北京市出院病例监测系统中急性卒中病例诊断质量评价

目的 评价北京市出院病例监测系统卒中的诊断质量，为利用数据库开展卒中发病流行病学监测提
供数据质量评价证据。
方法 采用按病历号完全随机抽样的方法在出院病例数据库中抽取15家医院1433份急性卒中病例
（2007年704份，2010年729份）。以专家组查阅原始病案资料重新诊断结果为金标准，评价诊断的一
致性。
结果 第一诊断与专家核查结果符合率为73.1%（95%可信区间：70.7%～75.4%）。缺血性脑血管病
的诊断符合率总体低于出血性脑血管病。第一诊断与专家核查结果不符病例的正确诊断主要是诊
断证据不足、头晕、既往脑梗死本次非急性发作等。
结论 出院病例数据库卒中诊断总体具有较高的阳性预测值，在利用出院病例数据库数据进行人
群卒中发病监测时应进行适当的纠正。

Validation of Stroke Diagnosis in Beijing Hospital Discharge Register Database

Objective To estimate the validation of stroke diagnosis in Beijing Hospital Discharge Register Database, and to provide evidence of data quality for epidemiology monitoring of incidence of stroke using this administrative database. Methods One thousand four hundred and thirty-three stroke patients （704 reported in 2007, 729 reported in 2010） in 15 hospitals in Beijing were randomly sampled from Beijing Hospital Discharge Register Database. The agreements of diagnosis were assessed using the results of expert re-diagnosis after reviewing the medical records in detail as the gold standard. 〈br〉 Results The agreement rate of ifrst diagnosis recorded in the database and result of expert re-diagnosis was 73.1% （95% confidence interval：70.7%~75.4%）. The agreement rate of ischemic cerebrovascular disease was lower than that of hemorrhagic cerebrovascular disease. The main correct diagnoses of diagnosis inconsistency cases were lack of evidence for diagnosis, dizziness, established ischemic stroke without attack before this hospitalization. Conclusion The positive predictive value of the hospital discharge register database is fairly good. Certain adjustments are required when the administrative register database is used for the monitoring of stroke incidence.