AIMS: This retrospective analysis sought to develop and validate a model using the measured diagnostic variables in cardiac resynchronization therapy (CRT) devices to predict mortality. METHODS AND RESULTS: Data used in this analysis came from two CRT studies: Cardiac Resynchronization Therapy Registry Evaluating Patient Response with RENEWAL Family Devices (CRT RENEWAL) (n = 436) and Heart Failure-Heart Rate Variability (HF-HRV) (n = 838). Patients from CRT RENEWAL were used to create a model for risk of death using logistic regression and to create a scoring system that could be used to predict mortality. Results of both the logistic regression and the clinical risk score were validated in a cohort of patients from the HF-HRV study. Diagnostics significantly improved over time post-CRT implant (all P < 0.001) and were correlated with a trend of decreased risk of death. The regression model classified CRT RENEWAL patients into low (2.8%), moderate (6.9%), and high (13.8%) risk of death based on tertiles of their model predicted risk. The clinical risk score classified CRT RENEWAL patients into low (2.8%), moderate (10.1%), and high (13.4%) risk of death based on tertiles of their score. When both the regression model and the clinical risk score were applied to the HF-HRV study, each was able to classify patients into appropriate levels of risk. CONCLUSION: Device diagnostics may be used to create models that predict the risk of death.