The Eating Disorder Inventory in a UK National Health Service Obesity Clinic and its response to modest weight loss

Abstract

Objective: Using the Eating Disorder Inventory (EDI-2) to identify a normative range in a UK obese population referred to a specialist hospital obesity clinic and to assess the response of the scale to the nonsurgical treatment of obesity. Methods: A total of 100 obese (BMI>30) patients attending an obesity clinic completed the EDI-2. In a separate sample, 45 obese patients participated in a dietary weight loss program consisting of an acute (10 weeks) and a long-term (2 year) weight maintenance phase. The EDI-2 was administered at baseline and at the end of the acute weight loss phase. Results: Most obese subjects showed elevated scores for body dissatisfaction (BD), which improved with weight loss. Eight percent had EDI scores suggesting psychological traits central to eating disorders. Conclusion: Most obese subjects do not have a trait central to an eating disorder as defined by the EDI-2, but do have high scores for BD that respond to weight loss. A few patients show scores suggesting an eating disorder that may benefit from psychological evaluation.

Introduction

The diagnosis of eating disorders is based upon clinical descriptives (DSM-IV) based on clinical interview or from questionnaires that may be self administered (BITE—Bulimic Investigatory Test, Edinburgh) or as part of a structured interview (EDI—Eating Disorder Inventory). It is well recognized that obesity can be associated with disordered eating (binge eating disorder, BED) (Telch & Agras, 1994), overeating, and night eating syndrome (NES) (Rand, Macgregor, & Stunkard, 1997). It is not known if the criteria used to diagnose eating disorders in a normal population are appropriate in obesity. Many obese patients underreport or undereat in response to dietary assessment (Goris, Westerterp-Plantega, & Westerterp, 2000), and this could further complicate the diagnosis of an eating disorder in obesity. Some studies indicate that emotional and psychological problems leading to overeating commonly contribute to the aetiology of obesity Fitzgibbon et al., 1993, Maddi et al., 1997, Riva et al., 1998. Obese subjects have higher scores for depression and significant psychopathology compared to normal weight controls particularly in those with a BED, as defined by the DSM-IV criteria Kuehnel & Wadden, 1994, Molinari et al., 1997, Sansone et al., 1996, Specker et al., 1994. Estimates of the prevalence of BED in obesity vary from 10% to 50% depending on the degree of obesity and the referral practice (Yanovski, 1997). Other studies contradict the suggestion that eating disorders are common in obesity O'Neill & Jarrell, 1992, Stunkard & Wadden, 1992, Wing & Greeno, 1994. The reasons for this discrepancy are unclear. Since it is generally agreed that treatment of the eating disorder should precede, or be given concurrently with, weight loss treatment Anderson et al., 1999, Wadden et al., 1994, Weintraub et al., 1992, it would be useful to have secure criteria for diagnosing an eating disorder in an obese patient. We therefore evaluated a widely used self-administered questionnaire for diagnosing eating disorders (EDI-2) in obese patients presenting to an obesity clinic.

The EDI-2 is a measure specifically designed both for diagnosis and assessment of eating disorders. It consists of a structured interview including a self-reported assessment of symptomatology associated with anorexia nervosa and bulimia (Garner, Olmstead, & Polivy, 1983). The inventory consists of 90 questions factored into 11 subscales. The subscales are Drive for Thinness (DFT), Bulimia, Body Dissatisfaction (BD), Ineffectiveness, Perfectionism, Interpersonal Distrust, Interoceptive Awareness, Maturity fears, Asceticism, Impulse Regulation (IR), and Social Insecurity. The first three subscales assess eating behaviour and perception of body shape. The next five subscales measure psychological aspects of eating disorders that are suggestive of anorexia nervosa. The last three subscales, recently added, are provisional subscales that overlap with borderline personality traits and have found to be abnormal in a small group of eating disorder patients (Garner, 1990). Each of the 90 questions has six possible responses—“always,” “usually,” “often,” “sometimes,” “rarely,” and “never.” The range of scores for each scale varies, with higher scores indicating greater symtomatology. The EDI is useful both as a tool for baseline assessment, as well as in planning treatment, because it is easy to administer and has been shown to be effective in detecting subclinical eating disorders (Garner, 1990). It is also a useful screening tool in research-based studies and is effective as an outcome measure and prognostic indicator. In patients being treated for eating disorders, the EDI is useful in monitoring progress and recovery (Garner, 1990).

Very few studies have evaluated the EDI in obesity Adami et al., 1995, Adami et al., 1994, Chandarana et al., 1988. Almost all of these studies concerned population groups referred for bariatric surgery. These studies showed a high incidence of eating disorders. They failed to predict weight loss outcome following surgery (Powers, Perez, Boyd, & Rosemurgy, 1999). Patients referred for bariatric surgery represent those with extreme obesity. In the UK, where bariatric surgery is rarely performed, most patients presenting to primary or secondary care will have more modest degrees of obesity. We aimed to define a normative range of EDI in obese subjects referred to a National Health Service Obesity Clinic in the UK. Although the EDI has been used to monitor progress and recovery in those with an eating disorder, no study has looked at the EDI response following treatment of obesity.