Commentary on Nower et al: The Pathways Model should apply to non-clinical gambling patterns.

Nower et al. [ ] report new data on the pathways model of problem gambling that substantiate the original 3 ‘subtypes’ and clarify some nuances of the original model. Despite a strong assertion that the model was intended to describe clinically relevant heterogeneity among those with gambling problems, we suggest this neo‐Kraepelinian assumption is superseded by contemporary research showing a continuum of gambling problems.

Since its original publication in 2002, the pathways model [1] has become a major framework for understanding the aetiology of problem gambling. This model quickly became influential, driving much clinical research into the subtyping and heterogeneity of problem gambling. In turn, this research profoundly impacted the treatment of gambling disorder, as many clinicians in the field tailor their interventions according to gambling subtypes. Of further significance, the pathways model integrates an array of etiological factors (biological, psychological and environmental) that are hypothesised to underpin transitions from casual to problematic gambling. In their latest study using a large sample of treatment‐seeking gamblers, Nower et al. [2] describe further evidence for their three subgroups, but their analyses also indicate some revisions to the model. Notably, the ‘antisocial‐impulsivist pathway’ (pathway 3) was clearly distinct from the ‘emotionally vulnerable’ pathway (pathway 2), whereas in the original model, the pathway 3 liabilities were conceptualized as additive upon pathway 2.

The recent paper [2] makes a strong assertion that the pathways model is intended to classify clinically relevant gambling patterns. Previous studies that either tested the model in non‐clinical groups (or mixed samples including only a minority with gambling problems) or used statistical approach beyond cluster or latent profile analyses are said to have misinterpreted or misapplied the model. As the originators of the model, Nower and colleagues [2] are entitled to say that the model was intended to describe clinically relevant gambling patterns, but it is an empirical question to what extent these factors are also manifested across the broader spectrum of gambling involvement. Their assertion implicitly adopts a neo‐Kraepelinian perspective [3] of a clear boundary between the ‘normal’ and the ‘pathological’. This traditional viewpoint has been challenged and largely superseded by dimensional approaches to psychopathology [4, 5]: in this case, a continuum of gambling involvement that further justifies the study of ‘normal’ individuals (i.e. healthy gamblers) to understand the etiological processes of disordered gambling [6, 7]. As a prototypical example, psychotic experiences such as hallucinations and delusions are common among individuals who do not reach a diagnostic threshold or suffer from clinically relevant functional impairment [8, 9]. Disease categories have been particularly contested in the case of personality disorders [10, 11], which is notable given that antisocial personality disorder is a feature of pathway 3. In fact, a large proportion of the evidence in psychopathology research result from studies conducted in the general population [5], and this point applies equally well to the field of gambling studies.

From a data‐analysis perspective, the pathways model has inspired a subfield of research looking to characterize the heterogeneity among gamblers with profiling approaches such as cluster analysis [12, 13] or latent class analysis [2, 14]. When applied to the pathways model, these procedures generate some specific issues. One pertains to the degrees of freedom that exists in supporting (or refuting) the pathways model. In principle, profiling approaches conducted with the relevant pathways variables should indicate that a 3 class solution provides best fit to the data. In practice, determining the number of classes results from a combination of goodness of fit statistics and theory, which may increase the likelihood of favouring 3 class‐solutions. In reality, a common scenario is for a profiling technique to identify more than 3 clusters as the optimal solution, where those clusters align with the pathways via a range of possible mappings [12, 14]. These techniques can also generate superficial classes, such as subgroups who score in the low (or high) range on all variables [15].

Mindful of both of these points, we assert that there is a need for alternative research designs and data‐analytic approaches that can characterize key factors present in the pathways model in a way that acknowledges both their dimensional nature (from non‐problematic to problematic gambling) and their heterogeneous expression. Expanding the remit of the pathways model is underscored by the very low rates of treatment seeking in people with gambling problems [16]. Progress would also incorporate statistical approaches such as regression models or network analytical approaches [17] that do not necessarily assume gambling pathways to be discrete and categorical entities, and lab and field studies relying on cognitive, emotional, behavioural and computational approaches.

DECLARATION OF INTERESTS

J.B. and C.B. have no disclosures. L.C. is the Director of the Centre for Gambling Research at University of British Columbia (UBC), which is supported by funding from the Province of British Columbia and the British Columbia Lottery Corporation (BCLC), a Canadian Crown Corporation. The Province of BC government and the BCLC had no role in the preparation of this commentary and impose no constraints on publishing. L.C. has received a speaker/travel honorarium from the National Association for Gambling Studies (Australia) and the International Center for Responsible Gaming (United States [US]), and has received fees for academic services from the International Center for Responsible Gaming (US), GambleAware (United Kingdom [UK]) and Gambling Research Exchange Ontario (Canada). He has not received any further direct or indirect payments from the gambling industry or groups substantially funded by gambling. H.B.J. is the Director of the National Problem Gambling Clinic funded by the National Health Service (NHS) and GambleAware. She is also the director of the NHS Young People's gambling services funded by the NHS.