初治涂阴肺结核胸片复读结果与诊断质量多因素分析

Analysis on Re-read Result of Chest X-ray and Diagnosis Quality of Initial Smear- negative Pulmonary Tuberculosis

目的分析在我国结核病防治规划实施中，县级结核病防治机构登记的初治涂阴肺结核患者诊断质量和影响因素。方法按照2006年全国初治涂阴肺结核患者的登记率水平，抽取4个省，每个省选择1个地（市），每个地（市）抽取2个县，采取整群抽样的方法，对各县抽取60例涂阴肺结核患者诊断时的x线胸片和收集患者有关临床资料。各县所有复读胸片经数码摄像后，采用国家级专家个人盲法复读和专家组复读。国家级专家包括结核病防治、结核病临床和放射专家各2位。国家级专家组由结核病防治、结核病临床和放射专家各2位组成。各小组和专家对x线胸片复读认为不符合活动性肺结核患者影像学改变者为过诊。并对所有过诊患者的影响因素（共14个变量）分别进行单因素及多因素logistic回归分析，探讨涂阴肺结核诊断质量与相关的影响因素。结果涂阴肺结核总过诊率20．78％（399／1920），8个县x线胸片诊断的平均过诊率为10．83％（26／240）32．50％（78／240）不等，各专家对胸片复核时，其复核后过诊率在10．21％（49／480）~33．13％（159／480）不等。所有的变量经logistic多因素回归分析，结果显示最终影响过诊的有胸片质量、性别、年龄、病例诊断医生、结核菌素检查、卡介苗接种史及患者来源7个因素。使涂阴肺结核过诊风险增加相关的因素包括：40岁-（OR=2．007，95％CI=1．165-3．535）及60岁以上（OR=1．710，95％CI=1．277-2．940）、女性（OR=I．941，95％CI=1．226-3．102）、门诊医生（OR=2．243，95％CI=I．793-4．27）、结核菌素检查史（OR=1．985，95％CI=1．106-3．564）、卡介苗接种史（OR=2．773，95%7I=1．023-7．515）、胸片质量丙（OR=2．402，95％Cy=1．006-6．995）及胸片质量丁（OR=10．694，95％CI=2．348-62．596）：涂阴肺结核过诊风险减少的相关的因素包括：诊断小组诊断（OR=O．471，95％CI=0．440-0．643）、转诊（OR=0．521，95%CI=O．313-0．865）、集中推荐（OR=O．437，95％C1=0．206-0．927）。结论我国涂阴肺结核过诊率较高，过诊的发生与多种因素有关。

Objective To analyze the diagnosis quality and related factors of initial smear-negative pulmonary tuberculosis （SN-PTB） registered in county dispensaries under National Tuberculosis Control Programme （NTP） in China. Methods Four provinces were selected according to the registration level of national initial SN-PTB registered in 2006, one prefecture and two counties were selected in each province.The X-ray films and relevant clinical data of 60 cases of SN-PTB were sampled through cluster sampling method in each county.Their chests X-ray were photographed by digital camera, which were blindly re-read by national expert individuals and national group of experts. National experts included two experts in TB prevention and control, two in clinical TB and two in radiological TB respectively; National group of experts consisted of two TB prevention and control experts, two clinical and two radiological experts.The over-diagnosis was defined as re-read results by experts or group of experts who considered the chest X-ray did not meet the radiological changes of active TB. Univariate and multivariate logistic regression analysis were made to analyze the factorsaffecting the over-diagnosis, and explore the quality of SN-PTB diagnosis. Results The overall over-diagnosis rate is 20.78% （399/1920） , the average over-diagnosis rate of eight counties ranged from 10.83% （26/240） to 32.50% （78/240）, after the re-reading by experts, the over-diagnosis rate ranged from 10.21% （49/480） to 33.13% （159/480）. Logistic regression analysis showed over-diagnosis was affected by seven factors including chest X-ray quality, gender, age, doctor, PPD examination, BCG vaccination history and source of patient. Risk factors for over-diagnosis included 40 〈 age 〈 60 （0/？=2.007, 95%CI=1.165-3.535） and over 60 years old （OR=l.710, 95%CI=1.277-2.940）, female （OR=1.941, 95%C/=1.226-3.102）, clinic doctor （OR=2.243, 95%CI=1.793-4.27）, PPD examination history （OR=1.985, 95%CI=1.106-3.564）, BCG vaccination history （OR=2.773, 95%CI=1.023- 7.515）, quality of chest film ＂C＂ （OR=2.402, 95%CI=1.006-6.995） and quality of chest film ＂D＂ （OR=10.694, 95%CI=2.348-62.596）; protective factors included diagnosis of diagnostic group （OR=0.471, 95%CI=0.440-0.643）, referral （OR=0.521, 95%CI=0.313-0.865） and cenlralized recommendation （OR=0.437, 95%CI=0.206-0.927）. Conclusion The over-diagnosis rate of smear negative pulmonary tuberculosis is high in China, the occurrence of over-diagnosis is affected by a variety of factors.