<![CDATA[Introduction. Community Acquired Pneumonia (CAP) is a lung parenchyma infection acquired outside of the hospital and is significantly associated with morbidity and mortality rates. Identifying patients with moderate-severe CAP at high risk of mortality through a combination of these variables is expected to serve as a basis for prompt and appropriate intervention, ultimately improving clinical outcomes for CAP patients. This study aimed to develop a scoring system using clinical characteristics, radiological findings, and serum biomarkers to assess the mortality risk in patients with moderate-severe CAP. Methods. This study employs a retrospective cohort design with moderate-severe CAP patients treated at Dr. Ciptomangunkusumo Hospital, Jakarta. The sample data were secondary, derived from medical records of patients diagnosed with moderate-severe CAP between January 2022 and December 2023. Predictor variables for mortality risk were obtained through multivariate analysis using Cox regression. Results. The study included 277 subjects, with 124 (44.77%) deaths and 153 (55.23%) survivors. Predictor variables consistently influencing mortality risk in moderate-severe CAP patients were low BMI (HR 1.609, 95% CI 1.047 – 2.472), procalcitonin (HR 1.778, 95% CI 1.200 – 2.634), and lactate levels (HR 1.451, 95% CI 0.994 – 2.119). The prediction model’s performance, based on the ROC curve analysis, showed moderate predictive ability (AUC = 0.641), and calibration performance, as assessed by the Hosmer-Lemeshow test, showed good validation (p = 0.082). Conclusions. There is an association between BMI, procalcitonin, and lactate levels with the mortality risk in moderate-severe CAP patients. A mortality risk scoring model for moderate-to-severe CAP patients has been established.]]>
