Dimensional models are commonly used as a supplement to the categorical model within the field of personality disorders. The purpose of this study was to examine personality dimensions as predictors of 5-year outcomes among women with bulimia nervosa. One hundred and thirty-four women with bulimia nervosa participated in a randomised psychotherapy treatment trial. Data was available for 109 out of the 134 participants at follow-up. Outcomes were the presence of any eating disorder (past year), the presence of a mood disorder episode (past year), and the global assessment of functioning at 5-year follow-up. Self-directedness was the only predictor of any eating disorder diagnosis (past year) at 5-year follow-up. Asceticism significantly predicted the presence of a mood disorder episode (past year) at 5 years. Borderline personality disorder symptoms predicted global functioning at 5 years. These results suggest that high self-directedness at pre-treatment may offer potential prognostic information regarding eating disorder status 5 years post-treatment. Furthermore, no single measure predicted outcome for all variables (any eating disorder diagnosis, a mood disorder episode (past year), or global functioning) at 5-year follow-up. This suggests that a comprehensive personality assessment using multiple measures is desirable for predicting outcomes.
The assessment of personality remains an area of ongoing scientific debate with no consensus regarding an optimal approach to categorical and dimensional models. Personality has been implicated in predicting global functioning and symptomatic expression in bulimia nervosa (BN); however, there is some debate about its impact on outcome (Grilo et al., 2003, Rossiter et al., 1993, Rowe et al., 2008 and Wonderlich et al., 1994). Some of this discrepancy can be attributed to the different models (categorical and dimensional), different assessment methods (clinical interviews and self-report), and different measures of personality used with this population.
The current categorical model of personality disorders has several limitations. Personality disorder symptoms in the categorical model are moderately heterogeneous, which result in high rates of co-occurring personality disorder diagnoses (Livesley, 1998). Furthermore, the stability of Axis II diagnoses and the reliability of Axis II assessment instruments are relatively poor (Trull and Durrett, 2005). Historically, the dimensional approach has existed alongside the categorical approach; however, these limitations have encouraged increased research investigating dimensional models of personality as an alternate approach (Goldner et al., 1999, Morey et al., 2007, Pukrop, 2002 and Widiger and Simonsen, 2005). One personality traits model that has been frequently examined is Cloninger's seven-factor model (Cloninger et al., 1993 and Cloninger et al., 1994).
The Psychobiological Model of Temperament and Character (TCI) (Cloninger et al., 1993) examines four dimensions of temperament: novelty seeking, harm avoidance, reward dependence, and persistence; and three dimensions of character: self-directedness, cooperativeness, and self-transcendence (Cloninger et al., 1993). Bulimia nervosa is characterised by high harm avoidance, low self-directedness and high novelty seeking (Fassino et al., 2001 and Fassino et al., 2004). In a previous study by Bulik et al., 1998a and Bulik et al., 1998b, self-directedness was a predictor of BN outcome at 1-year follow-up (Bulik et al., 1998a and Bulik et al., 1998b). However, this finding was not replicated in a study by Bloks et al. (2004) at 2.5 year follow-up (Bloks et al., 2004). Clarifying the role of self-directedness as a predictor of outcome may have important implications for the treatment of BN. For example, it has been suggested that increasing self-directedness as a pre-treatment intervention may be a useful approach (Anderson et al., 2002).
The Eating Disorder Inventory (EDI) is one of the most widely used self-report questionnaires in eating disorder research. In addition to the three eating disorder symptom scales (drive-for-thinness, bulimia and body dissatisfaction), the EDI has eight subscales assessing characteristics relevant to eating disorders, including personality (Garner et al., 1983). Elevated EDI subscales have been noted in those eating disorder patients with other comorbid disorders such as affective, anxiety, and personality disorders (Bizeul et al., 2003, Hinrichsen, 2004, Milos et al., 2004 and Sunday et al., 1993). Given the EDI was specifically designed to examine personality features and psychological traits in an eating disorder population, it is surprising that few studies have used this as an outcome measure for these themes.
Future editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) may consider the inclusion of a dimensional classification system as a new way of conceptualizing personality pathology. In relation to this potential change, it would be useful to clarify the link between personality dimensions and their impact on outcome. This article will explore personality predictors of outcome in BN using dimensional measures. The aims of the present study are to determine whether personality dimensions measured at pre-treatment by the TCI, EDI-2, and DSM-III-R personality symptoms among women with BN predicted the following at 5 years: (1) eating disorder outcome (past year) (2) mood disorder episode (past year) (3) global functioning. This is an exploratory study to investigate the extent to which personality dimensions might predict long-term outcome of BN and as such we did not have specific a priori hypotheses
