OBJECTIVE: To explore patients' use of the self-monitoring app Recovery Record during 26 weeks of naturalistic eating disorder treatment.

METHODS: Selected patient characteristics at baseline were explored as predictors of app use using linear regression. Patients were grouped according to diagnosis (anorexia versus bulimia), and mixed-effects analyses were used to explore differences in app use between diagnoses across four time periods (weeks 1-4; weeks 5-8; weeks 9-12; weeks 13-26).

RESULTS: Eighty-four patients were included of which 41 had anorexia and 43 had bulimia. The total number of logs varied greatly (mean (SD): 592 (628.50)), and patient app activity almost ceased at week 13. Increasing age and no previous eating disorder treatment predicted increased app activity (p = .007; p = .039, respectively). Patients with anorexia logged over four times more often than patients with bulimia in the last time period (median (CI): 4.27 (1.28;14.31); p = .018). Time predicted declining app use (all p ≤ .007).

CONCLUSION: Future research on long-term app engagement should investigate associations between patients' app use and changes in their eating disorder symptom severity over time.