Are Hoarding Disorder and Buying-Shopping Disorder Behavioral Addictions? Too Soon to Tell and a Roadmap for Future Investigations

When considering hoarding disorder (HD) and buying-shopping disorder (BSD) as behavioral addictions, Pickering and Norberg (2022) have provided a thoughtful and comprehensive review that argues for a proposed reconceptualization. The authors first outline one set of criteria used to define behavioral addictions, which include: (a) verifying that the behaviour consistently leads to functional impairment, (b) demonstrating that the behaviour is best explained by current addiction theory, and (c) establishing that the neurobiological or cognitive dysfunctions underlying the behaviour are similar to those identified for other behavioural and/or substance-related addictions. In particular, the authors highlight the Components Model of Addiction (Griffiths, 2005) to define a behavioural addiction based on the six core facets of salience, mood modification, tolerance, withdrawal, conflict, and relapse. The authors then conducted a narrative review of the HD and BSD literature, evaluating both conditions in light of the defining characteristics and etiological models of behavioural addictions. Pickering and Norberg conclude that the available data on BSD ranges from reasonable to convincing, while the available data on HD is more limited. The authors suggest that a reconceptualization of the HD and BSD classification is called for, and that both conditions “may more appropriately be viewed as falling under a behavioural addiction framework.” Critically, they outline an intriguing program of future research sparked by such a conceptual shift, including potential treatment modifications and etiological investigations.

Both HD and BSD are severe and intractable conditions associated with serious ramifications to the patient, their families, and, in the case of HD, even their communities (Müller et al., 2019; Timpano et al., 2016). This recognition – coupled with limited intervention options – signals that both HD and BSD should be acknowledged as serious public health concerns. It is true that research on HD and BSD is in its nascent stages. As a result, we know little about their etiological processes or maintaining factors, as well as effective treatment strategies. Pickering and Norberg are therefore commended for thinking beyond traditional boundaries imposed by the current DSM-5 nosology. We concur with their opinion that the field should revisit the validity of current classifications for HD and BSD. It is critical to do so since the classification of psychiatric syndromes and the elucidation of underlying etiological factors are inextricably intertwined: nosology forms the foundation of theory, and theory drives subsequent empirical investigations. In other words, if our understanding of HD and BSD is to grow beyond current models, a paradigm shift may be warranted.

To facilitate such a discussion, we highlight several considerations and raise a few counterarguments to the conclusions drawn by Pickering and Norberg. As hoarding researchers, we acknowledge that we have approached this critique with a greater emphasis on a HD perspective than a consideration of the BSD literature. We first draw attention to the fact that the definition of behavioural addictions is complicated and lacks a strong consensus (within the broader addictions field) regarding the specific criteria that should be employed. Alternatives to the Components Model of Addiction have relied on neuroscience-based models of substance-use disorders. These are grounded within a neuroadaptive and neurodevelopmental framework that stipulates both an in-the-moment intoxication/withdrawal/anticipation cycle, as well as a more protracted, allostatic process that shapes and shifts underlying neurocircuitry related to reward, stress, and executive function over time (Yücel et al., 2019). These models highlight excessive involvement with the substance (or behaviour), apparent tolerance or withdrawal, loss of control, cravings, cognitive salience, and mood regulation as core features of addiction. While some of these facets overlap with those presented by the Components Model of Addiction, they also differ on several important parameters, and thus reflect a more nuanced theoretical perspective. Further examination of the BSD and HD literatures may demonstrate that it is more or less (depending on the condition) challenging to find support for constructs such as craving, loss of control, cognitive salience, or the allostatic processes.

An additional critique of the Components Model of Addiction is that some of facets are difficult to operationalize in a “specific-enough” manner when shifting from a substance or chemical addiction to excessive behaviour (Billieux et al., 2015). Concern lies in the fact that the criteria may lead to an over-zealous characterization of any number of excessive behaviors as an addiction, even if they are not truly “addictive” processes. This critique is relevant when examining how well the BSD and HD literatures support the core facets of the Components Model of Addiction. Although we agree with Pickering and Norberg that the data is more convincing with respect to BSD, we argue that the supporting literature is markedly more problematic for HD, particularly when attempting to support the addiction facets of salience, mood modification, and tolerance/withdrawal.

