Construction of Nomogram prediction model for death based on new critical illness score scale for patients in intensive care unit

WANG Ruyue; ZHANG Xiumin; ZHANG Caifeng; TIAN Long; LIU Shuang

doi:10.16048/j.issn.2095-5561.2025.06.05

Abstract：Objective To construct the Nomogram prediction model based on the new intensive care unit(ICU)critical illness score scale and traditional score scale for the death of the patients in ICU.Methods A total of 1 200 patients in ICU of our Hospital from December 2019 to December 2023 were selected.Among the patients,there were 743 males and 457 females,with an average age of(62.5±8.3)years old,ranging from 51 to 74 years old.The patients were randomly divided into the modeling group(n=840)and validation group(n=360).A new critical illness score scale(hereinafter referred to as the'Table')was created by using the baseline data from the modeling group.Basing on the Table score,modified early warning score(MEWS),acute physiology and chronic health evaluation Ⅱ score(APACHEⅡ),sequential organ failure score(SOFA),Tilburg frailty indicator.Nomogram prediction model for the death of the patients in ICU was constructed and verified.Results High Table,MEWS,APACHEⅡ,and SOFA scores were independent risk factors affecting the death rate in the modeling group(P<0.05).Nomogram model was constructed based on the 4 predictive factors mentioned above.The model had good consistency and could provide clinical net benefits.Area under the curve predicted by the model for death in the modeling group and validation group were 0.883 and 0.874,respectively.Receiver operating characteristic curves of the two groups were well fitted(P>0.05).Conclusion The prediction performance of Nomogram model based on the new Table and traditional score scales for the death of the patients in ICU can meet the clinical requirements.The application of it can provide some reference for reducing the death rate of the patients.