新疆医科大学学报

目的 探讨ICU患者医疗器械相关性压力性损伤(Medical device related pressure injury, MDRPI)的危险因素，构建预测MDRPI的列线图模型。方法 选取2021年6月-2022年2月收治的ICU患者280例，根据是否发生MDRPI分为发生组(n=55)和未发生组(n=225),并根据患者的一般资料、临床资料及医疗器械使用资料进行单因素分析和Logistic回归分析，构建预测MDRPI风险的列线图模型。结果 280例研究对象中有55例发生MDRPI(19.6%)。单因素分析显示，两组患者的年龄、Braden量表评分、急性生理与慢性健康评分(APACHEⅡ评分)、血清白蛋白、血红蛋白、红细胞压积、手术、俯卧位、血管活性药使用、镇静药使用、肠内营养、气管插管使用、鼻胃管使用、无创通气面罩使用时间、器械使用总数差异均有统计学意义(P<0.05)。多因素Logistic回归分析显示，年龄、手术、俯卧位、APACHEⅡ评分、无创通气面罩使用时间是MDRPI的独立危险因素；血红蛋白是保护因素(P<0.05)。基于上述影响因素构建ICU患者MDRPI的列线图预测模型，校准度(x²=7.177,P=0.518)和区分度(AUC=0.953,95%CI 0.926～0.979)良好。决策曲线显示，当患者的阈值概率为0～1,使用列线图预测模型预测MDRPI风险的净收益更高。结论 本研究构建的ICU患者MDRPI风险列线图预测模型具有良好的风险识别能力，可以为尽早制定预见性干预策略提供理论指导。

Objective To explore the risk factors of medical device related pressure injury in ICU patients and to establish a risk prediction model for predicting the risk of pressure injury.Methods 280 patients in ICU admitted from June 2021 to February 2022 were divided into two groups by the occurrence of MDRPI.According to general data, clinical data and medical device use data, univariate analysis and Logistic regression analysis were carried out to construct a risk prediction model to predict the risk of MDRPI.Results Among the 280 subjects, 55 patients developed MDRPI(19.6%).Univariate analysis showed that there were significant differences in age, Braden scale score, APACHEⅡscore, serum albumin, hemoglobin, hematocrit, surgery, prone position, use of vasoactive agents, use of sedatives,enteral nutrition,use of endotracheal intubation,use of nasogastric tube,use of non-invasive ventilation mask and total use of devices between the two groups.Multivariate Logistic regression analysis showed that age,operation,prone position,Apache II score and duration of non-invasive ventilation mask were independent risk factors for MDRPI,and hemoglobin was a protective factor.Based on the above influencing factors,a nomogram prediction model for the risk of MDRPI in patients with ICU was constructed,and calibration(x²=7.177,P=0.518) and differentiation(AUC=0.953,95% CI 0.926～0.979) were good.The decision curve showed that when the threshold probability of the patient was 0～1,the net benefit of using the nomogram prediction model to predict the risk of MDRPI was higher.Conclusion The MDRPI risk prediction model of ICU patients constructed in this study has a good risk identification ability and can provide theoretical guidance for formulation of predictive intervention strategies as soon as possible.