机器学习模型在外科手术患者压力性损伤风险预测中的应用
打开文本图片集
中图分类号 R473.6 文献标识码A文章编号 2096-7721（2025）11-1958-08
Application of machine learning models in predicting pressure injury risk in surgical patients
YANG Jun'， CHEN Xiao’， WANG Zhijing2， WANG Yafei2， ZHANG Junqing3，TIAN Kun4，ZHANG Haiyan5 （1.ObstetricicalegoutralHspialngzuna;eptfsigngzualal CangzhouO61ooo，China;3.DeliveryRoomofObstetric Medical Center，CangzhouCentral Hospital，CangzhouO610oo，Cina; 4.DepartmentofCardiothoracic Surgery，Cangzhou CentralHospital，Cangzhou O61ooo，China;5.NursingTeachingand Research Section， Cangzhou Medical College，Cangzhou O61ooo， China）
AbstractObjective：Toinvestigateandanalyethepredictiveefcacyofmachinelearingmodelsinpredictingpressreinjuryrisks amongsurgicalpatients.Methods：40lsurgicalpatientsreadmitedtoCangzhouCentralHospitalfromDecember2022 toDecember2023 wereenrolldTheyweredividedintothebservationgroup（withpressureijury）andthecontrolgroup（withoutpresureinjurybased ontheourenceofpresureinjury.Clinicaldatafromthegroupswerecolectedviaelectronicmedicalrecords.Comparativeanalysis wasperformedusingregressionmodelinmachinelearingtoevaluatethepredictiveperformanceoftraditionalmodels，machineleaing models，andotaldeoultinosturebilityctiiilityfrtafe，totaladenregly turningfreqeyorarartas，oinstoycoinceenia，siesiogsdl serumalbuminlevelsweresignificantlyassociatedwithpressureinjury insurgicalpatients.TheAUCvaluesrankedasfollows：Random Forest>XGBoost>NuralNetwork>BinaryLogisticRegresion>Total Braden Score.Conclusion：Machinelearningmodelshassuperior predictiveperformanceforpressureinjuryriskcomparedtotradionalmodelsandtheBradenscore.Cliniciansshould prioritizefactors suchasskinmoisture，liitedmobilityoroaryartdisease，okinghistorycontinence，andbdriddentie>datoance preventive measures against pressure injuries.
Key wordsMachine Learning; Surgery; Pressure Injury; Risk Prediction
压力性损伤作为外科手术患者高发且严重的并发症，不仅威胁患者健康、延长住院时间，还会带来巨大的医疗负担[1-2]。（剩余9664字）
