复杂型先天性心脏病患儿术后应用危重程度评分工具的效果分析
网络出版日期: 2014-05-30
Analysis of applying scoring systems of severity of illness on children with complex congenital heart diseases after operations
Online published: 2014-05-30
目的 分析3种危重程度评分工具在评估复杂型先天性心脏病(CHD)患儿术后病情中的适用性。方法 按照小儿死亡风险评分(PRISM)、第3代小儿死亡风险评分(PRISM Ⅲ)及小儿危重病例评分法(PCIS)3种评分工具的使用要求,收集237例复杂型CHD患儿的一般资料及术后结局进行评分。采用组内相关系数(ICC)测量评定者间的信度;运用独立样本t检验、受试者工作特征(ROC)曲线下面积(AUC)、Hosmer-Lemeshow拟合优度检验、缺失值分析等方法评价3种评分工具的性能和适用性。结果 237例复杂型CHD术后患儿中,死亡24例,存活213例。ICC>0.8提示资料收集可靠。死亡组PRISM和PRISM Ⅲ得分均显著高于存活组,差异有统计学意义[(22.3±6.9)分vs(19.3±6.5)分,P<0.05;(15.7±7.1)分 vs(11.1±4.7)分,P<0.05];但死亡组PCIS得分与存活组比较差异无统计学意义[(79.6±5.5)分vs(81.9±7.6)分,P>0.05]。3种评分工具校准能力较好(P>0.05),但区分能力尚显不足,PRISM、PRISM Ⅲ和PCIS的AUC分别为0.695、0.724和0.220。3种评分工具均存在部分指标与国内监护室工作模式不匹配从而导致缺失的现象。结论 PRISM、PRISM Ⅲ及PCIS对复杂型CHD患儿术后病情评估的临床指导意义不强,适用性有待提高。在现有评分工具的基础上,优化适用于该类疾病敏感度更高的评价指标,从而形成特异度更强的评分工具。
关键词: 小儿死亡风险评分; 第3代小儿死亡风险评分; 小儿危重病例评分法; 先天性心脏病
张婷婷 , 庄周颖 , 傅丽娟 . 复杂型先天性心脏病患儿术后应用危重程度评分工具的效果分析[J]. 上海交通大学学报(医学版), 2014 , 34(5) : 700 . DOI: 10.3969/j.issn.1674-8115.2014.05.025
Objective To analyze the applicability of three scoring systems of severity of illness for the evaluation of children with complex congenital heart diseases (CHD) after operations. Methods The general data of 237 patients were collected and the results of operations were scored according to the requirements of three scoring systems, i.e. the pediatric risk of mortality score (PRISM), pediatric risk of mortality score Ⅲ (PRISM Ⅲ), and pediatric clinical illness score (PCIS). The intra-class correlation coefficient (ICC) was used to measure the inter-rater reliability. The performance and applicability of the three scoring systems were evaluated by the Ttest, area under the receiver operating characteristic (ROC) curve, Hosmer-Lemeshow goodness of fit test, and missing value analysis. Results Among 237 children with complex CHD, 24 died after operations and 213 survived. ICC>0.8 meant that the collected data was reliable. Scores of PRISM and PRISM III of the death group (22.3±6.9 and 15.7±7.1) were significantly higher than those of the survival group (19.3±6.5 and 11.1±4.7) and the differences were statistically significant (P<0.05). But the difference of scores of PCIS between the death group (79.6±5.5) and survival group (81.9±7.6) was not statistically significant (P>0.05). The calibration capability of the three scores was good (P>0.05), but the discriminatory ability was insufficient. The AUCs of PRISM, PRISM III, and PCIS were 0.695, 0.724, and 0.220, respectively. Some indexes of three scoring systems were missed due to incompatibility with the work pattern of care units of our country. Conclusion The clinical applicability of PRISM, PRISM III, and PCIS for the evaluation of children with complex CHD after operations was not ideal and need to be improved. It is necessary to optimize evaluation indexes that are more sensitive to these diseases to build more specific scoring systems based on existing scoring systems.
/
〈 |
|
〉 |