Abstract:

Objective To construct a prediction model for the length of hospitalization in children with pneumonia based on clinical characteristics. Methods A retrospective analysis of the clinical data of 1 255 children with pneumonia was conducted. The patients were divided into two groups based on the median length of hospitalization: the ≤7 days group (628 cases) and the >7 days group (627 cases). The differences between the two groups in demographic characteristics, past medical history, clinical manifestations, laboratory test results, imaging findings, treatment plans and other clinical data were compared. A multivariate stepwise Logistic regression analysis was performed to identify the factors influencing hospitalization for >7 days and to construct a prediction model. The model was evaluated using the receiver operating characteristic (ROC) curve and the clinical decision curve. Results Compared to the ≤7 days group, children in the >7 days group were younger in gae, had higher height, a higher proportion of preterm infants, a higher proportion of previous pneumonia history, and a higher body temperatures at admission. Furthermore, in the >7 days group, white blood cell count, neutrophil count, platelet count, C-reactive protein (CRP) and procalcitonin levels were elevated. The proportion of bilateral lesions, oxygen therapy, respiratory support and pleural effusion were higher, while lymphocyte count and hemoglobin levels were lower (P < 0.05). The results of the multivariate Logistic regression analysis showed that age (OR=0.979, 95% CI: 0.972-0.987), history of prematurity (OR=1.751, 95% CI: 1.216-2.521), previous history of pneumonia (OR=1.520, 95% CI: 1.037-2.228), admission temperature (OR=1.290, 95% CI: 1.097-1.518), serum CRP (OR=1.019, 95% CI: 1.013-1.025), pleural effusion (OR=1.980, 95% CI: 1.309-2.994) and oxygen therapy (OR=2.849, 95% CI: 1.851-4.385) were independent risk factors for a hospital stay >7 days in children with pneumonia. The model had an accuracy of 79.2%, and the area under the curve (AUC) was 0.919 (95% CI: 0.854-0.961). Conclusion The regression model constructed based on clinical characteristics can effectively predict the length of hospitalization in children with pneumonia. It provides scientific evidence for the early identification of high-risk children, optimization of treatment plans and shortening of hospital stays.

Key words: child, pneumonia, length of stay, Logistic models, prediction model