Analysis of knee replacements using data from the National Joint Registry for England and Wales

• Main Author: Baker, Paul Nicholas
• Published: University of Newcastle upon Tyne 2014
• Subjects: 617.5
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.627706

Introduction: Establishing best practice for knee replacement is important given the large number of procedures performed. Research into knee replacement is problematic given that implant failure is a rare event. The logistical and financial costs associated with prospective clinical trials are therefore high. Research using national arthroplasty registers may overcome some of these difficulties. Aim: To assess whether research performed on data recorded by the National Joint Registry for England and Wales has the ability to answer clinically relevant research questions relating to knee replacement surgery. To determine if registry research is able to answer specific clinical questions that are unsuited to prospective randomised clinical trial designs. Methods: Analyses was performed using combined data from the National Joint Registry for England and Wales (NJR) and the Department of Health Patient Reported Outcome Measures (PROMs) project. Results: Nine specific analyses investigated the ability of registry data to ask pertinent clinical questions relating to three areas of practice: unicondylar knee replacement (UKR), total knee replacement (TKR), revision knee replacement (RTKR). Discussion: Registry analyses are well suited to the analysis of rare outcomes such as implant revision and death. In comparison to prospective clinical trial designs they are cheaper, consume less time and resources and have the ability to identify associations and additional factors that may potentially influence outcome. As they use current national data they are more representative of “real-time” national practice and as such overcome some of the problems of generalisability associated with more rigidly designed clinical trials. However, as no information is collected about clinical decision making, drawing strong causal inferences from this type of data is problematic. Conclusion: Using registry data it is possible to answer a range of clinically important research questions. However, due to their limitations, it is necessary to combine information from these observational databases with clinical trial data before robust recommendations that influence clinical practice can be made. The key question researchers have to answer now is how registry data and clinical trial data can be effectively integrated.