论文部分内容阅读
目的基层医疗中有多达50%的尿路感染(UTIs)患儿被忽视。尿液培养是诊断的关键,但收集尿液通常很难。本文旨在通过以下方式对两步骤的临床规则进行内部验证:(1)按症状和体征来选择儿童进行尿液收集;(2)进行症状、体征和试纸测试以指导抗生素治疗。方法从英格兰和威尔士的基层医疗机构中选取5岁以下急性发病儿童233例。指数测试包括父母反映的症状、临床医生报告的体征、尿试纸结果以及关于尿路感染可能性的临床医生选择。以清洁尿样中的微生物培养结果作为确认尿路感染的参考标准。计算灵敏度、特异度、和基于系数(分级严重性)和分数(二分)的症状/体征的Logistic回归模型的受试者工作特征曲线下面积(AUROC),然后采用引导指令内部验证AUROC。结果 3 036例儿童提供了尿样,获得2 740份(90%)培养结果,其中60份(2.2%)呈阳性。临床诊断灵敏度为46.6%,AUROC为0.77。基于系数和分数的经验证模型的AUROC分别为0.87和0.86,在加入亚硝酸盐试纸、白细胞和血液后其分别增加至0.90和0.90。结论根据症状和体征的临床规则优于临床诊断,在识别儿童进行无创尿液取样方面表现良好。试纸结果为经验性抗生素治疗提供了进一步的诊断价值。
As many as 50% of urinary tract infections (UTIs) in primary care are neglected. Urine culture is the key to diagnosis, but collecting urine is often difficult. This article aims to internally validate a two-step clinical rule by: (1) selecting children for symptom and signs of urine collection; (2) conducting symptoms, signs, and test strip tests to guide antibiotic therapy. METHODS: 233 acute onset children under 5 years of age were selected from primary care facilities in England and Wales. Index tests include symptoms reported by parents, clinicians reported signs, urinalysis results, and clinician options on the likelihood of urinary tract infection. Microbial culture in urine samples was used as a reference standard to confirm urinary tract infection. The area under the receiver operating characteristic curve (AUROC) of the logistic regression model based on the sensitivity (sensitivity), specificity, and based on the coefficient (severity of grade) and the score (dichotomous) symptoms / signs were calculated and then AUROCs were internally verified using the lead instructions. Results In 3 036 children, urine samples were obtained and 2 740 (90%) of the cultures were obtained, of which 60 (2.2%) were positive. Clinical diagnostic sensitivity was 46.6% and AUROC was 0.77. The AUROCs based on the coefficients and scores of the validated models were 0.87 and 0.86, respectively, which increased to 0.90 and 0.90, respectively, after addition of nitrite test strips, leukocytes and blood. Conclusions Clinical rules based on symptoms and signs are superior to clinical diagnosis and performed well in identifying non-invasive urine sampling in children. Test strip results provide further diagnostic value for empirical antibiotic therapy.