Recent Meta-Analysis

Treatment Outcomes for Anorexia Nervosa: Our Recent Meta-Analysis

Daniel Le Grange, Ph.D., FAED, Katharine Loeb, Ph.D., FAED, and Stuart Murray, Ph.D.

Every study that the scientific field conducts contributes, metaphorically speaking, only one piece to a very large puzzle. This is particularly true for complex disorders like anorexia nervosa. Even a meta-analysis represents just one piece of such a puzzle. With that in mind, below is a summary to highlight some of what our recent meta-analysis has underscored for us as both scientists and practitioners in the field of eating disorders.

First, the positive or negative nature of a study should be judged by the scientific integrity of the research, rather than the implications of the results.  Meta-analytic designs are well-suited to investigate pooled patterns of findings across multiple, separate studies.  The intrinsic statistical limitations of a meta-analysis, and the corresponding risks of over- or mis-interpretation of the results, are mitigated by study’s clarity of presentation regarding: the specific, targeted research question(s); the criteria for selection of studies to include in the analysis; and the conclusions that can be drawn from the results.  It is always beneficial to read an original research paper rather than relying on media-based summaries, regardless of whether a study’s findings appear to support or refute scientific and popular beliefs.

Our study aimed to examine aggregate outcomes in RCTs for AN by delineating two different symptom dimensions:  the physical (i.e., body weight indices) and the psychological (i.e., fear of weight gain and other clinically significant body concerns).  The distinction between these two domains is important because clinical and empirical observations suggest that there is often a disconnect – in timing if not in mechanisms – between pathways to recovery in each.  Our research questions were not directed at a particular intervention, but rather at treatment studies in general, relying on the original investigators’ designation of specialized vs. comparator modalities.  Our analyses address these interacting sets of questions, that is, how do specialized vs. comparator interventions fare in achieving positive change in weight vs. psychological symptoms for AN?

Our results indicate that weight and psychological AN symptoms follow distinct trajectories throughout treatment.  Our results also highlight that as a field, we need to continue to focus on precision methods to achieve and sustain psychological relief for our patients.  While we believe such relief is predicated on sufficient weight gain, reliably realizing durable remission across both symptom domains will require further study of the core mechanisms underpinning psychological AN psychopathology.  Research efforts to accomplish this goal, and to apply findings to shape more potent treatment strategies (e.g., within a transdiagnostic framework), are ongoing in the field.  To facilitate further investigation of these questions, researchers would ideally always report outcomes from treatment studies of AN as both separated (weight vs. psychological symptoms) and combined indices.

What to do in the meantime, while AN scientists continue to try to solve an endlessly shifting jigsaw puzzle?  This meta-analysis does not try to answer that.  For an answer, we turn to other pieces of the puzzle, particularly to the findings from well-designed RCTs.  FBT, CBT-E, and other available specialized intervention modalities are indeed efficacious and reflect current best practices for the treatment of eating disorders. 

1Murray, S., Quintana, D., Griffith, S., Loeb, K., & Le Grange, D.Treatment Outcomes for Anorexia Nervosa: A Systematic Review and Meta-Analysis of Randomized Controlled Trials, Psychological Medicine, 2018 Aug 13:1–10. Doi: 10.1017/S0033291718002088.

2Fairburn CG, Cooper Z, Doll HA, O'Connor ME, Palmer RL, Dalle Grave R.  Enhanced cognitive behaviour therapy for adults with anorexia nervosa: A UK-Italy study.  Behaviour Research and Therapy 2013; 51: 2-8.