Salience is when a substance/activity dominates a person’s day-to-day experience; reverse salience is when the substance/activity dominates one’s thinking, feelings, and behaviour only when the substance/activity is no longer accessible. Pickering and Norberg postulate that hoarding patients’ reactions to losing their objects is a demonstration of reverse salience. Even though one can draw parallels to reverse salience in addictive behaviours, one could also draw connections to other conditions that experience distress when cycles of negative reinforcement are disrupted. Two examples would be if an anxious patient experiences fear when they are no longer able to rely on safety aids, or an individual with obsessive compulsive disorder who reports distress if they cannot complete a ritual. Neither example would be characterized as a behavioural addiction to safety aids or rituals, respectively. We maintain that the underlying mechanisms that contribute to heightened negative-emotional reactivity in response to losing or discarding one’s possessions are likely more complex than is captured by the negative salience component of an addictions framework. Experimental work has demonstrated that individuals with HD symptoms experience a range of emotions in response to possession loss, including both negative and positive emotions. It has been theorized that mixed emotions may reflect a more complex and insecure attachment process in HD that does feel neatly into a salience conceptualization (Yap & Grisham, 2021).

A similar argument can be used for the HD data the authors presented as support for the mood modification facet. Whereas the available data in BSD and HD populations is more convincing for acquiring – insofar that collecting behaviours can serve to upregulate negative emotions, or are reinforced through maintaining positive emotions – the literature is once again much less clear when it comes to saving behaviours. Pickering and Norberg describe investigations that have found that distress and related negative-mood states are associated with increased saving behaviours. Available research, however, has not yet demonstrated the reverse: that saving behaviours increase positive mood. In fact, HD is consistently associated with high, negative affect and high rates of depression. We thus do not agree that mood plays a critical role in saving behaviours, and instead reason that the process by which mood and saving behaviours are related has been insufficiently examined, to date, to draw any firm conclusions.

The third facet of the Components Model of Addiction that is challenging to apply to the HD literature is the notion of tolerance and withdrawal. These constructs are grounded in our understating of the physiological effects that chemical substances have, and are generally less applicable to excessive behaviors. More time spent pursuing the behaviour is the most common way to interpret tolerance, whereas withdrawal is frequently interpreted as frustration at not being able to pursue the behaviour. Pickering and Norberg apply a similar rationale to interpreting the HD literature. To the contrary, we assert that greater time spent on something is likely driven by myriad factors, some (or all) of which might not be linked with “tolerance” (whether loosely or broadly defined). For example, the progression of hoarding symptoms across the lifespan could be a factor of greater earning capacity impacting the ability to acquire more belongings, the deterioration of executive functions with age, or the slow accumulation of more and more clutter that is simply a function of time.

A final critique of the etiological and phenomenological support presented for BSD and HD as behavioural addictions relates to Pickering and Norberg’s discussion of impulsivity. The authors state that impulsivity and compulsivity are core addictive processes that are typically viewed as two ends of a spectrum. More recent models have instead suggested that compulsivity and impulsivity should be conceptualized as being orthogonally related to one another: that is, they represent two constructs that are separate, but linked by shared neuropsychological mechanisms involving dysfunctional inhibition of thoughts and behaviours (Fineberg et al., 2010). Moreover, considering that impulsivity – particularly emotion-related impulsivity – has been implicated in any number of affective conditions, it seems over-general to conclude that impulsivity is a key marker for addiction.

After considering the extant literature, we believe that it would be premature to reclassify HD as a behavioural addiction, which could potentially allow an “addiction perspective” to overshadow critical empirical work outside of the addictions literature. That being said, we concur that our understanding of HD may certainly benefit from an addiction perspective, as has been previously considered (see Grisham et al., 2010), but we would encourage researchers to remain open to alternative frameworks. There is now ample data that the DSM-5 categorizations do not always “carve nature at its joints” and sometimes fail to reflect the true nature of psychopathology. DSM-5 diagnostic categories do not reflect the continuous nature of most psychopathology and are marked by low reliability, diagnostic instability, and inability to capture comorbidity and symptom heterogeneity. DSM-5 categories further do not consider underlying etiology, either with respect to diagnosis-specific risk or transdiagnostic factors. As such, it would perhaps be most beneficial to set the question of classification aside for the time being, and instead turn our focus to core etiological mechanisms (Michelini et al., 2021).

