Many-models medicine: diversity as the best medicine

Many medicines have been proposed to cure various ills of biomedicine including evidence-based medicine, evolutionary medicine, narrative medicine, and complexity medicine, among others. To the extent that all models are idealizations or abstractions, all of these model medicines are imperfect in some respects. In the absence of a single unified model, if indeed unification is possible or even desirable, and despite the relative advantages of one model or another, in practice many models and methods are necessary in medicine. In this article, I consider the value of such diversity in models and methods. I briefly describe several models. Then I discuss simulations of agents who use diverse models. Advocates of models such as those discussed here typically claim that we should use their preferred model because it is the best. Evidence-based medicine, for instance, has been promoted as the single best model of medicine while other models have been cast as lesser models or in opposition to it and each other. But isolated models and methods may never be as good as groups of models and methods. Debates about various individual models may result in better outcomes, but explicitly choosing to use many models is likely to produce even better outcomes.