Finding the Best Match - a Case Study on the (Text-)Feature and Model Choice in Digital Mental Health Interventions

With the need for psychological help long exceeding the supply, finding ways of scaling, and better allocating mental health support is a necessity. This paper contributes by investigating how to best predict intervention dropout and failure to allow for a need-based adaptation of treatment. We systematically compare the predictive power of different text representation methods (metadata, TF-IDF, sentiment and topic analysis, and word embeddings) in combination with supplementary numerical inputs (socio-demographic, evaluation, and closed-question data). Additionally, we address the research gap of which ML model types - ranging from linear to sophisticated deep learning models - are best suited for different features and outcome variables. To this end, we analyze nearly 16.000 open-text answers from 849 German-speaking users in a Digital Mental Health Intervention (DMHI) for stress. Our research proves that - contrary to previous findings - there is great promise in using neural network approaches on DMHI text data. We propose a task-specific LSTM-based model architecture to tackle the challenge of long input sequences and thereby demonstrate the potential of word embeddings (AUC scores of up to 0.7) for predictions in DMHIs. Despite the relatively small data set, sequential deep learning models, on average, outperform simpler features such as metadata and bag-of-words approaches when predicting dropout. The conclusion is that user-generated text of the first two sessions carries predictive power regarding patients' dropout and intervention failure risk. Furthermore, the match between the sophistication of features and models needs to be closely considered to optimize results, and additional non-text features increase prediction results.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41666-023-00148-z.

Converging vulnerability factors for compulsive food and drug use

Highly palatable foods and substance of abuse have intersecting neurobiological, metabolic and behavioral effects relevant for understanding vulnerability to conditions related to food (e.g., obesity, binge eating disorder) and drug (e.g., substance use disorder) misuse. Here, we review data from animal models, clinical populations and epidemiological evidence in behavioral, genetic, pathophysiologic and therapeutic domains. Results suggest that consumption of highly palatable food and drugs of abuse both impact and conversely are regulated by metabolic hormones and metabolic status. Palatable foods high in fat and/or sugar can elicit adaptation in brain reward and withdrawal circuitry akin to substances of abuse. Intake of or withdrawal from palatable food can impact behavioral sensitivity to drugs of abuse and vice versa. A robust literature suggests common substrates and roles for negative reinforcement, negative affect, negative urgency, and impulse control deficits, with both highly palatable foods and substances of abuse. Candidate genetic risk loci shared by obesity and alcohol use disorders have been identified in molecules classically associated with both metabolic and motivational functions. Finally, certain drugs may have overlapping therapeutic potential to treat obesity, diabetes, binge-related eating disorders and substance use disorders. Taken together, data are consistent with the hypotheses that compulsive food and substance use share overlapping, interacting substrates at neurobiological and metabolic levels and that motivated behavior associated with feeding or substance use might constitute vulnerability factors for one another. This article is part of the special issue on 'Vulnerabilities to Substance Abuse'.

Elucidating knowledge and beliefs about obesity and eating disorders among key stakeholders: paving the way for an integrated approach to health promotion

Background

Understanding the knowledge and beliefs of key stakeholders is crucial in developing effective public health interventions. Knowledge and beliefs about obesity and eating disorders (EDs) have rarely been considered, despite increasing awareness of the need for integrated health promotion programs. We investigated key aspects of knowledge and beliefs about obesity and EDs among key stakeholders in Australia.

Methods

Using a semi-structured question guide, eight focus groups and seven individual interviews were conducted with 62 participants including health professionals, personal trainers, teachers and consumer group representatives. An inductive thematic approach was used for data analysis.

Results

The findings suggest that, relative to obesity, EDs are poorly understood among teachers, personal trainers, and certain health professionals. Areas of commonality and distinction between the two conditions were identified. Integrated health promotion efforts that focus on shared risk (e.g., low self-esteem, body dissatisfaction) and protective (e.g., healthy eating, regular exercise) factors were supported. Suggested target groups for such efforts included young children, adolescents and parents.

Conclusions

The findings indicate areas where the EDs and obesity fields have common ground and can work together in developing integrated health promotion programs.

Treating the binge or the (fat) body? Representations of fatness in a gold standard psychological treatment manual for binge eating disorder

This article reports the results of a Foucauldian-informed discourse analysis exploring representations of fatness embedded within an empirically based psychological treatment manual for binge eating disorder, a condition characterized by overvaluation of weight and shape. Analyses indicate that the manual prioritizes weight loss with relatively less emphasis placed on treating the diagnostic symptoms and underlying mechanisms of binge eating disorder. We raise critical concerns about these observations and link our findings to mainstream psychology's adoption of the medical framing of fatness as obesity within the "gold standard" approach to intervention. We recommend that psychology as a discipline abandons the weight loss imperative associated with binge eating disorder and fat bodies. We recommend that practitioners locate the problem of fat shame in society as opposed to the individual person's body and provide individuals with tools to identify and resist fat stigma and oppression, rather than provide them with tools to reshape their bodies.

