Commentary on Karhulahti et al. (2022): exploring gaming disorder from the harmful dysfunction analysis perspective

I read with great interest the well-designed pre-registered study of Karhulahti et al. (2022). Their findings pose an important question regarding the ability of different selfreport instruments (Lemmens et al. 2009; Kir aly et al. 2017; Pontes et al. 2021) to converge in the identification of individuals with gaming disorder (GD) as shown by different prevalence rates and overlap of participants with GD. The authors concluded that “the scales of addictive gaming behaviors—standardly studied as a single construct—seem unable to identify mutual groups with shared problems” (p.9). The reply of Billieux et al. (2022) further discussed those results and, encouraged by and relying on the open science approach of Karhulahti et al. (2022), conducted a network analysis using all items of the instruments assessing GD and concluded “that the various GD measurement tools are not reliably distinct and that their content strongly overlaps” (p.2). Billieux et al. (2022) considered that all four instruments examine GD, despite doing so based on somewhat different approaches, criteria and, thus, item content (and response modality). Yet the authors (Billieux et al. 2022; Karhulahti et al. 2022) agreed on the need for improvements and harmonization in the assessment and screening of GD. This reply aims to apply Wakefield’s Harmful Dysfunction Analysis (HDA) (Wakefield 1992a, 1992b, 2013, 2015, 2020) to the study of GD to test whether it improves convergence in prevalence rates and identification of participants with GD as well as detects health differences between participants with GD and those without GD, using data and codes shared by Karhulahti et al. (2022) at https://osf.io/ v4cqd/. HDA may provide a useful framework for the assessment of addictive disorders (Wakefield and Schmitz 2014, 2015). According to HDA, mental disorders are harmful dysfunction: a “disorder requires both dysfunction – that is, failure of some mechanism to perform a function that it was biologically designed to perform [“when within an appropriate environment that matches the range of conditions for which it was selected” (Wakefield 2017b) (p. 60)] [... ] – and harm, where the dysfunction causes harm to the individual as judged by social values” (Wakefield 2017a) (p.40– 41). In the case of addictive disorders, dysfunctions may be caused by evolutionarily novel stimuli (e.g. technological creations) for which the brain and other biological systems were not designed (Wakefield 2017a) and that lead to failures of designed regulatory systems (Wakefield 2017b). The dysfunction that results from the novel input has been referred to as a dysfunction in self-regulation, a dysfunction of the desire/deliberation/choice system, a pathological degree of diminution of control (Wakefield 2009, 2013, 2017a, 2017b) or a motivational dysfunction (Wakefield 2018, 2020). Details on the methods of this secondary analysis are available at https://osf.io/kprj5/. In comparison to the findings of Karhulahti et al. (2022), prevalence rates of GD based on HDA have slightly increased (see Supplementary material) due to the requirement of endorsing a minimum of two criteria: dysfunction and harm. Instruments convergence is mostly unaffected, further supporting previous explanations about the role of different approaches and criteria on which each instrument relies (e.g. item content/formulation), response modality and population studied, in influencing the detection of GD (Jo et al. 2019; Higuchi et al. 2021; Billieux et al. 2022; Yen et al. 2022). Indeed, based on the ICD-11 criteria and GDT items, impaired control over gaming (alongside its negative consequences) is central to GD; whereas according to the DSM-5 and IGDT10, GD is studied under the broader addiction framework in which impaired or loss of control is explored alongside classic symptoms of dependence. In line with the results of the original study, the results of between-group differences adopting the HDA perspective confirm that GDT and GAS7 (but not IGDT10) based GD groups reported poorer mental health compared to the general population. Further, the GD group based on the GDT showed poorer mental health compared to the GD group based on the IGDT10. To conclude, this secondary analysis, despite its limitations (Supplementary material), encourages researchers and clinicians to consider the HDA as a promising approach to identifying individuals with GD most in need of support because of the presence of both dysfunction and harm. Future research could examine whether HDA criteria may help in enhancing the detection of GD by minimizing false negatives