Developments in motion analysis technology over the last two decades have enhanced our understanding of human locomotion. However, such advances in knowledge are futile if no practical use is made of them. Scientists and engineers need to make the most of these developments by forging stronger links with orthopaedic surgeons and applying further advances in their knowledge to clinical problems for the long-term benefit of patients. This need has been identified by many in the field of biomechanics and a “serious attempt [has been made] to take gait analysis out of the research laboratory and into the clinic” (Whittle, 1996 pp.58). For this reason, the aim of this research is to develop an objective and quantitative classification tool that uses motion analysis to aid orthopaedic surgeons and therapists in making clinical decisions. Practical applications of this tool would include joint degeneration monitoring; diagnostics; outcome prediction for surgical intervention; post-operative monitoring and functional analysis of joint prosthesis design. The classification tool (Jones, 2004), based around the Dempster-Shafer theory, is logical and visual; as the progression from obtaining clinically relevant measurements to making a decision can be clearly followed. The current study applies the tool to identify knee osteoarthritis (OA) and post-operative recovery following total knee replacement (TKR) surgery. Knee function data from 42 patients (22 OA and 20 normal (NL)) were collected during a clinical knee trial (Holt et al., 2000). Nine of the OA patients were followed at 3 stages following TKR surgery. Using the tool, a subject’s knee function data are transformed into a set of belief values: a level of belief that the subject has OA knee function, a level of belief that the subject has NL knee function and an associated level of uncertainty. These three belief values are then characterized in a way that enables the final classification of the subject, and the variables contributing to it, to be represented visually. Initial studies using this technique have provided encouraging results for accuracy, validity and clinical relevance (Jones, 2004). The tool was able to differentiate between the characteristics of NL and OA knee function with 98% accuracy. The belief values and simple visual output showed the variation in the extent to which patients had:

developed OA and;

Furthermore, the visual output enabled straightforward comparison between subjects and indicated the variables that were most influential in the decision making process for comparison with clinical observations and quality of life scores. The tool is generic, and, as such, would be applicable to a wide range of pathological classification and predictive problems.

Results Holt, C.A. et al. (2000). Computer Methods in Biomechanics and Biomedical Engineering 3. Lisbon. Gordon and Breach Science Publishers SA. pp.289–294. Jones L. (2004). The development of a novel method for the classification of osteoarthritic and normal knee function. PhD Thesis. Cardiff University Whittle, M.W. (1996). Gait analysis: an introduction. 2nd Edition. Oxford; Boston: Butterworth-Heinemann.