Predictors of psychological therapy treatment outcome for anorexia nervosa: a systematic review; and, Evaluation of the reliability and validity of the English version of the schema mode inventory for eating disorders-short form for adults with dysfunctional eating behaviour
Tait, Dorothy Jane
Background/Aims: Eating disorders (EDs) are considered one of the most difficult and serious psychiatric illnesses to treat, due to high levels of complexity, chronicity and comorbidity. Anorexia nervosa (AN) is particularly resistant to treatment, with the majority of studies reporting high attrition rates and limited treatment success. Gaining a better understanding of treatment outcome predictors in AN will facilitate fine-tuning of current evidence-based treatments, potentially leading to better outcomes. One predictor of ED outcomes is high levels of comorbidity, especially rigid personality traits, Cluster C personality disorders and entrenched core beliefs (early maladaptive schemas). There is good evidence that psychological problems in adulthood are associated with patterns of thinking and feeling that begin in response to a complex interplay of temperament and early life experiences. We call these patterns of thinking and feeling ‘core beliefs’ or ‘schemas’. Higher levels of early maladaptive schemas (EMS) and schema modes have been found in people with EDs. Understanding the schemas and mode states relevant to EDs is important in order to facilitate case conceptualisation, which takes account of personality/trait-based complexity and comorbidity alongside ED symptoms. The 190-item Schema Mode Inventory for eating disorders (SMI-ED) was adapted from the original SMI, in order to fine-tune measurement of EMS/modes in the ED population. A shortened version of the SMI-ED (SMI-ED-SF) was developed for regular clinical use and has been validated on an Italian ED population. This thesis consists of two studies: a systematic review (Journal Article 1) and an empirical study (Journal Article 2). The systematic review investigated the predictors of both dropout and outcome in outpatient psychological therapy for AN. The empirical study explored the psychometric properties of the English version of the SMI-ED-SF and tested how well this new, shorter measure performs. Relationships between schema modes, ED symptoms, personality traits and childhood emotional neglect were explored in this process. Methods: In Journal Article 1, a systematic search of three electronic databases, a quality assessment of included studies and a subsequent narrative synthesis were conducted. In Journal Article 2, 655 adults aged 16 and over who had ED symptomatology, alongside healthy adults were recruited from clinical services and through advertisements placed on social media and various local and international notfor- profit eating disorder organisations and support groups. Participants completed relevant psychological measures via the ‘Jisc Online Survey’ tool (historically run by Joint Information Systems Committee). The survey included the SMI-ED-SF, followed by re-testing two months later. The psychometric properties of this new measure were explored using correlational and hierarchical regression analyses. Results: In Journal Article 1, 16 studies consisting of 11 trials were reviewed using a quality criteria tool. Several psychological predictors (e.g. longer duration of illness, poor body image, emotional avoidance, employment status, age, higher anxiety, OCD symptoms) of outpatient psychological therapy outcome (e.g. ED/mental health symptoms, weight change) were identified at the end of treatment and follow-up. Predictors of dropout were also identified. In Journal Article 2, the SMI-ED-SF showed adequate concurrent, convergent, discriminant and incremental validity, and good test-retest reliability. It also showed that specific schema modes are significantly linked to ED behaviours and personality traits. In addition, it showed ED severity can be predicted by personality, childhood neglect and schema modes. Conclusions: Overall, taken together, the two studies provide evidence of how complex EDs are to treat, and that many factors must be taken into account. The systematic review showed that there was little consistency between the studies with regard to methodologies and measures used, and the way results were reported, which partly explained why consistent predictors were sometimes not identified. However, it showed how many factors need to be taken into consideration when treating AN. The empirical study tested the psychometric properties of the SMI-ED-SF which has broadened its utility for both research and clinical settings. The findings represent a step forward in developing an easy-to-use psychometrically sound instrument to identify and explore mechanisms through which schema modes can be expressed by those with EDs. Developing a more precise measure of mode states within an ED population will enhance therapeutic case conceptualisation and treatment planning. This study also showed that schema modes are significantly correlated with specific ED behaviours and personality traits. To enhance the effectiveness of treatment and reduce risk of relapse, ED sufferers with schema-level beliefs may require a treatment model that specifically addresses both eating and personality pathology, as well as childhood trauma. The new SMI-ED-SF offers patients a tool to quickly and easily begin to explore the early origins of underlying schema level representations that will increase their chances of success in therapy.