Estimating time-to-death and determining risk predictors of heart failure patients

Congestive heart failure (CHF) is one of the global health issues affecting millions, and is a leading cause of hospitalization worldwide. Despite extensive research, accurately estimating survival times, and identifying risk factors for CHF patients remains an ongoing challenge. To overcome this challenge, the Accelerated Failure Time (AFT) model is employed to identify key prognostic factors for CHF using retrospective data from Saint Peter's Specialized Hospital. Specifically, the assumptions of the Cox model were checked, and it was found that they were violated. This led to the comparison of AFT models to identify the one that provides better parameter estimates. The Akaike Information Criterion (AIC), and the Bayesian Information Criterion (BIC) were used to evaluate model performance. In this dataset, the Weibull AFT model decisively outperformed the Cox model, providing the best fit with the lowest AIC/BIC, highest log-likelihood, more accurate parameter estimates, and reduced bias. The key prognostic factors identified includes pneumonia, diabetes mellitus, and electrolyte imbalance, which are strongly associated with reduced survival in CHF patients. These findings lead to the need for urgent, targeted interventions, and establish the Weibull AFT model as a robust tool for analyzing complex survival data.