Construction Of Disease Early Warning Model For Children With Sepsis In PICU

ObjectiveSepsis has the characteristics of rapid disease progression,rapid disease change,untimely discovery and poor prognosis.It is a common critical illness in clinical practice.If potential patients cannot be identified in time,severe sepsis often progresses when the disease is discovered.Syndrome and multiple organ failure,seriously affecting the prognosis of patients.Therefore,early detection,early diagnosis and timely intervention are the keys to improving the prognosis of sepsis.The purpose of this study is to screen for risk factors that may be related to early warning of diseases in children with sepsis,to build an early warning model of diseases in children with sepsis,and to determine graded care based on the cut-off value of the early warning model and formulate individualized care measures.Caring for children in a targeted manner to ensure early detection and timely treatment of children with sepsis,improve the prognosis of children as much as possible,and reduce the adverse effects of sepsis on children's long-term development.MethodsIn this study,children who were admitted to the pediatric intensive care unit(PICU)of a top three hospital in Qingdao from November 8,2017 to September 18,2019 were selected as the research object.and screened them according to the diagnostic criteria of sepsis Children with conditional sepsis were used as the observation group,and children with inflammation that could be clearly identified by imaging and laboratory tests were used as the control group.Through single factor correlation analysis to determine indicators that may be related to early warning of sepsis in children,through multivariate logistic regression analysis to determine independent risk factors for predicting sepsis,form a regression equation,determine the best critical point for prediction,and use Receiver Operating Characteristic Curve(ROC curve)and Area Under the Curve(AUC)evaluate the model's resolving power,and use Hosmer-Lemeshow goodness-of-fit test to evaluate the disease early warning model The degree of fit.Determine the level of care through the obtained critical points,and formulate personalized nursing measures based on the existing nursing aids based on literature review and interviews.Select the children from 30 days to 18 years old who will go to the pediatric emergency department of a top three hospital in Qingdao from October 15th,2019 to January 15th,2020 as the object of testing the early warning model of children with sepsis.The value is substituted into the equation to verify the model equation obtained from the study.The validity of the predictive model is judged based on the positive rate of the child's diagnosis(whether it is sepsis).Results1.In the first part of this study,a total of 485 cases were collected.After excluding cases that did not meet the inclusion criteria and missing items,417 cases were finally included in the study,including 93 children with sepsis and non-septic324 children were infected.After independent factor analysis and multivariate logistic regression analysis,the independent risk factors for the early warning model of children with sepsis are C-reactive protein(CRP),procalcitonin(PCT),and albumin level(ALB).Among them,CRP and PCT are positively correlated with the formation of sepsis,the higher the value,the more likely it is to develop sepsis;the albumin level is negatively correlated with the formation of sepsis,the lower the albumin level It is possible to develop sepsis.The resulting model equation is Z=6.688+0.021(CRP)+0.050(PCT)-0.080(ALB).The optimal critical point is the Yoden index(0.647)when the sensitivity is 78.5%and the specificity is 86.2%.The area under the ROC curve of the model was 0.879,and the 95%confidence interval was 0.836-0.922,suggesting that the model has a good ability to distinguish between children with sepsis and children with general infection.The Hosmer-Lemeshow goodness-of-fit test indicates that the model has good Goodness of fit(C~2=7.419,P=0.492).2.In the second part of the study,a total of 145 cases were collected.After excluding the cases that did not meet the inclusion criteria and the study object fell out during the study,125 cases were finally included in the study.Substitute the CRP,PCT and ALB of the child into the equation to predict sepsis.The sensitivity,specificity,positive predictive value,negative predictive value and diagnostic coincidence rate of the sepsis disease early warning model finally obtained were52.5%,81.2%,56.8%,78.4%and 72%,respectively.The disease early warning model equation of children with sepsis has high sensitivity and specificity,and the nurse can effectively predict the occurrence of sepsis,and implement the nursing intervention program for children with sepsis as early as possible,which greatly reduces the number of nursing staff The evaluation time avoids delays in treatment of children.Conclusion1.The disease early warning model of children with sepsis composed of CRP,PCT and ALB levels can early warning of the occurrence of sepsis in children in PICU.The equation of the early warning model is Z=6.688+0.021(CRP)+0.050(PCT)-0.080(ALB).The higher the levels of CRP and PCT in the equation,the more likely sepsis occurs,and the lower the ALB level,the more likely the child will develop sepsis.2.The early warning model obtained by the study has good fitting goodness,high sensitivity and specificity,and can effectively predict the possibility of children with inflammation developing into children with sepsis,and can be used as early detection,early diagnosis,early medical and nursing intervention for medical staff in accordance with.