Assessment on validity,Reliability and Simplicity of Stroke Scales and Design ora New Scale
Background:A lot of rating systems have been devcloped, but none of them was univcrsally accepted in clinical stroke trials. Objective: The aim of this study was to authenticatc the four stroke scales, which has becn generally used, by comparing their validity, reliaoiiity and simplicity. We also designed a new scalc and assessed it. Methods:l.Predictive validity was investigated in 103 patients with acute stroke within 72 hours after onset. The neurological deficit was mcasurcd by a observer with CNS, NIHSS, ESS and MESSS. The correlative analysis was performed between their scores and thc Barthel Index(BI) at the end of the second month and the third month after onset. The logistic regression was used to examine whether the four scales were predictive for mortality. 2.To establish the reliabiltiy and simplicity of scales, twenty patients with acute stroke were investigated. The neurological deficit was scored independently by two raters using four scales at an interval of two hours. Time needed to complete each scale was recorded. The degree of interrater agreement for each items on the four scales was determined by calculation of the kappa statstic and wcighted k statistic. 3. Items that can typically prcdict functionaloutcome were identified by stepwisc regression and correlativc analysis between BI and 43 items on the four scalcs. Thc ncw scale was composed of these items with their weight givcn a new. 4. The predictive validity of the ncw scale was determined by correlative, stcpwisc regression and logistic regression analysis with its original data. The reliability and simplicity of the new scale were assessed in another 15 patients with acute stroke. Results: The correlative values between BI and each scale of MESSS, CNS and NIHSS or ESS werc -0.804-0.833, 0.694-0,812 and 0.72 1-0.827 individuallv. The values of MESSS wcrc sigificantly higher than that of CNS(P<0.05), but there were no significant differences among those of NIHSS, ESS and MESSS(P>0.05). It was proved by logistic regression that all of the stroke scales were suficiently predictive for mortality. The interobserver agreement was almost prefect (K>0.80) for 5 items (62.5%) on CNS and for I item (12.5%) on MESS.It was substantialfor 3 items(37.5%) on CNS and 7 items (87.5%) on MESSS. There Were 5 items with K,0.60 on NIHSS ot ESS, and none on both CNS and MESSS. Thus,the interobscrver reliablity for MESSS was worse than that for CNS<and better than that for NIHSS ot ESS, The mean time±Sdfor completion of total scales was 4.46±0.98 mintutes for CNS 7.70± 1.88 mintutes for MESS,12.55 ± 3.59 minutes for NIHSS and 11 .05 ± 2.58 mintues for ESS. The time for MESSS was longer than that for CNS but shorter than that for N1HSS or ESS. Multiple stepwisc gression showed that level of consiousness, comprchcnsion, language, extraocular movement, visual fields, facial palsy, motor arm proximal, motor hand and motor leg wcre typical predictors of BI. Correlative analysis vith their original data showed that the corelative values between BI and CDSS were significiently higher than that for the other scales. Logistic regression showed that CDSS can predict mortality sufficiently. Interobscrver agreement for CDSS was similar to that for CNS, but better than that for MESSS. The time needed to complete CDSS was as long as that for CNS, but shorter than that for MESSS. Conclusion: 1.CNS has high reliability and simpliticity, butlow validity. MESS has high validity, but lower rclability and simplicity than CNS. Both NIHSS and ESS have low validity, reliability and simpliaty. Thus,all of the four general used scales have their own shortages.It is suggested that MESSS be used in stroke trials in the condition fo that there are no ideal scales to be choiced 2.The items, which can properly be used to design a new scale ,are level of consciousness, comprehension, language, extraocular movement, visula fields, facial palsy. Motor arm proximal, motor hand and motor leg. 3.The reliatility and simplicity of CDSS are similar to that of CNS and better than that of MESSS.The validity of CDSS is higher than that of both MESSS and CNS. Thus. CDSS is a valid. Reliable and simple new stroke scale. It is worthy to be used and farther assessed in clinical trial.
logistic regression、neurological deficit、correlative analysis、acute stroke、regression analysis、clinical trial、Barthel Index、used in
8
R74;R3
2005-07-07(万方平台首次上网日期,不代表论文的发表时间)
共1页
69