Can Statistics give Better Understanding of Existing Clinical Data?
Statistics can be used to reveal patterns and relationships within data that may be unknown to the clinician. By looking at several large orthopaedic data sets, novel methods for approaching these data sets have been developed. Results from these methods will be presented with specific relevance to knee and hip replacement. New visual techniques allow an intuitive visualisation of the progress of different patient groups on different variables over time. Variable reduction strategies can identify a small group of key variables which convey the majority of the information of the data allowing the focus to be placed on these important variables. Finally, the techniques of graphical modelling and chain graphs allow for the development of complex statistical models which can be easily interpreted and provide valuable information about the inter-relationships within the data.
