June 30, 2019. Patients with pleural effusion were randomly enrolled into the clinical decision tree group and the integral model group. Two kinds of comprehensive diagnosis models were used for diagnosis, and the diagnostic efficacy of the two methods was compared. Results The clinical decision tree group included 127 patients, 90 male, 37 female, 18–88 years old, with an average age of (49.33±19.63) years. Finally, 108 cases of tuberculous pleurisy and 19 cases of non tuberculous pleurisy were diagnosed. In the integral model group, 120 patients were included, including 89 males and 31 females, 13-84 years old, with an average age of (46.41±20.01) years. 94 cases of tuberculous pleurisy and 26 cases of non tuberculous pleurisy were finally diagnosed. The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy of clinical decision tree method were 57.4%, 78.9%, 93.9%, 24.6% and 60.6% respectively; and of integral model method were 75.5%, 96.2%, 98.6%, 52.1% and 80.0% respectively. Conclusion The sensitivity, specificity and accuracy of the integral model method in the diagnosis of tuberculous pleurisy were higher than that of the clinical decision tree method. The method is simple and can be widely used in the clinical diagnosis of tuberculous pleurisy. 摘要：目的 探讨临床决策树法与积分模型法对结核性胸膜炎的诊断效能。 方法 采用前瞻性随机对照研究方 法, 以2019年1月1日一2019年6月30日武汉市肺科医院收治的胸腔积液患者为研究对象。符合人组标准的胸腔积液 患者随机人组临床决策树法组和积分模型法组, 分别采用两种综合诊断模型进行诊断, 对比两组方法的诊断效能。 结果 临床决策树组共纳人127例患者, 其中男性90例, 女性37例, 年龄18~88岁, 平均(49.33±19.63)岁, 最后诊断结 核性胸膜炎108例, 非结核性胸膜炎19例；积分模型组共纳人120例患者, 其中男性89例, 女性31例, 年龄13 ~ 84岁, 平均(46.41±20.01)岁 , 最后诊断结核性胸膜炎94例 , 非结核性胸膜炎26例。临床决策树法诊断结核性胸膜炎的敏感 度57.4%, 特异度78.9%, 阳性预测值93.9%, 阴性预测值24.6%, 诊断准确率60.6%;积分模型法诊断结核性胸膜炎的敏 感度75.5%, 特异度96.2%, 阳性预测值98.6%, 阴性预测值52.1%, 诊断准确率80.0%。 结论 积分模型法诊断结核性胸 膜炎敏感度、特异度及诊断准确率均髙于临床决策树法, 方法简单, 可推广应用于结核性胸膜炎的临床诊断