Table_2_Development and Validation of a Nomogram for Predicting Postoperative Delirium in Patients With Elderly Hip Fracture Based on Data Collected on Admission.DOCX

Delirium is a common postoperative complication in elderly hip fracture patients that seriously affects patients’ lives and health, and early delirium risk prediction, and targeted measures can significantly reduce the incidence of delirium. The purpose of this study was to develop and validate a nomogram for the prediction of postoperative delirium (POD) in elderly hip fracture patients. A total of 328 elderly patients with hip fractures enrolled retrospectively in department 1 of our hospital were randomly divided into the training set (n = 230) and the internal validation set (n = 98). The least absolute shrinkage and selection operator (LASSO) regression analysis was used for feature variable selection, and multivariate logistic regression with a backward stepwise method was used to construct a nomogram in the training set. The discrimination efficacy and calibration efficacy of the nomogram were evaluated through the receiver operating characteristic (ROC) curve and calibration curve, respectively. The clinical usefulness was estimated through decision curve analysis (DCA) and clinical impact curve (CIC) analysis. Another validation set from department 2 of our hospital, containing 76 elderly patients with hip fractures, was used for external validation of the nomogram. A total of 43 (13.1%) and 12 (15.8%) patients had POD in department 1 and department 2, respectively. The nomogram was constructed by three predictors, including dementia, chronic obstructive pulmonary disease (COPD), and albumin level. The nomogram showed good discrimination efficacy and calibration efficacy, with the AUC of 0.791 (95% CI, 0.708–0.873), 0.820 (95% CI, 0.676–0.964), and 0.841 (95% CI, 0.717–0.966) in the training set, the internal validation set, and the external validation set, respectively. Both DCA and CIC demonstrated that this nomogram has good clinical usefulness. The nomogram constructed by dementia, COPD, and albumin level can be conveniently used to predict POD in patients with elderly hip fractures.

谵妄（Delirium）是老年髋部骨折患者常见的术后并发症，严重影响患者的生活与健康，早期开展谵妄风险预测并采取针对性干预措施，可显著降低谵妄的发生率。本研究旨在开发并验证一款用于预测老年髋部骨折患者术后谵妄（postoperative delirium, POD）的列线图（nomogram）。研究回顾性纳入我院一科的328例老年髋部骨折患者，将其随机分为训练集（n=230）与内部验证集（n=98）。采用最小绝对收缩和选择算子（least absolute shrinkage and selection operator, LASSO）回归分析进行特征变量筛选，并在训练集中通过向后逐步法构建多因素logistic回归列线图。分别通过受试者工作特征（receiver operating characteristic, ROC）曲线与校准曲线评估该列线图的区分效能与校准效能；通过决策曲线分析（decision curve analysis, DCA）与临床影响曲线（clinical impact curve, CIC）评估其临床实用性。另纳入我院二科的76例老年髋部骨折患者作为外部验证集。我院一科与二科中分别有43例（13.1%）与12例（15.8%）患者发生术后谵妄。该列线图由3个预测因子构建，包括痴呆（dementia）、慢性阻塞性肺疾病（chronic obstructive pulmonary disease, COPD）与白蛋白水平。该列线图展现出良好的区分效能与校准效能：在训练集、内部验证集与外部验证集中的AUC分别为0.791（95%CI：0.708~0.873）、0.820（95%CI：0.676~0.964）与0.841（95%CI：0.717~0.966）。决策曲线分析与临床影响曲线均证实该列线图具备良好的临床实用性。综上，由痴呆、慢性阻塞性肺疾病及白蛋白水平构建的列线图可便捷用于预测老年髋部骨折患者的术后谵妄风险。