This article, written by Fojan Faghiri and Akram Kohansal, tackles the issue of survival prediction by combining the traditional Cox proportional hazards model with the newer Bayesian neural network model. The hybrid model (BDPLCM) uses the Cox approach for linear effects while the Bayesian neural network handles the nonlinear relationships and also quantifies the associated uncertainty to predict survivorship over time. Kohansal and Faghiri found that the BDPLCM outperformed several existing models including the standard Cox model, Partially Linear Additive Cox Model (PLACM) and Deep Partially Linear Cox Model (DPLCM), as shown by the Cordance index which analyzes how well a model predicts survival. Even though the training time was longer than traditional methods, the accuracy and reliability of the BDPLCM model outweighed the cost.