Tracking Self-Image and Treatment Outcom

Vol. 30 / No. 5 — Eating Disorders Review

A patient’s initial self-image had an impact on outcome a year later.

Self-image, or the way we view ourselves, has real predictive value in treatment outcome, according to findings at the Karolinska Institute in Stockholm, Sweden.  A trio of researchers led by Dr. Emma Forsén Mantilla came to this conclusion after a year of following a large clinical sample of adult women with eating disorders (n=2221) (J Eat Disord. 2019; 7:15). Participants were from five diagnostic groups: 457 women with restrictive-type anorexia nervosa (AN-R); 228 women with binge-purge type AN (AN/BP); 861 women with BN; 505 with other specified eating disorders (OSFED), and 170 with binge-eating disorder (BED).

The authors used data from the interpersonal Structural Analysis of Social Behavior(SASB) self-image or introject, which captures both the valuation of self-esteem and self-directed actions, or how one treats oneself as a result of interpersonal learning. Previous studies had shown that a negative self-image predicted poor treatment outcome, dropping out from treatment, and suicidal behavior. Participants also completed number of questionnaires, including the Eating Disorder Examination Questionnaire (EDE-Q).

Dr. Forsén Mantilla and colleagues’ goal was to extend the findings by Birgegård et al. (Int J Eat Disord. 2009; 42:522) by studying the predictive value of the SASBon 12-month outcome among patients with DSM-5eating disorder diagnoses. Data were taken from the STEPWISE clinical database, an Internet-based data collection system for specialized eating disorders care in Sweden. STEPWISE has been in use since 2005 (Eur Eat Disord Rev. 2010; 18:251). The authors hypothesized that self-control, self-blame and, inversely, self-acceptance, could be used to strongly predict outcome in AN, whereas variables related to affiliation, as in self-love/attack would moderately predict outcome among patients with BN.

Results for AN/BP, AN/R

For patients with AN/BP, lower self-neglect scores predicted a more positive clinical outcome. For those diagnosed with AN/R, higher self-love and lower self-blame scores predicted a more positive clinical outcome at 12 months. Among those with OSFED, higher scores on self-love and lower self-blame and self-control scores predicted a better outcome; for those with BED, higher self-love scores predicted a better outcome at 12 months.

The authors also reported that AN/R patients with less severe baseline eating disorders, younger ages, and higher self-love scores do better after one year in specialist care. One intriguing finding was that higher scores on Self-examination were associated with still having an eating disorder after a year of treatment. The authors pondered whether accepting  oneself “as is” when ill, by allowing the self to remain static and to follow current impulses, can be detrimental even though the SASB variable is positive at face value. This might be a reflection of the “anorexic self” rather than the “authentic self,” and might encourage resistance to change.

Compared to previous findings, the authors’ results among AN patients were not clearly related to Self-control. Instead, the dimension of affiliation was prominent and led to the addition of Self-control in those with AN/BP and OSFED. A strength of this study, compared with earlier reports, was the authors’ attempt to include the entire eating disorders diagnostic spectrum, as represented in the DSM-5. Thus, AN subtypes were included as well as previous EDNOS (eating disorders not otherwise specified) patients in AN and BN groups.  Two limitations of the study were: (1) that relevant follow-up data were not available for 40% of the complete patient sample follow-up data (a large number), and (2) as a minor limitation, the authors attempted to construct the DSM-5diagnostic categories based on DSM-IVdiagnostic data.

The authors concluded that self-image aspects may provide clinically useful information at the beginning of treatment, especially in those with AN, because self-image are clinically relevant and may predict outcome.