The Evidence Based Medicine paradigm (EBM), in which the evidence from randomised controlled clinical trials drives treatment protocols to optimise health while maximising economic efficiency, has revolutionised treatment is many fields of medicine. Treatment resistant psychosis carries a very high social and economic burden, but I will argue that EBM has failed to improve the lives of people with treatment resistant psychosis. The first problem has been the concept of treatment resistant psychosis itself. While it is a clinically recognisable syndrome, it is poorly defined and validated, and is unattractive to pharmaceutical companies as it is imbued with a sense of therapeutic pessimism. Secondly, I will argue that the model of EBM overvalues the findings of randomised clinical trials (RCTs). However, RCTs have several drawbacks which are accentuated in TRP. These include selection bias, short time horizons and outcome measures that are poorly measured, can fail to detect clinically meaningful effects and do not align with patient priorities. Thirdly, non-RCT evidence for the efficacy of clozapine in this population, while compelling, is undervalued by the EBM paradigm. After presenting these arguments, I will make suggestions as to how we can move forward with research in this field.