ICU病人譫妄風(fēng)險預(yù)測模型的應(yīng)用價值比較

打開文本圖片集

Comparison of application value of prediction models in the prediction of delirium risk in ICU patients

ZUO Xianfang， ZHANG Aiqin

Medical School of Nanjing University， Jiangsu 210008 China

Corresponding Author  ZHANG Aiqin， E?mail： aq09z@126.com

Abstract  Objective：To compare and analyze the ability to predict delirium and application value of Zhu Xiaoying model，Lanzhou model，PRE?DELIRIC and E?PRE?DELIRIC in ICU patients.Methods：Patients who were admitted to the neurology ICU， neurosurgery ICU，general surgery ICU， cardiothoracic surgery ICU and respiratory medicine ICU in Eastern Theater General Hospital from December 2020 to April 2021 were selected as the research subjects.When the patient was admitted to the ICU，we applied E?PRE?DELIRIC as the assessment tool，the other three prediction models were used when patients were admitted to ICU within 24 hours to assess the risk of delirium.Using AUROC and Hosmer?Lemeshow goodness?of?fit test to evaluate the discrimination and calibration of four models in the prediction of delirium risk in ICU patients.Results：A total of 482 ICU patients were included，of which 150 had delirium，the incidence of delirium in ICU patients was 31.12%.The AUROC value of Zhu Xiaoying's model was 0.806 （95%CI：0.768?0.840），the AUROC value of the Lanzhou model was 0.853 （95%CI：0.818?0.883）， and the AUROC value of PRE?DELIRIC was 0.918（95%CI：0.890?0.941）.The AUROC value of E?PRE?DELIRIC was 0.891（95%CI：0.860?0.918）.Delong's method was used to compare the AUROC values of the four models，and pairwise comparison showed statistically significant differences.The results of the Hosmer?Lemeshow goodness?of?fit test showed that PRE?DELIRIC and E?PRE?DELIRIC had a high degree of consistency between actual occurrence risk and model predicted risk（P>0.05）.Conclusion：The four models had a high degree of discrimination in predicting the risk of delirium，and PRE?DELIRICand E?PRE?DELIRIC had a high degree of calibration，the overall prediction effect of PRE?DELIRIC was good.ICU staff can select the appropriate delirium prediction model to apply according to the clinical situation， and take corresponding nursing measures combined with the risk stratification of the model to improve the clinical outcome of patients.

Keywords    delirium； intensive care unit，ICU； prediction model； risk of occurence

摘要  目的：比較并分析祝曉迎模型、蘭州模型、譫妄預(yù)測模型和早期譫妄預(yù)測模型預(yù)測重癥監(jiān)護室（ICU）病人發(fā)生譫妄的能力和應(yīng)用價值。（剩余14736字）