Testing the relative associations of different components of dietary restraint on psychological functioning in anorexia nervosa and bulimia nervosa

Abstract

Although empirical evidence identifies dietary restraint as a transdiagnostic eating disorder maintaining mechanism, the distinctiveness and significance of the different behavioural and cognitive components of dietary restraint are poorly understood. The present study examined the relative associations of the purportedly distinct dietary restraint components (intention to restrict, delayed eating, food avoidance, and diet rules) with measures of psychological distress (depression, anxiety, and stress), disability, and core eating disorder symptoms (overvaluation and binge eating) in patients with anorexia nervosa (AN) and bulimia nervosa (BN). Data were analysed from a treatment-seeking sample of individuals with AN (n = 124) and BN (n = 54). Intention to restrict, food avoidance, and diet rules were strongly related to each other (all r's > 0.78), but only weakly-moderately related to delayed eating behaviours (all r's < 0.47). In subsequent moderated ridge regression analyses, delayed eating was the only restraint component to independently predict variance in measures of psychological distress. Patient diagnosis did not moderate these associations. Overall, findings indicate that delayed eating behaviours may be a distinct component from other indices of dietary restraint (e.g., intention to restrict, food avoidance, diet rules). This study highlights the potential importance of ensuring that delayed eating behaviours are screened, assessed, and targeted early in treatment for patients with AN and BN.

Introduction

Dietary restraint is considered an important transdiagnostic maintaining mechanism across the eating disorders (Fairburn, Cooper, & Shafran, 2003). Although “general” levels of dietary restraint are elevated in most eating disorder patients, there is good evidence that certain diagnoses differ in the extent to which they endorse certain behavioural and cognitive components of restraint (Elran-Barak et al., 2015). For example, individuals with anorexia nervosa-restricting type are usually successful in their adherence to rigid dietary behaviours (e.g., fasting for long periods), whereas individuals with bulimia nervosa (BN) and binge eating disorder (BED) engage in a pattern of dieting that is more chaotic and inconsistent (e.g., “on” and “off” dieting days; Fairburn & Harrison, 2003). Despite the importance of dietary restraint in eating disorders, its conceptualisation remains ill-defined. That is, while numerous distinct facets of dietary restraint are discussed in the literature (e.g., intention to restrict overall food intake, avoidance of certain foods, multiple self-imposed diet rules, fasting-related behaviours), researchers have typically assessed or used these facets interchangeably under a more global “restraint subscale” (Hagan, Forbush, & Chen, 2017). Thus, the differentiated effects of some of these various behavioural and cognitive components of dietary restraint in eating disorders remains unclear.

Fairburn and colleagues provided a useful model for understanding the different “types” of restraint components, why each of these restraint component should be specifically targeted in treatment, and what role each restraint component has in maintaining other eating disorder symptoms (Fairburn, 2008, 2013; Fairburn, Marcus, & Wilson, 1993). This model underpins cognitive-behavioural therapy (CBT), and although CBT is one the leading evidence-based eating disorder treatments (Linardon, Wade, De la Piedad Garcia, & Brennan, 2017b) recent calls have been made to improve its effectiveness (Linardon, 2018). Testing aspects of CBT's underlying model is one important avenue toward improving its effectiveness (Pennesi & Wade, 2016). According to Fairburn (2013), there are “three forms” of dietary restraint, each of which involve highly specific and inflexible rules about eating. The first form is delayed eating (i.e., synonymous with fasting), which occurs when individuals delay eating for as long as possible during the day, often not eating anything until the evening. The second form is dietary restriction. Dietary restriction may be conceptualised in one of two ways, either as (1) actual undereating or (2) the intention to restrict food intake, whether or not an individual is successful in their attempt. The third form is food avoidance, which is where certain foods are completely avoided because they are perceived as “unhealthy” or “fattening”, and are hence a major trigger for binge eating. Each form of dietary restraint is suggested to be highly distressing to the individual, a cause of considerable anxiety, and assumed to interrelate with and maintain other symptoms of eating disorders (e.g., shape and weight overvaluation, binge eating; Fairburn, 2008).

Some research has examined the significance of these specific components of dietary restraint in eating disordered samples. For example, previous work has found that experimentally manipulating fasting behaviours for either six or 14 h was associated with greater binge eating severity in women with AN (De Young et al., 2014), BN (Telch & Agras, 1996), and BED (Agras & Telch, 1998). Ecological momentary assessment studies have also shown dietary restriction (undereating) to predict a greater severity of binge eating in BN (e.g., Zunker et al., 2011), and greater levels of body image concerns, negative affect, stress, and anxiety in AN (Haynos et al., 2015; Lavender, De Young et al., 2013a; Lavender, Wonderlich, et al., 2013b). Reductions in the intention to restrict food intake during the early weeks of CBT has been shown to predict favourable outcomes in BN (Wilson, Fairburn, Agras, Walsh, & Kraemer, 2002). Several cross-sectional studies have also reported associations between self-reported dietary restriction behaviours (i.e., skipping meals, eating only very small meals) and binge eating behaviour in eating disordered samples (e.g., Elran-Barak et al., 2015; Masheb, Grilo, & White, 2011). Evidence from early research demonstrated that consumption of, and exposure to, “forbidden” foods was associated with greater negative affect, stress, and disinhibited eating in eating disordered samples (Ruggiero, Williamson, Davis, Schlundt, & Carey, 1988; Soetens, Braet, Van Vlierberghe, & Roets, 2008). Together, these findings demonstrate that different components of dietary restraint may bear a unique clinical significance in eating disorders.

These purportedly distinct components of dietary restraint have not yet been compared in the same sample. Thus, it remains unclear to what extent each component co-varies and whether each component bares a unique or independent clinical significance. According to the DSM, clinical significance is defined as the relationship between a symptom and (a) marked psychological distress, or (b) impairment in functioning (American Psychiatric Association, 2013). For these reasons, we examined these restraint components relationship to a set of outcomes that reflect this definition of clinical significance. In particular, depressive, anxiety, and stress outcomes were used as our measure psychological distress, because each are known to be elevated in patients with eating disorders (Godart et al., 2015). Disability severity was selected as our measure of functional impairment, not only because there is a dearth of research examining the relationship between eating disorder features and disability severity, but also because, clinically speaking, patients with eating disorders usually present to treatment because of the debilitating effect their condition has on physical, psychological, and social functioning (Engel, Adair, Hayas, & Abraham, 2009). Certain restraint components may be bare more of a clinical significance than other components in terms of their ability to predict disability severity, thereby serving as important treatment targets. Because clinical significance may also be conceptualised as the relationship between one symptom with a set of other disorder-specific symptoms (Mitchison et al., 2017), we also included shape and weight over-evaluation and binge eating frequency as additional outcome measures.

Consequently, the present study has two aims: (1) to examine the clinical significance and relative associations of different components of dietary restraint (i.e., intention to restrict, delayed eating, food avoidance, and eating rules) on measures of psychological distress, disability, and core eating disorder symptoms; (2) to examine whether any of the observed relationships are moderated by patient diagnosis (AN or BN). Given that this is the first study to have examined these different restraint components in the same sample, no formal a priori hypotheses for these two aims were derived.