Establishment And Evaluation Of Big Data And Machine Algorithm Risk Prediction Model For Acute Kidney Injury In Patients With Severe Pneumonia

Objective:To analyze the risk factors of acute kidney injury(AKI)in patients with severe pneumonia,establish a nomogram risk prediction model and evaluate the model.Methods: The clinical data of 284 patients with severe pneumonia in our hospital from January 2018 to December 2019 were retrospectively analyzed,according to the occurrence of AKI,the patients were divided into AKI group and non AKI group,the risk factors of AKI were screened by univariate analysis and multivariate logistic regression analysis,the regression equation was established,the risk factors were included in R software,and AKI nomogram prediction model was established,the discrimination of AKI was predicted by ROC,and the accuracy of the evaluation model was evaluated by calibration curve and goodness of fit.Results: Among 284 patients with severe pneumonia,136 had AKI,the incidence rate was47.89%(136/284);the RRT rate at admission and mortality of AKI group were higher than those of non AKI group(P< 0.05),and the length of hospital stay was longer than that of non AKI group(P < 0.05);the mortality of AKI in different stages was statistically significant(P< 0.05),the mortality of AKI stage 1 was significantly lower than that of AKI stage 2 and AKI stage 3(P <0.05);Logistic regression analysis showed that chronic kidney disease(OR = 3.430,95% CI:1.236-9.518),PCT level(OR = 1.681,95% CI: 1.137-2.485),APACHEⅡscore(OR = 1.127,95% CI: 1.055-1.203)and history of mechanical ventilation(OR = 2.378,95% CI: 1.366-4.140)were independent risk factors for AKI(P< 0.05);the nomogram prediction model based on the risk factors screened by logistic regression model showed that the area under the ROC curve(AUC)was 0.830(95% CI: 0.781-0.872),the sensitivity was 73.72%,and the specificity was86.39%;the slope of calibration curve was close to 1;Hosmer-lemeshow goodness of fit test showed that =10.879,P = 0.209.Conclusion:The nomogram prediction model based on the risk factors of AKI in patients with severe pneumonia has good discrimination and consistency,which can provide some guidance for the prevention of AKI in patients with severe pneumonia.