Association between incubator standards and newborn nosocomial infection with machine-learning prediction.

Background: Newborns have a high incidence of nosocomial infection (NI). We conducted a logistic regression to analyze different incubator standards and other risk factors for newborn NI, which could better help clinical choice of incubator standard.

Methods: Newborns with complete necessary clinical data were included. We collected the demographic and incubator data of 76 patients (40 uninfected and 36 infected) at the Heping Hospital Affiliated to Changzhi Medical College. An analysis of variance, Pearson correlation matrix analysis, and logistic regression analysis were conducted to explore the different incubator standards and other risk factors for neonatal hospital infections. In addition, 4 machine-learning algorithms were used to predict neonatal hospital infections.

Results: We found differences in the gestational age, incubator type, paternal age, and maternal age between the 2 groups. The correlation analysis only revealed a correlation between paternal age and maternal age. The logistic regression showed that gestational age [odds ratio (OR)= 0.77574, 95% confidence interval (CI): 0.583513-0.996354] and the new standard incubator (OR =0.011639, 95% CI: 0.000958-0.067897) may be protective factors for infant infection during hospitalization. Among the extreme gradient boosting (XGBoost), random forest (RF), support vector machine (SVM), and decision tree (DT) algorithms, XGBoost had the best performance in terms of accuracy, sensitivity, specificity, and precision.

Conclusions: We found early gestational age and incubator standards may be risk factors for the NIs of newborns, which might help clinicians to improve the health and safety standards for incubators. XGBoost can be used to predict newborn NIs.