摘要
目的对缺血性脑卒中TOAST病因分型进行研究者内部和研究者间的一致性评价。方法采用前瞻性队列研究方法,连续性登记2002年3月-2003年3月间入院的缺血性脑卒中患者,按照TOAST标准进行入院时和出院时病因分型,分析TOAST分型研究者内部和研究者间一致性。结果321例患者纳入试验。(1)研究者内部一致性:辅助检查前后,TOAST分型符合率为64.2%(206/321),检查后有35.8%(115/321)的患者分型发生改变,且各型间相似,K=0.50,中度一致。(2)研究者间一致性:2位神经科医师诊断的20例患者中,16例(80.0%)一致,K=0.72,高度一致。(3)除其他明确病因型外(病例数少),各型敏感性、特异性均较高。结论缺血性脑卒中TOAST病因分型在研究者间一致性较高,研究者内部一致性稍差,在临床使用中不能根据患者入院时的分型决定患者的治疗策略、预防复发的方法和是否纳入以某种分型为主的临床试验。辅助检查后的分型一致性高,可以用来指导临床处理和预防。
Objective To investigate interphysician and intraphysician agreement on the etiological typing of ischemic stroke according to TOAST criteria. Methods Data were collected prospectively from consecutive inpatients with stroke from March 2002 to March 2003. All the patients were divided into 5 major ischemic stroke subtypes based on TOAST criteria before and after completing ancillary diagnostic studies in order to assess the interphysician and intraphysician agreement about TOASY subtypes. Remits 321 patients enrolled in the trial. ( 1 ) Initial clinical impression of typing agreed with final determination in 64.2% (206/321) of patients, the rate was medium and similar to all stroke subtype (K = 0.50). (2)The two physicians agreed in their final diagnosis of subtypes of ischemic stroke in 16 out of 20 patients, the overall agreement was high (80.0%)( K=0.72). (3)The sensitivity and specificity of subtypes of ischemic stroke were all high except for the stroke of other definite etiological subtype. Conclusion Interphysician agreement is higher than that of intraphysician in the diagnosis of subtypes of ischemic stroke according to the TOAST criteria. Early identification of subtype of ischemic stroke couldn't direct acute management, long-term strategies to prevent recurrence and the design of clinical stroke trials targeting or excluding patients with specific subtypes of stroke, but the identified subtypes after completing ancillary diagnostic studies Could guide management and prevention of ischemie stroke.
出处
《疑难病杂志》
CAS
2006年第2期89-92,共4页
Chinese Journal of Difficult and Complicated Cases