Braden评分联合微循环评价指标构建创伤重症患者压力性损伤风险可视化预后模型

目的:基于Braden评分，联合局部皮肤温度和局部组织氧饱和度作为局部微循环指标，构建创伤重症患者压力性损伤风险预后模型，并开发可视化列线图。方法:选取浙江省某三甲医院2020年06月1日至2020年08月31日期间所有进入急诊重症监护室（EICU）的创伤患者，采取Braden量表评估患者压力性损伤发生风险，测量患者骶尾部受压部位皮肤温度与血氧饱和度，作为微循环评价指标。采用多因素Logistic回归构建基于Braden评分联合微循环评价指标的创伤重症患者压力性损伤预后模型及可视化列线图，并与单纯Braden量表构建的预测模型比较。通过计算受试者操作特征曲线下面积判断预测模型区分度。C指数对最佳模型进行内部验证、H-L拟合优度检验、整体鉴别指数观察预测模型效果。结果:本研究共纳入152例患者，其中33例发生压力性损伤，占21.71%。Braden量表联合局部皮肤温度与局部组织氧饱和度后曲线下面积为0.866，模型内部验证C指数为0.847。H-L拟合优度检验结果为8.37， P=0.051，即模型具有良好区分度与一致性。模型整体鉴别指数=0.144， P=0.023。 结论:Braden量表联合局部微循环指标构建创伤重症患者压力性损伤预后模型具有良好区分度与一致性，较单纯Braden量表预测效力提高14.4%，据此构建列线图，能够为临床提供快捷、方便可靠的预测工具。

A pressure injury risk prognosis model for severe trauma patients based on Braden scale combined with microcirculation indicators

Objective:Based on Braden scale, combined with local skin temperature and local tissue oxygen saturation as microcirculation indicators, to construct a pressure injury risk prognosis model for severe trauma patients, and develop a visual nomogram.Methods:All the trauma patients in the Emergency Intensive Care Unit (EICU) of a tertiary hospital in Zhejiang Province from June 1, 2020 to August 31, 2020 were selected. The Braden scale was used to assess the risk of pressure injury in the patient and measure the patient’s sacral injury. The skin temperature and blood oxygen saturation at the pressure site were used as indicators for microcirculation evaluation. Multivariate Logistic regression was used to construct a prognostic model and visual nomogram for severe trauma patients with stress injury based on Braden score combined with microcirculation evaluation indicators, and compared with the prediction model constructed by the Braden scale alone. The discrimination degree was judged by the area under the receiver operating characteristic curve. The C index performed internal verification of the model, H-L goodness-of-fit test, and the overall discrimination index to observe and predict the effect of the model.Results:A total of 152 patients were enrolled in this study, of which 33 (21.71%) had pressure injuries. The area under the curve of the Braden scale combined with local skin temperature and local tissue oxygen saturation was 0.866, and the internal verification C index of the model was 0.847. The H-L goodness of fit test result was 8.37 ( P=0.051), indicating that the model had good discrimination and consistency. The overall discrimination index of the model was 0.144 ( P=0.023). Conclusions:The Braden scale combined with local microcirculation indicators to construct a prognostic model of stress injury in severe trauma patients has good discrimination and consistency, and the predictive power of the Braden scale is improved by 14.4%. The construction of a nomogram can provide clinical rapid convenient and reliable forecasting tool.