Treatments for the spine still lag behind other orthopaedic interventions in terms of their technical maturity and success rates. The use of computational models for pre-clinical testing of new spinal devices is attractive but, as yet, these models have not reached their full potential for the evaluation of treatments across a patient population. This chapter examines the development of subject-specific models of the spine with a particular focus on the analysis of fracture fixation using vertebroplasty. The development of specimen-specific models allows direct comparison to be made between the model predictions and the behaviour of the same specimens in laboratory tests, thus allowing the features and sensitivity of the models to be thoroughly examined. The progress made in this field is examined and key results presented. Current work and future challenges are also discussed in the context of extending the technique to simulate the performance of new treatments across a population.