[Clinical application of the Chinese version of Cornell assessment of pediatric delirium: a pilot study].

Objective: To evaluate the clinical application of Chinese version of Cornell assessment of pediatric delirium (CAPD) scale in children by a pilot study. Methods: A prospective observational study. From June 2017 to December 2017, the original CAPD screen was translated into Chinese and debugged cross-culturally according to the guidelines, which was further applied in the Department of Critical Care Medicine, Children's Hospital of Chongqing Medical University. The clinical data of 250 patients who were evaluated with the Chinese version of CAPD scale and Richmond agitation-sedation scale were extracted and analyzed with chi-square test or Mann-Whitney U test. Cronbach's α coefficient,Split-half coefficient and intra-group correlation coefficient were used to evaluate the reliability, while exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were used to evaluate the structural validity of the scale. The result was compared with the gold standard diagnostic results to evaluate the diagnostic efficacy of the CAPD. Results: Sixty-one (24.4%) patients had delirium during pediatric intensive care unit hospitalization,and significant differences existed between them and 189 patients without delirium in age (χ(2)=11.220, P=0.011), rates of mechanical ventilation (χ(2)=6.691, P=0.010) and length of PICU hospitalization (Z=10.656, P=0.001). The Chinese version of the CAPD scale had high discrimination among the 8 items in the main table without skewed distribution and discrete items. The internal and external reliability of the scale were preferable as the Cronbach's α coefficient was 0.819 and ICC value was 0.835. The KMO value of EFA was 0.834 and Bartlett spherical test showed statistical significance (χ(2)=661.440, P<0.01). CFA constructed a two-factor structural equation model with favorable fit index:Chi-square/degree of freedom (DF) ratio was 1.786 (χ(2)=33.930, P=0.019), goodness of fit index (GFI) was 0.967,adjusted GFI was 0.938,root mean square of the approximate error (RMSEA) was 0.056. A good diagnostic efficacy of this scale was demonstrated by receiver operating characteristic curve analysis as when the cut-off was 10, the area under curve was 0.99 (95%CI: 0.97-0.99), sensitivity was 96.7% (95%CI: 88.7-99.6), specificity was 93.1% (95%CI: 88.5-96.3), positive likelihood ratio was 14.0 (95%CI: 13.2-14.9), and the negative likelihood ratio was 0.035 (95%CI: 0.008-0.200). Conclusion: The Chinese version of the CAPD Scale has favorable reliability,validity,diagnostic efficacy, as well as feasibility, which should be applied in evaluation of pediatric delirium in clinical observation and research in the future.目的：探讨中文版康奈尔儿童谵妄量表（CAPD）的临床应用。 方法：前瞻性观察性研究。自2017年6月至2017年12月，根据指南对CAPD原表进行汉化及跨文化调试。在重庆医科大学附属儿童医院重症医学科（PICU）进行初步临床应用，联合躁动-镇静评分，使用中文版CAPD量表对患儿进行谵妄评估，收集250例儿科重症监护室患儿的临床资料，组间比较采用χ(2)检验或秩和检验。采用Cronbach′s α系数、Split-half系数和组内相关系数（ICC）评价量表信度以及探索性因子分析和验证性因子分析评价量表结构效度，与金标准诊断结果进行比对，评价量表的诊断效能。 结果：61例（24.4%）患儿在PICU住院期间发生谵妄，与189例未发生谵妄的患儿在年龄（χ(2)=11.220，P=0.011），机械通气（χ(2)=6.691，P=0.010）以及PICU住院时间（Z=10.656，P=0.001）三个因素上差异有统计学意义；中文版CAPD量表评估主表8个条目区分度较高，无明显偏态分布、离散的条目；量表内部信度Cronbach′s α系数为0.819，外部信度ICC值0.835；探索性因子分析（EFA）KMO值为0.834，Bartlett球形检验差异有统计学意义（χ(2)=661.440，P<0.01）；验证性因子分析（CFA）构建两因子结构方程模型：χ(2)/自由度为1.786（χ(2)=33.930，P=0.019），拟合优度指数为0.967，调整的拟合优度指数为0.938，近似误差均方根为0.056，模型拟合指数良好；受试者工作特征（ROC）曲线分析示截断值为10时，曲线下面积（AUC）为0.99（95%CI：0.97~0.99），灵敏度为96.7%（95%CI：88.7~99.6），特异度为93.1%（95%CI：88.5~96.3），阳性似然比为14.0（95%CI：13.2~14.9），阴性似然比为0.035（95%CI：0.008~0.200），中文版CAPD量表具有良好的诊断效能。 结论：中文版CAPD能够有效地评估儿童谵妄的发生，可行性强，可以用于今后儿童谵妄相关临床观察研究的评估工具。.