Development and validation of the Detail and Flexibility Questionnaire (DFlex) in eating disorders

Abstract

Whilst neuropsychological testing provides the most accurate profile of cognitive functioning, the time consuming nature of individual assessment deems it impossible for many research and clinical settings. This paper presents the development and validation of the Detail and Flexibility Questionnaire (DFlex), a 24-item self-report scale measuring two aspects of neurocognitive functioning; cognitive rigidity (difficulty with set-shifting/flexibility) and attention to detail (weak coherence). Exploratory factor analysis extracted two subscales, further confirmed and refined by item response analysis. Both subscales showed high internal reliability, construct validity (as compared to relevant subscales of the Autistic-Spectrum Quotient) and strong discriminant validity with large effect sizes found between both lifetime eating disorder and healthy control groups, and between current and recovered anorexia nervosa. We suggest using the cognitive rigidity and attention to detail subscales independently to give a rough approximation of these two aspects of cognitive style as they manifest in the context of everyday life.

Introduction

New models to explain anorexia nervosa (AN) have focussed on underlying mechanisms that cover vulnerability in addition to maintaining factors. The Maudsley Maintenance Model of AN (Schmidt & Treasure, 2006) details four domains that contribute to the maintenance of AN, the first of which is obsessive–compulsive personality traits. It has long been known that obsessive–compulsive traits are common in the eating disorders (Halmi et al., 2005, Kaye et al., 1993) and impact on recovery (Crane, Roberts, & Treasure, 2007), however assessment has largely focussed on the behavioural and diagnostic aspects of obsessive–compulsive disorder such as ordering and cleaning behaviours. Over the last few years, focus has started to shift from visible behaviours of those with an eating disorder to underlying cognition by investigating neurocognitive profile (Treasure, 2007, Treasure et al., 2007). This paper will focus on two cognitive styles identified through neuropsychological assessment that may fall under the obsessive–compulsive umbrella.

Cognitive flexibility or set-shifting, the ability to be flexible with one's mindset in adapting to new task demands or changes in situations, is a key aspect of executive function. A meta-analytic review identified 15 studies in the literature assessing set-shifting in the eating disorders, with consistent difficulties seen across AN, bulimia nervosa (BN) and recovered AN groups using neuropsychological tasks such as the Wisconsin Card Sorting Test, Trail Making Task, and the Brixton Test (Roberts et al., 2007, Tchanturia et al., 2005). Set-shifting has been implicated as an endophenotype of eating disorders (Holliday et al., 2005, Roberts et al., 2010, Treasure et al., 2007), and proposed as part of a cognitive neuroscience hypothesis of AN (Steinglass & Walsh, 2006). This trait can manifest both in aspects of the patients' everyday life (e.g. rigid housekeeping routines) and in terms of illness symptomology (e.g. rules around food preparation/choice of food).

More recently, the concept of weak coherence (formally referred to as “weak central coherence”, see Happe & Booth, 2008) has been explored in the eating disorders. In the main this tendency to focus on intricate detail rather than the general tendency to integrate parts in their global context has been extensively studied in the autism literature (Happe, 2005). Research evidence suggests that this inherent bias toward detail is also present in the eating disorder population, as measured by tasks such as the Embedded Figures Task, Matching Familiar Figures Task, and the Rey-Osterrieth Complex Figure (Lopez et al., 2008, Lopez et al., 2008a, Lopez et al., 2008b, Oldershaw et al., 2011, Roberts et al., submitted for publication, Southgate et al., 2008, Tokley and Kemps, 2007). A focus on detail is exemplified by the obsessive attention to precise calorie content of food items in AN.

It is of clinical interest to examine how these neurocognitive styles may shape behaviours. This has been a focus of interest in the literature on autism. Baron-Cohen, Wheelwright, Skinner, Martin, and Clubley (2001) developed the Autism-Spectrum Quotient (AQ) which incorporates domains that represent both social cognition and information processing biases seen in autism. Two of the domains represented in the AQ approximate cognitive flexibility (attentional shifting subscale) and weak coherence (attention to detail subscale), where people with autism report elevated scores on both subscales (Baron-Cohen et al., 2001). Pilot research suggested that whilst the attentional shifting subscale of the AQ discriminated between AN patients and controls, the attention to detail subscale did not (Hambrook, Tchanturia, Schmidt, Russell, & Treasure, 2008). On further investigation, it is likely that the detail subscale of the AQ is conceptually insensitive in what is a majority female clinical group, for example a large number of items are biassed toward masculine traits (e.g. “I usually notice car number plates, or similar strings of information”). Therefore the aim of the present study is to design and validate a self-report measure of behaviours possibly linked to poor set-shifting and weak coherence relevant to the eating disorder population. It is hypothesised that scores on this measure will be reliably higher amongst eating disorder patients compared to controls.