Building on the intriguing review presented by Pickering and Norberg, our recommendation is to use the core domains highlighted as a catalyst for studying psychobiological processes associated with BSD and HD, regardless of DSM-5 classification. By examining whether or not particular processes fail to distinguish HD or BSD from behavioural addictions or other affective disorders, we can begin to map out the shared and unique vulnerability mechanisms that play a critical role in the development of each condition. Indeed, a number of the facets highlighted by Pickering and Norberg are included in the Research Domain Criteria (Michelini et al., 2021). In addition, a number of the domains discussed overlap with those identified as primary targets for addictions research (Yücel et al., 2019), and efforts have been made to discern impulsivity and compulsivity (Brooks et al., 2017). We believe that a productive roadmap to guide additional research would focus on core emotional and cognitive mechanisms that both HD and BSD share with each other, and with addictions or affective disorders. These include, but are not limited to, the negative valence system (e.g., negative emotionality, emotion-related impulsivity), positive valence system (e.g., incentive salience, positive reward and appetitive mechanisms), arousal and regulatory systems (e.g., mood regulation, preoccupation/anticipation), cognitive systems (e.g., decision-making, cognitive flexibility), and social processes (e.g., interpersonal versus object attachment). We should aim to improve understanding of both transdiagnostic processes and vulnerabilities, as well as specific risk factors and mechanisms for both HD and BSD, thereby more effectively “carving nature at its joints.”

A final reflection on the review is that we concur with Pickering and Norberg that treatment research on HD and BSD has largely stalled and that this is concerning given the modest treatment gains made by patients (Müller et al., 2019; Timpano et al., 2016). As described in the previous paragraph, a more sophisticated understanding of underlying mechanisms for these disorders may drive the development and integration of novel treatment strategies that better target specific, key pathological processes, thereby improving treatment outcomes for these populations. We further agree with the authors that the addictions literature presents an intriguing opportunity to either consider potential ways to augment current treatment techniques, or provide an avenue to consider approaches not yet tested with HD or BSD patient populations. Pickering and Norberg discuss a range of such suggestions, from exposure activities, counterconditioning approaches, contingency management, remediation of interpersonal dysfunction, and attentional bias modification. In reflecting on these suggestions, we would contend that several of the examples provided are also not necessarily specific or unique to intervention efforts for addictions (e.g., attentional bias modification), and are furthermore consistent with current best practice treatment approaches for treating HD (e.g., exposure techniques discussed). We would also highlight an additional challenge in the treatment of HD—issues of insight and motivation—for which the addictions literature might be particularly relevant, and which was only briefly discussed by Pickering and Norberg. Specifically, recent advances in the treatment of addictions literature addressing motivational issues and accommodation behaviors may provide a particularly insightful and fresh perspective for the development of novel approaches for HD and BSD. Some of these addiction-derived approaches may also correspond with a recent shift in hoarding research from a dominant focus on the role of the negative emotions associated with discarding possessions to greater consideration of the reward system and role of positive emotions in relation to acquiring (Yap & Grisham, 2021).

Our overarching recommendation for the field is to consider factors or processes that are similar in HD (or BSD) and in other conditions, and to simultaneously continue investigating mechanisms that differentiate HD (or BSD) from other conditions. Knowing which parameters have parallel processes in other disorders and which are unique to HD or BSD will enrich future treatment research. From there we could examine questions of how to maximize intervention effects, consider features not currently addressed with current treatment models, and investigate additional ways to maintain treatment gains. For example, should research clarify whether and in what manner attentional biases or reward sensitivity play a role in HD or BSD, specific treatments could be designed to strengthen acquiring interventions. Similarly, efforts to incorporate better interventions to help patients exercise self-regulation may constitute an important new avenue of research, as well are studies using cognitive techniques to target cognitive control and ongoing efforts to understand interpersonal functioning and loneliness. Finally, any consideration of addictions treatments would be remiss to consider the application of motivational interviewing, and whether those efforts would need to be tailored to more effectively be applied to HD and BSD.

In conclusion, we appreciate the thought-provoking review by Pickering and Norberg and agree with our colleagues that there is much to be gained from ensuring that HD and BSD researchers are not functioning in a silo, but tap into other rich empirical frameworks, including behavioural addictions. We also note that given patterns of comorbidity in HD, as well as consistent findings that emotional disorders share similar cognitive-affective, behavioural, and interpersonal maintaining factors, hoarding research has and will continue to benefit from exciting work in mood, anxiety and trauma-related disorders. There are strands of research in each of these research fields that have potential to advance our understanding of BSD and HD and improve our inadequate treatment outcomes. We urge caution, however, in a premature reclassification of BSD or HD without a more complete understanding of the underpinning psychological mechanisms that drive the collecting and saving of possessions in these complex and burdensome disorders.