Prospective Predictors of Eating Disorder Duration and Transition to a Second Distinct Eating‐Disorder Family in Adolescent Girls and Young Women

OBJECTIVE: This hypothesis-generating study examined whether baseline risk factors and baseline eating-disorder symptoms predicted first eating-disorder episode duration and time from first incident eating-disorder onset to transition to a second distinct eating-disorder family among participants who developed an eating disorder during follow-up. METHOD: Data came from four prevention trials of participants at elevated risk because of body image concerns. The analytic sample included 223 participants who were eating-disorder-free at baseline and developed an eating disorder during 36-month follow-up. First-episode duration was defined as the duration of the first continuous eating-disorder episode. Age- and condition-adjusted Cox proportional-hazards models examined predictors of time to remission. For Aim 2, condition-adjusted Cox proportional-hazards models examined time from first eating-disorder onset to the first subsequent diagnosis from a different eating-disorder family, with participants who did not transition to a second family right-censored at their last observed diagnostic month. RESULTS: Higher baseline BMI z-score (zBMI) was associated with faster remission from the first continuous eating-disorder episode (HR = 1.170, 95% CI [1.005, 1.362], p = 0.043). Higher baseline dieting was associated with a higher rate of transition to a second distinct eating-disorder family (HR = 1.421, 95% CI [1.034, 1.952], p = 0.030). DISCUSSION: Higher baseline zBMI was associated with faster remission from the first continuous eating-disorder episode, whereas higher baseline dieting was associated with a higher rate of transition to a second distinct eating-disorder family. These exploratory findings warrant replication in future studies.