Problems with Models in Psychiatry

Over the past decade or two, psychiatry has largely adopted scientific ways. Due to intellectual and budgetary pressure, operational diagnoses and randomised trials now prevail. These methods are demonstrably reliable and useful; but often appear far removed from clinical practice, where patients need to be treated as individuals, and idiosyncratic details can be the key to therapeutic success. There are continuing concerns about the failure of research methods to convincingly capture such complexity (Nunn et al. 2000; Szasz 2001; Timimi 2004; reviewed by Kandel 1998; and Midgley 2004). This essay argues that bridging the gap requires researchers not just to incrementally improve models of mental function, but to become far more rigorous in the ways we create and use such models. Here we are using the term model as a convenient shorthand for simplifi