Abstract
OBJECTIVES: To analyze clinical risk models using epidemiological surveillance and hospital data (2022-2024) in Cali, Colombia to identify drivers of respiratory failure in patients hospitalized with severe acute respiratory infection. METHODS: A cross-sectional analytical study was conducted using secondary data from notification file 348, which is part of the Colombian National Public Health Surveillance System. A total of 2 354 hospitalized cases of severe acute respiratory infection were analyzed after quality control and exclusion of inconsistent records. Bivariate and multivariable analyses were performed using chi-squared tests and generalized linear models, respectively, to determine the factors associated with respiratory failure (outcome variable). RESULTS: Among all cases, 7.3% had respiratory failure. Not having health insurance and having a high multidimensional poverty index score were significantly associated with respiratory failure (both p<0.05). Co-morbidities such as asthma (p<0.01), chronic obstructive pulmonary disease (p<0.01), heart disease (p<0.05), and cancer (p<0.05) also significantly increased the risk of respiratory failure. Sepsis was the only complication significantly associated with the outcome (p<0.01). The final model had good sensitivity (81.9%) and several high-risk profiles were identified. CONCLUSIONS: The study highlights the relevance of co-morbidities and social vulnerability in the development of respiratory failure among cases of severe acute respiratory infection. These findings support the implementation of clinical audit tools and targeted risk management strategies to improve timely care and reduce adverse outcomes in high-risk populations.