Examining a psychosocial interactive model of binge eating and vomiting in women with bulimia nervosa and subthreshold bulimia nervosa

Abstract

The current study tested a psychosocial interactive model of perfectionism, self-efficacy, and weight/shape concern within a sample of women with clinically significant bulimic symptoms, examining how different dimensions of perfectionism operated in the model. Individuals with bulimia nervosa (full diagnostic criteria or subthreshold) completed measures of bulimic symptoms, multidimensional perfectionism, self-efficacy, and weight/shape concern. Among those who were actively binge eating (n=180), weight/shape concern was associated with binge eating frequency in the context of high perfectionism (either maladaptive or adaptive) and low self-efficacy. Among those who were actively vomiting (n=169), weight/shape concern was associated with vomiting frequency only in the context of high adaptive perfectionism and low self-efficacy. These findings provide support for the value of this psychosocial interactive model among actively binge eating and purging samples and for the importance of considering different dimensions of perfectionism in research and treatment related to bulimia nervosa.

Introduction

A relation between perfectionism and bulimic symptoms has been supported in meta-analytic work, along with the suggestion that perfectionism might operate in interaction with other variables in predicting disordered eating (Stice, 2002). In order to better understand under which circumstances perfectionism is associated with bulimic symptoms, an interactive model of perfectionism, self-efficacy, and body dissatisfaction has been proposed (Bardone, Vohs, Abramson, Heatherton, & Joiner, 2000). Here, we test this theoretically-driven interactive model for the first time in a sample of women with bulimia nervosa (full diagnostic criteria or subthreshold). We also add to the literature by examining different dimensions of perfectionism in the interactive model and by separately examining binge eating and purging behaviors.

As initially conceptualized, this interactive model posits that the combination of the vulnerabilities of high perfectionism and low self-efficacy and the stress of high body dissatisfaction identifies individuals with bulimic symptoms in non-clinical populations (Bardone et al., 2000). A person with this combination of variables would be in the position of having high standards and/or feeling like a failure if she does not live up to standards (high perfectionism), being aware that she is not meeting one of her standards (body dissatisfaction), and doubting that she could reach her standards (low self-efficacy). Theoretically, this confluence of variables would generate aversive self-awareness and negative affect, which may motivate binge eating as temporary relief according to escape theory and mood modulation theory (Fairburn, Cooper, & Shafran, 2003; Heatherton & Baumeister, 1991), and as suggested by ecological momentary assessment data (Engelberg, Steiger, Gauvin, & Wonderlich, 2007).

To date, only one study has tested these three factors, finding support for their interaction in predicting binge eating but not purging in a non-clinical sample (Bardone-Cone, Abramson, Vohs, Heatherton, & Joiner, 2006). Specifically, the combination of high perfectionism, low self-efficacy, and feeling overweight was associated prospectively with greater number of weeks of binge eating. Work on a related model using self-esteem in place of self-efficacy has also found support (Holm-Denoma et al., 2005; Vohs, Bardone, Joiner, Abramson, & Heatherton, 1999; Vohs et al., 2001). Two studies using self-esteem have not replicated these interactive findings (Shaw, Stice, & Springer, 2004; Steele, Corsini, & Wade, 2007). Given conceptual work suggesting that self-efficacy, rather than self-esteem, is the better fit in interaction with failing to meet high standards (Bardone et al., 2000), and given empirical work demonstrating that increases in bulimic symptoms are associated with the self-efficacy component of self-esteem (Bardone, Perez, Abramson, & Joiner, 2003), the more robust interactive model may require self-efficacy rather than self-esteem as a key construct.

Whether perfectionism, self-efficacy, and body dissatisfaction combine meaningfully in a clinical sample is unknown. While in unselected non-clinical samples there is some support for this model predicting presence of bulimic symptoms, the question is different in samples where all individuals are exhibiting bulimic symptoms. We propose that the more appropriate question in clinical samples is: Does the model demonstrate the expected pattern of moderation, namely, that among women with the joint vulnerabilities of high perfectionism and low self-efficacy, it is those experiencing higher levels of body dissatisfaction who will demonstrate higher binge eating and purging frequencies?

Although most eating disorder research, including all but one of the prior studies of this interactive model, have used a unidimensional assessment of perfectionism, there is substantial theoretical and empirical support for multidimensional perfectionism (Frost, Marten, Lahart, & Rosenblate, 1990; Hewitt & Flett, 1991). Factor analyses of the most well-established multidimensional measures of perfectionism support an underlying structure of “maladaptive” perfectionism (e.g., interpreting mistakes as failures) and “adaptive” or “benign” perfectionism (e.g., having very high standards), with maladaptive perfectionism having stronger associations with psychopathology and distress (Dunkley, Blankstein, Halsall, Williams, & Winkworth, 2000; Frost, Heimberg, Holt, Mattia, & Neubauer, 1993). Interestingly, among individuals with bulimia nervosa there is evidence that both maladaptive and adaptive perfectionism are elevated compared with healthy controls (Lilenfeld et al., 2000).

The one study of the interactive model testing maladaptive and adaptive dimensions of perfectionism did not find an interactive effect with maladaptive perfectionism in their non-clinical sample (Steele et al., 2007). Unexpectedly, they found that high adaptive perfectionism, high self-esteem, and feeling overweight was associated with elevated bulimic symptoms, which the authors interpreted as suggesting that self-esteem does not serve as a buffer in the context of perfectionistic standards and feeling overweight. How these perfectionism dimensions operate in conjunction with self-efficacy and in clinical samples warrants investigation.

This work expands upon prior work involving an interactive model of perfectionism, self-efficacy/self-esteem, and body dissatisfaction in several important ways. It is the first study to test the relation between the confluence of perfectionism, self-efficacy, and body dissatisfaction and bulimic symptoms in a sample with clinically significant bulimic symptoms. It also contributes to the understanding of multidimensional perfectionism and bulimia nervosa by examining both maladaptive and adaptive dimensions of perfectionism. Additionally, it is only the second study of the interactive model to separately examine binge eating and purging and to focus on self-efficacy. Based on the existing literature and theory, it is predicted that the impact of body dissatisfaction (indexed in the current work by weight/shape concern) on both binge eating and purging will be greatest among individuals with the vulnerabilities of high perfectionism and low self-efficacy.