Breaking the cycle: Identifying key symptom pathways of eating disorders and the influence of childhood maltreatment

Background Most network analyses on central symptoms in eating disorders (EDs) have been cross‐sectional. Longitudinal within‐person analyses of therapy processes are scarce. Our aim was to investigate central change processes in therapy in a transdiagnostic sample, considering the influence of childhood maltreatment. Method We employed dynamic time warping analyses to identify clusters of symptoms that tended to change similarly across therapy on a within‐person level. Symptoms were measured by a 28‐item Eating Disorder Examination Questionnaire (EDE‐Q). Furthermore, we examined the temporal direction of symptom change to identify symptoms that tended to precede and predict other symptoms. Finally, we estimated two directed, temporal networks in patients with and without a history of childhood maltreatment. Results Our analysis included 122 ED patients (mean age = 30.9, SD = 9.7; illness duration = 14.2 years, SD = 8.9; prior treatment = 5.6 years, SD = 5.1). The initial network revealed three robust clusters of symptoms over time: (1) ED behavior, (2) inhibition, and (3) cognitions and feelings about body and weight. Overvaluation of shape had the highest out‐strength preceding and predicting other symptoms. Dissatisfaction with weight preceded and predicted other symptoms in the maltreatment network. The non‐maltreatment network showed a similar structure to the transdiagnostic network. Conclusion Targeting and monitoring feelings and cognitions related to shape may be crucial for achieving lasting symptom improvement in a transdiagnostic sample. Furthermore, our findings highlight the need for further investigation into the different processes driving EDs based on maltreatment status. Public significance There is limited understanding of the processes that occur for patients with eating disorders between admission and discharge in therapy, especially for patients with a history of childhood maltreatment. Our analyses suggest that changes in cognitions regarding shape precede and predict changes in cognitions about weight. Different processes may be driving the eating disorder according to maltreatment status, which might further illuminate the riddle of dropout and relapse in therapy for patients with a history of childhood maltreatment. These findings suggest the need for further investigation into the specific dynamics occurring during therapy for individuals with a history of childhood maltreatment.