Aram P, et al. - In view of the significance of tools providing personalized risk prediction of the outcomes after surgical procedures for patients to make preference-based decisions amongst the available treatment options, researchers constructed and compared several parametric and non-parametric models for predicting prosthesis survivorship after knee replacement surgery for osteoarthritis. Using 430,455 patient-procedure episodes between April 2003 and September 2015 from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man, they found the flexible parametric model as superior in overall performance compared to other tested parametric methods, and to have better discrimination compared to the random survival forest approach.