Total knee replacement benefits patients who suffer from severe knee pain or joint stiffness and other joint related illnesses that limit everyday activities. There has been an increase in the number of procedures performed each year and a need to evaluate the performance of these implants during specialized activities such as kneeling. Most computational studies lack insight into inter-patient variability and the results do not apply to large population. This study developed:​ (1) three-dimensional explicit finite element (FE) models to investigate natural and implanted knee joint kinematics and bone strain and (2) a platform to enable population-based evaluation by combining statistical model and joint function. Verification of a finite element model confirmed a strong agreement between model predicted and in-vitro kinematics of specimen-specific patellofemoral (PF) joints of four cadaveric knees in simulated kneeling. Three different commonly used PF implants were employed in an additional broader patellar bone strain study to assess the relative performance of these implants during highly demanding activities. This study predicted that the medialized dome design achieves the optimal balance of sufficient congruency between PF articular surfaces while still facilitating sagittal plane tilt to reduce isolated loading of the distal nose of the patella. A combined statistical shape model and FE method were utilized to successfully identify the most important shape characteristics affecting joint performance during kneeling. Scaling in the knee joint has minimal effect on PF joint kinematics but greatly affects joint contact mechanics. Knee soft tissue dimensions alter the kinematics. The patellar bone strain model described here provides a novel platform for further implant performance analyses. The statistical shape-function model is a tool for population based studies to help predict the clinical outcome of joint replacement