A clinician's critique of rheumatoid arthritis health economic models.

Modelling cost-effectiveness of new drugs for RA has become increasingly prevalent and sophisticated. This situation has arisen largely because regulatory agencies, such as the National Institute for Health and Clinical Excellence, have demanded models from industry and have commissioned independent models. Many technical aspects of health economic models have converged-yet the results of models differ greatly. These differences can be accounted for in large part by differences in assumptions about the nature of patients likely to be treated; likely treatment sequences; likely responses to treatment; likely continuation on drug and likely disease progression, in particular. Such parameters cannot be fixed and evolve with changing practice and are ideally captured by contemporary data. Importantly, data from the local setting to which a health economic problem is applied are necessary, but in the absence of ideal sources, for the many contributions needed, considerable differences in opinion and biases are commonplace. In the final analysis, all models are just that, models, and as such an approximation of real life. Thus, although considerable heat is generated during debates about model parameters, model outputs may just yield sufficient light for regulatory agencies allocating resources.