Heart Failure (HF) is a condition where the heart is unable to pump sufficient blood to meet the body's needs. When the body cannot compensate for HF symptoms, patients require urgent and intensive treatment to prevent fatal outcomes. In this study, we investigated features related to rhythm stability and ventricular activation stability and their association with HF patient mortality. We utilized data from 10,800 hospitalized HF patients from the MIMIC-IV ECG and MIMIC-IV datasets. All records were automatically processed to extract information about specific QRS complexes and heart rhythm. We evaluated 17 extracted features for their ability to differentiate patients' mortality within a 3-year follow-up using survival analysis. Next, we employed Lasso regression for feature selection and to build a multivariate model to predict patient outcomes. Finally, we determined hazard ratios (HR) using the log-rank method for data split by the model. Our results revealed a high association between mortality and features describing the presence of other QRS morphologies and RR-interval instability (strongest at RR-interval variation range, χ2 68.1, p<0.0001). However, none of these features surpassed the differentiating ability of age (χ2 476.7, p<0.0001). Lasso regression (using training subset, N=6,256) automatically selected 7 features and produced a model. The effect of combined features for mortality prediction in the test set (N=4,172) resulted in an HR of 2.26 (95% CI 2.03-2.52, p<0.0001), while the effect of age alone resulted in an HR of 2.09 (95% CI 1.88-2.33, p<0.0001). This study demonstrates that instability in heart rhythm and ventricular conduction is significantly associated with mortality in HF patients. However, the strongest predictor remains the patient's age.