Mixture modeling for pulmonary thromboembolism; A latent class analysis study

Abstract

Acute pulmonary embolism (PTE) is a major cause of death and severe disability. There are many clinical parameters to evaluate the prognosis of PTE. The most common is the Pulmonary Embolism Severity Index (PESI). Also, the availability of some blood parameters can also help us determine the prognosis of PTE. This study was designed and implemented to perform a modeling for PTE mortality risk factors with the help of latent class method. Methods The current study was a cross-sectional study. All PTE patients who were admitted to Shahid Madani Hospital and their information was recorded in the PTE registry of this center and also died during hospitalization or underwent thrombolytic therapy were included in this study after obtaining the necessary ethical patient information was extracted from the PTE registry system and were analyzed. Modeling was done with the help of Mplus software and other analyzes were done with the help of SPSS software. Results In general, 263 patients were included in this study. 49.4% of them were men and 50.6% of them were women. The average age of the patients was 69.30 Finally, a model with two hidden classes was selected that obtained the highest percentage of prevalence, the highest percentage of change and the best fit for lower BIC (9.4158) and higher entropy (0.986). The classes were named with low risk profile and high risk profile labels. Patients who were members of the high-risk class had significantly lower hemoglobin, higher systolic blood pressure, and more history of heart failure. Finally, according to the logistic regression results, the determined high-risk class had a significant relationship with the mortality of PTE patients.