Screening and comprehensive prediction model construction and validation of serological diagnostic indicators for elderly patients with stroke-associated pneumonia

Objective: To screen the serological diagnostic indicators of elderly patients with stroke-associated pneumonia (SAP), and analyze the construction and validation of comprehensive prediction model. Methods: A total of 91 elderly SAP patients were selected as SAP group, and stroke patients without SAP were selected as non-SAP group.The clinical related indicators and serological indicators of the two groups were compared, and the risk factors of elderly SAP were analyzed, and the comprehensive prediction model of elderly SAP was constructed through the risk factors. Results: Compared with non-SAP group, the proportion of dysphagia, hypoproteinemia, interleukin-6, procalcitonin, C-reactive protein, neutrophil to lymphocyte ratio were higher in SAP group (P < 0.05 to P < 0.01).Dysphagia, interleukin-6, procalcitonin, C-reactive protein and neutrophil to lymphocyte ratio were the risk factors for SAP in the elderly (P < 0.05 to P < 0.01).The AUC of ROC curve was 0.915, the Youden index was 0.714, the sensitivity and specificity were 82.4% and 89.0%, respectively, the 95%CI was 0.873-0.956, and the actual application accuracy was 84.6%. Conclusions: The comprehensive prediction model constructed through clinical relevant indicators and serological indicators has a high predictive value for the occurrence of SAP in the elderly.