Krarup Lars-Henrik, Boysen Gudrun, Janjua Huma, Prescott Eva, Truelsen Thomas
Department of Neurology, Copenhagen University Hospital Bispebjerg, Copenhagen, Denmark.
Neuroepidemiology. 2007;28(3):150-4. doi: 10.1159/000102143. Epub 2007 May 3.
Many registers containing routine medical information have been developed for research and surveillance purposes. In epidemiological research assessment of endpoints is often conducted via registers. In the present study we validated stroke and transient ischemic attack (TIA) diagnoses in the Danish National Register of Patients (DNRP).
Subjects from a Danish cohort study, the Copenhagen City Heart Study (n = 19,698), were crosslinked with the DNRP. The following International Classification of Disease 10th revision codes were used to identify possible strokes and TIAs: I60-I69 and G45. Two independent raters reviewed all cases. Positive predictive values of stroke, TIA and stroke subtypes were estimated by dividing the confirmed cases by the total number of cases located in the DNRP. Interrater reliability was tested using kappa statistics.
Of 236 possible cerebrovascular events, 1 in 6 stroke diagnoses did not meet study criteria. The majority of events in the DNRP were registered as unspecified stroke (I64), n = 105 (44%), of which two thirds were diagnosed as ischemic stroke events by the raters. Intracerebral hemorrhage and ischemic stroke had a positive predictive value from 74 to 97%, respectively.
Our results show that the DNRP tends to overestimate the number of cerebrovascular events, while ischemic stroke is underestimated.
许多包含常规医疗信息的登记册已被开发用于研究和监测目的。在流行病学研究中，终点评估通常通过登记册进行。在本研究中，我们在丹麦国家患者登记册（DNRP）中验证了中风和短暂性脑缺血发作（TIA）的诊断。
来自丹麦队列研究哥本哈根市心脏研究（n = 19,698）的受试者与DNRP进行了交叉链接。使用以下国际疾病分类第10版代码来识别可能的中风和TIA：I60 - I69和G45。两名独立的评估者审查了所有病例。通过将确诊病例数除以DNRP中定位的病例总数来估计中风、TIA和中风亚型的阳性预测值。使用kappa统计量测试评估者间的可靠性。
在236例可能的脑血管事件中，每6例中风诊断中有1例不符合研究标准。DNRP中的大多数事件被登记为未指定的中风（I64），n = 105（44%），其中三分之二被评估者诊断为缺血性中风事件。脑出血和缺血性中风的阳性预测值分别为74%至97%。
我们的结果表明，DNRP倾向于高估脑血管事件的数量，而缺血性中风则被低估。