The Obesities: An Overview of Convergent and Divergent Paradigms

The study of obesity lends itself to difficulties not only due to our imprecise ability to measure body composition, food consumption, and physical activity but also, even more important, due to complexities involved in defining and conceptualizing obesity. For centuries, obesity has been considered a disease, although researchers and clinicians cannot agree on definitions of "disease" or, if it is one, whether obesity is a disease of metabolism, inflammation, brown fat, chronobiology, the blood-brain barrier, the right brain, or even of infectious origin. The concept of "obesity" as a disease remains controversial to some because not everyone who has excess adipose tissue has any evidence of disease. Obesity, though, has also been considered a sin, a crime against society, an aesthetic crime, a self-inflicted disability, an example of body diversity, a failure in the regulation of energy balance, an appropriate or even inappropriate adaptation to our increasingly obesogenic environment, a genetic disorder, and a psychological/behavioral disorder of overeating involving self-regulation or even addiction. Five major paradigms-medical, sociocultural, evolutionary, environmental, and psychological/behavioral, all with their own subcategorical models-have been identified. All 5 paradigms are required because we are dealing not with "obesity" but with a plurality, the "obesities."

An Overview of Links Between Obesity and Mental Health

The American Medical Association recently recognized obesity as both an illness and a leading cause of preventable death and chronic disease. This association is not only linked to physical health outcomes, however, as obesity has also been extensively associated with mental illness as well. Both obesity and severe mental illness decrease quality of life and are associated with an increase in disability, morbidity, and mortality, and when they occur together, these adverse health outcomes are magnified. Despite educational campaigns, increased awareness, and improved treatment options, the high prevalence of mental illness and comorbid obesity remains a serious problem. This review examines this overlap, highlighting clinical and biological factors that have been linked to this association in order to improve our understanding and help elucidate potential therapeutic avenues.

Goal-Directed Resilience in Training (GRIT): A Biopsychosocial Model of Self-Regulation, Executive Functions, and Personal Growth (Eudaimonia) in Evocative Contexts of PTSD, Obesity, and Chronic Pain

This paper presents a biopsychosocial model of self-regulation, executive functions, and personal growth that we have applied to Goal-Directed Resilience in Training (GRIT) interventions for posttraumatic stress disorder (PTSD), obesity, and chronic pain. Implications of the training for the prevention of maladaptation, including psychological distress and health declines, and for promoting healthy development are addressed. Existing models of attention, cognition, and physiology were sourced in combination with qualitative study findings in developing this resilience skills intervention. We used qualitative methods to uncover life skills that are most salient in cases of extreme adversity, finding that goal-directed actions that reflected an individual's values and common humanity with others created a context-independent domain that could compensate for the effects of adversity. The efficacy of the resilience skills intervention for promoting positive emotion, enhancing neurocognitive capacities, and reducing symptoms was investigated in a randomized controlled trial with a veteran population diagnosed with PTSD. The intervention had low attrition (8%) and demonstrated improvement on symptom and wellbeing outcomes, indicating that the intervention may be efficacious for PTSD and that it taps into those mechanisms which the intervention was designed to address. Feasibility studies for groups with comorbid diagnoses, such as chronic pain and PTSD, also showed positive results, leading to the application of the GRIT intervention to other evocative contexts such as obesity and chronic pain.

Obesity--a neuropsychological disease? Systematic review and neuropsychological model

Obesity is a global epidemic associated with a series of secondary complications and comorbid diseases such as diabetes mellitus, cardiovascular disease, sleep-breathing disorders, and certain forms of cancer. On the surface, it seems that obesity is simply the phenotypic manifestation of deliberately flawed food intake behavior with the consequence of dysbalanced energy uptake and expenditure and can easily be reversed by caloric restriction and exercise. Notwithstanding this assumption, the disappointing outcomes of long-term clinical studies based on this assumption show that the problem is much more complex. Obviously, recent studies render that specific neurocircuits involved in appetite regulation are etiologically integrated in the pathomechanism, suggesting obesity should be regarded as a neurobiological disease rather than the consequence of detrimental food intake habits. Moreover, apart from the physical manifestation of overeating, a growing body of evidence suggests a close relationship with psychological components comprising mood disturbances, altered reward perception and motivation, or addictive behavior. Given that current dietary and pharmacological strategies to overcome the burgeoning threat of the obesity problem are of limited efficacy, bear the risk of adverse side-effects, and in most cases are not curative, new concepts integratively focusing on the fundamental neurobiological and psychological mechanisms underlying overeating are urgently required. This new approach to develop preventive and therapeutic strategies would justify assigning obesity to the spectrum of neuropsychological diseases. Our objective is to give an overview on the current literature that argues for this view and, on the basis of this knowledge, to deduce an integrative model for the development of obesity originating from disturbed neuropsychological functioning.

From passive overeating to "food addiction": a spectrum of compulsion and severity

A psychobiological dimension of eating behaviour is proposed, which is anchored at the low end by energy intake that is relatively well matched to energy output and is reflected by a stable body mass index (BMI) in the healthy range. Further along the continuum are increasing degrees of overeating (and BMI) characterized by more severe and more compulsive ingestive behaviours. In light of the many similarities between chronic binge eating and drug abuse, several authorities have adopted the perspective that an apparent dependence on highly palatable food-accompanied by emotional and social distress-can be best conceptualized as an addiction disorder. Therefore, this review also considers the overlapping symptoms and characteristics of binge eating disorder (BED) and models of food addiction, both in preclinical animal studies and in human research. It also presents this work in the context of the modern and "toxic" food environment and therein the ubiquitous triggers for over-consumption. We complete the review by providing evidence that what we have come to call "food addiction" may simply be a more acute and pathologically dense form of BED.