There have been considerable criticisms of the CGM over the years. Many of the criticisms are valid but it is also important to recognise the considerable strengths. Clearly any attempt to address the weaknesses will only be beneficial if this is done in such a way that the strengths are presereved.


It is essential that modern clinical services work within a strict clinical governance framework. Amongst other things this requires services to implement tried and trusted techniques. In the light of subsequent developments the earlier papers of the Helen Hayes pioneers in particular (Kadaba et al. 1989, 1990; Ramikrishnan et al., 1991 were ahead of well ahead of their time in model validation. The papers included studies of repeatability, provision of normative reference data (and comparison of this against previous publications) and a sensitivity analysis of the most common measurement artefact (that which arises from misplacement of the thigh wands).

The CGM is also one of the few models to have had been subjected to serious validation work outside the centres in which it was developed. 15 of the 23 papers included within the classic systematic review of repeatability studies of kinematic models (McGinley et al. 2009) used a variant of the CGM and Pinzone's more recent study demonstrated that two of the most respected clinical services, in the USA and Australia, have collected essentially similar normative reference data using the model.

As well as a considerable volume of formal validation data the CGM is also backed by over 35 years of use in some of the most respected clinical services across the world and there is a considerable body understanding of its clinical application within this user community. This is further consolidated by the volume of academic literature generated by users of the model (the key paper of Davis et al. has now been cited over 1700 times and that or Kadaba et al. over 1900 times).

Simplicity and comprehensibility

The extent of these pages should make it clear that the CGM is not trivial but it is about as simple as possible for clinically useful model. The model is deterministic and thus the effects of marker misplacement and soft tissue are entirely deterministic. It is thus possible for people with a sufficient understanding of the model to think through how it will behave in a range of different circumstances. This understanding can also be used to suggest how the model can be best adapted to give clinically useful outputs in patients who might have a range of bone or joint deformities or be wearing a range of prostheses or orthoses.

The CGM has acquired a unfair reputation for being poorly documented and difficult to understand. It is possible that this stems from the early developers only over describing the model verbally rather than releasing the source code. That description was assumed to be only partial but the emergence of a number of exact clones at a fairly early stage ([Baker et al, 1999](( is an example) falsify that assumption.

Proceed to analysis of weaknesses.