Response to McKay: The stakes are high


We are pleased to be in the company of a commentator [1] who has argued previously for expanding the scope of the existing efficacy paradigm [2, 3]. Our succinct response is that the bar is high because the stakes are high. The stakes are high because while the explosion of clinical trials from the 1980s to the present has resulted in a range of moderately effective addiction treatments (e.g. [4-7]), moderate effectiveness calls for improvement and scarce resources call for efficiency. In addition, the absence of clear comparative efficacy of one treatment over another suggests that developing and testing new treatments is no longer yielding substantive knowledge on how to optimize treatment delivery [8]. The key question here is: when the stakes are high, should the direction for the future be more of the same?

A paradigm shift in both research and practice that places tests of a treatment's causal theory at the same level of importance as tests of treatment efficacy is something different. It requires that not just the treatment, but the treatment's essential elements be validated in order to achieve the highest level of empirical support. Dr McKay identifies a number of sound arguments on why it may be difficult to specify active treatment ingredients. We agree that therapy process probably works in complex and multi-determined ways. However, this complexity argues for an incremental approach to process research, ‘peeling layers’ of ingredients and mechanisms and the contexts in which they are operative [9]. Indeed, research methods in the addictions often reduce the therapy process to the effect of an intervention on a client response, and the effect of that client response on an outcome: in essence, a unidirectional mediation model. This mediational process probably accounts for a portion of causality, but we recognize that more nuanced analytical methods are also needed. These nuances may be at the level of speaker interaction or characterizing the nature of ‘sudden gains’ in addiction treatment, as Dr McKay suggests. Along similar lines is Greenberg's method of task analysis, which emphasizes an event-based approach to identifying key moments in psychotherapy that result in new learning outside the therapy context [10]. In sum, we agree that behavioral addiction treatments work in complex ways, which is the very reason further understanding is needed, and only through that understanding can deeper and more complex methodological frameworks be incorporated.

Dr McKay also highlights the role of research on common processes of change and queries the implications of evidence demonstrating that the majority of patient change is non-specific to a given modality. Such evidence may be an unfortunate fall-out of the clinical trial paradigm. Specifically, it may be that many treatments have only ‘artificial uniqueness’ and thus fail in tests of comparative efficacy. Many treatments may also impact common client mechanisms of change [8]. Common therapy processes are common because they matter. It is the task of treatment developers to show where their treatments rely upon common ingredients and where they incorporate truly unique interventions that are active ingredients and thus critical to the model; or perhaps a future treatment will demonstrate differential efficacy because it has a unique combination of efficient, yet entirely common, ingredients. Yes, active ingredients should account for a meaningful portion of client change, but identifying this portion may first require a far better understanding of changes produced via assessment and control group exposure [9].

Therapy is indeed very complicated and unpredictable, yet it is only through intensive and sustained study of treatment processes, active ingredients and client mechanisms of change that it will ever become less complicated and more predictable. We are not overly optimistic, but rather pessimistic that this will happen unless the noted paradigmatic shift is located not only in research, but also in our service infrastructure.

Declarations of interest



The present work is supported by a grant awarded by National Institute on Alcohol Abuse and Alcoholism (NIAAA; K23AA018126) to Dr Molly Magill. However, both authors have contributed equally to the completion of this manuscript. This work is the responsibility of the authors and reflects the positions of the authors, and does not represent an official position of NIAAA or the National Institutes of Health.