How can we (and why should we) develop better models of recovery?

Abstract

Why is it that we know so little about addiction despite a vast quantity of research and theorising? Why is it so striking that our research-derived knowledge often fails to predict people's drug-using and other addictive behaviour and their response to treatment interventions, certainly in any long-term way? Having recently been involved in the development of National Institute of Health and Clinical Excellence (NICE) guidelines for psycho-social interventions for substance misuse, I was forcefully reminded that research findings typically point in different directions, and do not convincingly generalise, and that attempts to synthesise findings through systematic review and meta-analysis produce for the most part very modest combined effect sizes and reveal considerable heterogeneity. A tough minded Dodo bird might be inclined to announce that nobody has won and nobody shall have prizes, except for a few old friends like Cognitive Behaviour Therapy (CBT), who always get a small medal just for turning up, and there might be a special prize for the tall enigmatic stranger Contingency Management.There is clearly a wide range of possible responses to such questions, but it might be useful to explore how inadequate theory, combined with an orthodox and restrictive hierarchy of evidence, produces research which fails to tell us anything meaningful and at worst perpetuates inaccurate ideas. One problem lies with the development of hypotheses which rely on accepted opinion, when this opinion is based on serious misunderstanding or is out of date (Marshall 2003).It seems as though outdated models are quite tenacious and play a considerable part in blocking the advancement of our understanding. The picture is not entirely bleak: all but one of the major matching hypotheses of Project MATCH were not supported, but the causal chain analyses conducted afterwards have clarified the glaring gaps in the