儿童阑尾切除术研究中的缺失数据处理:现状与NSQIP分析
本研究先系统检索2023年儿童阑尾切除术文献评估缺失数据报告情况,再回顾分析NSQIP-P数据库2015—2022年142,129例病例,比较五种缺失数据处理策略对术后感染并发症危险因素识别的影响。结果显示文献报告不足,且处理策略会改变部分变量的显著性。
要点
- 儿童阑尾切除术文献中54.3%未提及缺失数据,仅7.3%使用高级统计方法。
- NSQIP-P队列39.3%记录存在缺失值,主要涉及身高和种族。
- 完全病例分析最保守,识别出的显著预测因素最少。
核心结论
- 儿童外科研究对缺失数据的方法学透明度严重不足,数据处理策略的选择会显著改变临床危险因素的识别,应优先采用多重插补而非简单排除方法。
研究设计与样本
- 系统检索PubMed与Web of Science评估2023年儿童阑尾切除术文献的缺失数据报告;回顾分析NSQIP-P数据库2015—2022年142,129例病例,比较五种缺失数据处理策略。
主要发现
- 文献综述纳入116篇符合条件的文章,54.3%未提及缺失数据,86.4%报告所用方法依赖完全病例分析。
- NSQIP-P队列中39.3%记录存在缺失值,主要为身高和种族。
- ASA分级和手术时长等核心预测因素在多变量logistic回归中保持稳定,而术前白细胞计数和人体测量指标随处理方式变化出现显著性波动。
- 完全病例分析最保守,识别出的显著预测因素最少。
临床怎么用
- 研究者在分析儿童阑尾切除术数据时应优先采用多重插补而非简单排除,并规范报告缺失情况。
局限性
- 文献综述仅覆盖2023年发表的文章;NSQIP-P为回顾性数据库分析,结果外推需谨慎。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Appendectomy is a leading pediatric surgical procedure, yet the impact of missing data handling on clinical conclusions remains unexplored in pediatric surgery. This study aimed to identify current reporting practices in the pediatric appendectomy literature and quantify how different missing-data strategies influence the identification of postoperative infectious-complication risk factors. METHODS: A systematic review of pediatric appendectomy literature from 2023 was conducted across PubMed and Web of Science to assess missing data reporting. Subsequently, a retrospective analysis of the National Surgical Quality Improvement Program-Pediatric (NSQIP-P) database (2015-2022) was performed on 142,129 cases. Five strategies were compared: available case analysis, threshold-based exclusion, complete case analysis, simple imputation, and multiple imputation. Multivariable logistic regression models identified risk factors for five infectious outcomes, including combined infectious complications, organ space infections, deep incisional infections, superficial surgical site infections, and wound dehiscence. RESULTS: The literature review identified 116 eligible articles, of which 54.3% failed to mention missing data. Only 7.3% used advanced statistical techniques, while 86.4% reported that a method relied on complete-case analysis. In the NSQIP-P cohort, 39.3% of records had missing values, primarily in height and race. While core predictors like American Society of Anesthesiologists (ASA) class and operative duration remained stable in multivariate logistic regression analysis, variables such as preoperative white blood cell count and anthropometric measurements fluctuated in significance depending on the handling method. Complete case analysis was the most conservative, identifying the fewest significant predictors. CONCLUSION: Methodological transparency regarding missing data is critically lacking in pediatric surgical research. Because the choice of data-handling strategy significantly alters the identification of clinical risk factors, researchers should prioritize multiple imputation over simple exclusion methods. Journals must mandate rigorous reporting of missingness to ensure the reliability of evidence-based surgical guidelines.
引用
- Tsutsumi A, Si T, Salas J, et al. Missing data handling in pediatric appendectomy research: current practices and National Surgical Quality Improvement Program (NSQIP) analysis. Surg Endosc. 2026. PMID: 42509504. doi:10.1007/s00464-026-13123-7
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。