Effectiveness and efficiency in the treatment of gambling disorder: reflections on the Dodo Bird Conjecture

Problem gambling and its treatment

Despite the health problems associated with gambling, ¹ the gambling business and its adverse impacts continues to grow nationally and internationally. ² In Australia, as in other high income countries, around 2% of the population develop significant gambling problems that affect themselves, as well as people who are close to them.^3–5 Electronic gaming machine losses in Australia amount to over $12 billion a year, with 40% of these losses being experienced by people with recognized gambling problems. ⁶ Such losses lead to significant family and community breakdown, social and psychological problems and suicide. The Coroner's Court of Victoria, Australia, identified 126 suicides between 2000 and 2012 of people who had gambling problems along with two suicides of people who were adversely affected by a partner s problem gambling. ⁷

Problem gambling is classified as an addiction disorder, ⁸ which has led to the creation of a range of treatment programmes and is attracting the interest of different psychotherapy groups eager to provide treatment options. ⁹ Some treatments have been demonstrated to be successful as judged by clinical outcome measures linked to the DSM criteria,^10–12 with Cognitive and Behaviour Therapy (CBT) being the pre-eminent one. ¹¹ These treatment successes notwithstanding, and in spite of many years of research, we are still not much wiser as to which particular components of treatment support improvements in the longer term. Most studies do not deal with well-defined populations, and self-reported outcomes are documented over the short-term only with limited longer term follow-up.

Limitations in the evidence on effectiveness

A recent randomized controlled trial of cognitive therapy (CT) and behavioural or exposure therapy (ET) using a dismantling approach to investigate the core treatment components found no statistically significant differences in outcomes, with both found to be efficacious. ¹³

We still do not know what components of treatment bring about documented improvements or whether a common element, such as a social intervention factor, might have led to the benefits. ¹⁴ Overall, there is little difference between the clinical outcomes achieved through a wide range of prescribed approaches to treatment. ¹⁵ As it is known that non-specific effects of therapy in mental disorders account for most of the variance in outcomes,^15–17 this may also be true for the treatment of gambling disorders.

Re-emergence of the Dodo Bird Conjecture

What are the pathways to success in the treatment of gambling disorders? Does it matter which therapy is used or is the fact that someone is offered a trusted and valued treatment by a caring health professional more important than the actual treatment itself? Does treatment consist of the promulgation of an elaborate myth through which clients are guided to a form of self-awareness that enables them, in varying degrees, to overcome their personal crisis in relation to their gambling problem ¹⁵ even if the clinician is none the wiser as to the processes through which such changes might be wrought? In other words, are we approaching this problem in a sound scientific manner or does the field remain mired in historical bias and subjectivity as is so often shown to be the case in the process of epistemological evolution? ¹⁸

The claim remains that treatments for gambling disorder, across a range of formats, are successful to some extent, even if extent cannot be determined with any confidence. So what are the implications of the claim that all treatments are effective?

Saul Rosenzweig introduced the Dodo Bird Conjecture in 1936 as a construct to focus attention on the application of psychological therapies used to treat people experiencing mental health problems. In the case of psychotherapies, he argued that common factors were more important than specific technical differences, and that all therapies produce effective outcomes for a range of reasons. That is, all therapies are winners (from the Dodo Bird in Alice’s Adventures in Wonderland).

If all treatments are achieving similar results, something very interesting is occurring. Either any kind of intervention is as good as any other or mechanisms outside of the formally documented treatment processes are working to produce improvements in clinical outcomes. If treatments target different pathways to recovery and are predicated upon different theoretical underpinnings, surely it is not realistic to expect that all of these approaches and pathways could be similarly effective unless some of the factors leading to such changes are not those being addressed specifically through the treatment. For example, while it is recognized that paracetamol may help reduce fever, it is not prescribed as a cure for infection. The two problems – managing fever and treating infection – require different approaches matched to the known pathways of both the problem and the cure.

