Research On Predictive Model Of Sepsis In Preterm Children Based On Sample Entropy

The incidence of premature children is increasing year by year,and its complications especially sepsis are one of the main reasons of long-term adverse development and even death.Because the lack of clinical symptoms specificity,it is difficult for early accurate identification.But sepsis before onset often appears short heart rate deceleration,blood pressure,blood oxygen,body temperature,breathing and other abnormal physiological data changes.Studies have shown that the final incidence is closely linked with indicators especially the changes in heart rate variability.To achieve early warning of infection through signal processing,algorithm design,with medical diagnosis experience,neonatal doctors can avoid blind antibiotic treatment as well as improve the cure rate,children can also minimize the pain of laboratory tests.The purpose of this study is to investigate the predictive model of sepsis in Chinese infants.Through the algorithm based on sample entropy and optimization of the parameters,the continuous heart rate,ECG sequence should be analyzed,generating a predictive model composed of sample entropy absolute value,change value,global trend and inflection point.Then to analyze multiple cases,for health,infection and cure,infection and death,summarize the characteristics of different model results.The model shows the consistency of heart rate variability no matter of the input is ECG or HR.The healthy infant's absolute value of the entropy is stable,no sharp increase or steep decrease,the trend of the curve was gentle and the range of fluctuation was small.As the time progressed,the amplitude of fluctuation became narrower.For infection and cure of infants,the downward trend corresponds to adverse clinical results,the upward trend corresponds to better clinical results;from the numerical point of view,From the numerical point of view,the curve after the inflection point relative amplitude is greater,drugs and withdrawal are often accompanied by the sample point of maximum absolute value,when the disease or before the onset may also have a minimum absolute value.For morbidity and mortality of infants,before the onset the absolute value usually falls to the lowest point,but also the inflection point of the significant decline.At the time of drug administration,the sample entropy increased with the same time;the overall trend continued to decline until death.