The new Medical Device Regulation (MDR) addresses patient safety by mandating supporting clinical evidence for the sale and clinical use of medical devices. However, it places pressure on medical device manufacturers to supply regulators with robust clinical evidence to obtain regulatory approval of new devices they develop. Regulatory compliance is therefore more stringent, requiring more effort on part of both regulators and manufacturers. This thesis project explored the possibility of developing a new method to be used as a regulatory decision-making tool for the approval of newly developed knee implants. A Bayesian probabilistic method was developed and proposed. This method utilises existing arthroplasty registry and arthroplasty registry annual report data to determine if a newly released implant variant is “good”, where a “good” implant is one which meets standard regulatory thresholds in term of revision risk. This method was implemented on two datasets which were selected based on conditions of granularity, measures of implant performance, and accessibility for the thesis project. Some advantages and limitations of this method as a regulatory tool were proposed, the limitations being based on a clinical appraisal framework. Based on this it was concluded that the method indeed showed potential as such a regulatory tool, but needs further improvement due to considerations of reliability, precision, and the probabilistic nature of the results it yields. Both surgeons and patients can also potentially stand to benefit from this method. Aside from the regulatory context, also, the method has limitations in the fact that it does not account for other factors that may be responsible for complications related to knee arthroplasty surgery. Overall, however, this thesis presents a novel approach towards estimating the performance of knee implants from existing clinical data.