Furthermore, what of the nature of the patient’s condition? ¹⁹ Some people develop their problem as a consequence of social problems or as the result of compensation for personal limitations while others may simply be high-level risk-takers without a reliable regulatory compass at a particular time in their life. ² There are many variables affecting individual responses to these interactions including differences in the way men and women respond to their gambling problems. ²⁰

Given the diversity of backgrounds of people with gambling problems and the different aetiologies, some must be more likely to respond to one treatment rather than another. If a person’s addiction is due to behavioural processing and conditioning, it may be best for them to attempt to remediate their problem through desensitisation or forms of ET. Should their particular problem be due to false beliefs and consciousness (not in the Marxist sense) ²¹ about the possibility of winning, like ‘Granny’ in the novel The Gambler1), such people might respond best to cognitive restructuring. Perhaps, a deft combination of both approaches would be ideal for others with different contextual causes for the development of their problems or substituting their gambling for another form of regular social activity might be the answer for other clients who have drifted into gambling as a result of relationship breakdowns or a spiritual crisis. Every individual may need a tailored programme for their disorder rather than a standard intervention as is the case with patient-centred care for long-term illnesses.^22,23

Two models of psychotherapy

The dilemma of regarding all treatments as equal is encapsulated in Wampold’s delineation of ‘contextual’ and ‘medical’ models of psychotherapy. ²⁴ The former relies on a range of factors for its effectiveness (client context and background, therapeutic alliance, client and therapist belief in the treatment process and social support) while the latter predicts that specific treatment components determine the extent to which any regimen might be expected to be successful. In this context, the general or non-specific effects of treatment are secondary to the specific therapeutic processes that are seen as responsible for the outcome.

The medical model predicts that when the specific ingredients of a treatment are removed, the treatment will be significantly less effective, whereas the contextual model, which does not give primacy to the particular ingredients, predicts that removing one or a few ingredients will not attenuate efficacy. The contextual model also predicts that adding a theoretically crucial ingredient, which is tested with an additive design, will not augment the benefits of treatment. ¹⁵

The challenge of measuring outcomes

It may also be that the ways in which outcomes are measured are not sensitive enough to determine what is changing over time in relation to a person’s problematic gambling. Outcome measures rely on subjective and self-reported information because there is no objective and physiological measure. Outcomes and effect sizes may also be influenced by a Hawthorne effect rather than being a treatment effect. ²⁵ The reported data may be due to changes not in problematic gambling behaviour but in the way that treated individuals perceive and report their understanding of their problematic behaviour as their personal contexts change. Alternatively, the ‘therapy’ process may provide a necessary human connection that genuinely assists in remediating gambling problems irrespective of the nature of the treatment being provided.

Engagement in a form of treatment may also change an individual’s perception of what is problematic about their gambling and perhaps it even changes the truthfulness of the information they provide to therapists and researchers about their condition. In any event, if almost identical outcomes are reported for treatments as different in form and function as chalk and cheese, then it is highly likely that the treatment is masking some other factors that are either leading to a genuine improvement or a change in the way that the outcome is perceived by individuals.^26–28 It is an ongoing challenge for science to disentangle these factors and ultimately sharpen the debate about which factors in this complex treatment process actually contribute to real improvement in outcome.^29,30

At the same time, however, the message from the gambling research community remains that all treatments work and work to a similar extent. ³¹ Such a view undermines the extensive research that has been underway since 1990 and which has been designed to identify and isolate the actual mechanisms of change associated with various treatments. To admit to there being no level of precision at all in relation to treatment efficacy is alarming in the current environment.

A phoenix rising?

There are many helpful approaches to treatment available. To date, researchers have concentrated on dismantling the components of treatments in order to identify the elements that lead to successful outcomes for specific clients,^13,32 but it is also claimed that many treatments have similar outcomes in response to the most researched and documented treatment for gambling disorder, CBT.^11,12,33

In this context, where the impact of treatment can be mainly due to non-specific effects and where the tools for assessing and diagnosing disorders and measuring outcomes are blunt, there is a tendency for interventions to rely on the ‘black box’ effect. This is resulting in a re-emergence of the Dodo Bird Conjecture in relation to treatment for gambling disorder and an acceptance in the field that one particular treatment is as useful as any another.

Given the uncertainty about treatment effectiveness, and the re-emergence of the Dodo Bird Conjecture, a more useful tack would be to focus on the cost of different treatments while further research is carried out to identify key clinical impacts of treatment. That is, we should be seeking to discern the most cost-effective treatment for gambling disorder while, at the same time, continuing to focus research on the components of therapy and treatment effectiveness in an effort to dispatch the Dodo to its proper place in the scheme of things. This would allow us to hasten slowly towards the overall goal of identifying the effective components of treatment through the deconstruction and analysis process while providing as much positive support to people with gambling problems as possible, within existing funding constraints, by applying the treatments that are most cost effective.