Validation of musculoskeletal models: the importance of trend validations

Introduction Validation of musculoskeletal models is an important issue and is getting more and more attention, which is also apparent in this symposium. Most probably the reasons for this huge attention for validation are advancements in scientific research, but also by the appearance of more general commercial and open source musculoskeletal modeling systems like Lifemodeler ("LifeModeler, 2010), OpenSIM (Delp et al., 2007), and AnyBody (Damsgaard, Rasmussen, Christensen, Surma, & de Zee, 2006), which makes musculoskeletal modeling available to science and industry. We are at a transition from science to clinical applications where erroneous predictions by the models could have serious consequences. However, compared to some other areas in engineering there are no formalized processes of Verification & Validation (V&V) in musculoskeletal modeling. Only recently a review article has attempted to describe V&V and sensitivity studies in relation to computational biomechanics (Anderson, Ellis, & Weiss, 2007), though the focus was solely on finite element modelling and tissue engineering.