基于最小绝对收缩和选择算子回归的经外周中心静脉导管置管部位医用粘胶相关性皮肤损伤风险预测模型构建与验证OA
Construction and validation of a risk prediction model for medical adhesive-related skin injury at peripherally inserted central catheter insertion sites based on least absolute shrinkage and selection operator regression
目的 基于最小绝对收缩和选择算子(LASSO)回归构建经外周中心静脉导管(PICC)置管部位医用粘胶相关性皮肤损伤(MARSI)风险预测模型并验证.方法 回顾性分析504例PICC置管患者的临床资料,按7∶3比例将患者分为训练集(352例)与验证集(152例),其中训练集用于模型构建,验证集用于内部验证.采用LASSO回归方法筛选PICC患者发生MARSI的潜在风险因素,再通过单因素和多因素Logistic回归分析筛选独立影响因素,据此构建列线图预测模型并验证其效能.结果 LASSO回归筛选出性别、年龄、吸烟史、肿瘤类型、化疗次数、置管部位、敷料更换频率、置管次数、维护方法、医用粘胶类型、皮肤潮湿、皮肤水肿、红细胞、白蛋白、血小板、天冬氨酸氨基转移酶为保留变量.单因素Logistic回归分析结果显示,年龄、置管部位、敷料更换频率、置管次数、维护方法、医用粘胶类型、皮肤潮湿、皮肤水肿是PICC患者发生MARSI的影响因素(P<0.05).多因素Logistic回归分析结果显示,年龄、敷料更换频率、置管次数、维护方法、医用粘胶类型、皮肤潮湿、皮肤水肿是PICC患者发生MARSI的独立影响因素(P<0.05).基于此构建列线图模型,受试者工作特征(ROC)曲线显示,该模型在训练集和验证集中的曲线下面积(AUC)分别为0.926(95%CI:0.887~0.964)、0.888(95%CI:0.817~0.959),区分度良好.结论 年龄增长、敷料更换频率>7 d、置管次数>2次、维护方法不标准、使用透明敷料、皮肤潮湿及皮肤水肿是PICC置管部位MARSI发生的独立危险因素;据此构建的列线图模型具有良好的区分能力与校准效果,可为临床早期风险分层提供参考.
Objective To construct and validate a risk prediction model for medical adhesive-re-lated skin injury(MARSI)at peripherally inserted central catheter(PICC)insertion sites based on least absolute shrinkage and selection operator(LASSO)regression.Methods The clinical data of 504 patients with PICC insertion were retrospectively analyzed.Patients were divided into training set(352 cases)and validation set(152 cases)at a ratio of 7∶3,with the training set used for model construction and the validation set used for internal validation.LASSO regression was used to screen potential risk factors for MARSI in patients with PICC insertion,and univariate and multivariate logistic regression analyses were then performed to identify independent influencing factors.A nomo-gram prediction model was constructed accordingly,and its performance was validated.Results LASSO regression identified sex,age,smoking history,tumor type,the number of chemo-therapy sessions,insertion site,dressing change frequency,the number of catheter insertions,main-tenance method,medical adhesive type,skin moisture,skin edema,red blood cell count,albumin,platelet count,and aspartate aminotransferase as retained variables.Univariate logistic regression a-nalysis showed that age,insertion site,dressing change frequency,the number of catheter inser-tions,maintenance method,medical adhesive type,skin moisture,and skin edema were influencing factors for MARSI in patients with PICC insertion(P<0.05).Multivariate logistic regression analy-sis showed that age,dressing change frequency,the number of catheter insertions,maintenance method,medical adhesive type,skin moisture,and skin edema were independent influencing factors for MARSI in patients with PICC insertion(P<0.05).A nomogram model was constructed based on these factors.The receiver operating characteristic(ROC)curve showed that the area under the curve(AUC)of the model was 0.926(95%CI:0.887-0.964)in the training set and 0.888(95%CI:0.817-0.959)in the validation set,indicating good discrimination.Conclusion Increased age,dressing change frequency>7 d,number of catheter insertions>2,nonstandard maintenance meth-od,use of transparent dressings,skin moisture,and skin edema are independent risk factors for MARSI at PICC insertion sites.The nomogram model constructed based on these factors has good dis-crimination and calibration,and can provide a reference for early clinical risk stratification.
司亮亮;赵文利;郭亚茹;姜鑫;刘琰;刘瑞青
河南省人民医院,静脉治疗中心,河南郑州,450000河南省人民医院,静脉治疗中心,河南郑州,450000河南省人民医院,静脉治疗中心,河南郑州,450000河南省人民医院,静脉治疗中心,河南郑州,450000河南省人民医院,静脉治疗中心,河南郑州,450000河南省人民医院,肿瘤介入科,河南郑州,450000
医药卫生
最小绝对收缩和选择算子回归经外周中心静脉导管置管部位医用粘胶相关性皮肤损伤影响因素列线图风险预测模型Logistic回归分析
least absolute shrinkage and selection operator regressionperipherally inserted central catheterinsertion sitemedical adhesive-related skin injuryinfluencing factorsnomo-gramrisk prediction modellogistic regression analysis
《实用临床医药杂志》 2026 (19)
20-29,10
河南省医学科技攻关计划联合共建项目(LHGJ20240028)
评论