Empirically informed symptom severity cutoffs for body dysmorphic disorder

Abstract

Background

Symptom severity cutoffs for body dysmorphic disorder (BDD) are lacking, hindering communication between professionals and with the patient community.

Method

We pooled data from 11 clinical trials or high-quality cohort studies from specialist clinics, totaling 804 individuals with BDD (80 % girls/women, 67 % adults). All participants had baseline scores on the adult or adolescent versions of the Yale-Brown Obsessive-Compulsive Scale Modified for BDD (BDD-YBOCS) and the Clinical Global Impressions–Severity scale (CGI-S). Receiver-operating characteristic analyses were used to identify BDD-YBOCS severity cutoffs, using the CGI-S as the benchmark measure. The classification performance of the cutoffs was evaluated in a holdout sample, consisting of 20 % of randomly selected participants.

Results

No participants had subclinical symptoms and the cutoff for clinical versus subclinical cases was not computed. A BDD-YBOCS score ≥24 distinguished moderate from mild cases (area under the curve [AUC] = 0.72 [0.64–0.81]; accuracy: 77 %), a score ≥30 distinguished severe from moderate cases (AUC = 0.83 [0.90–0.87]; accuracy: 77 %), and a score ≥37 distinguished extreme from severe cases (AUC = 0.82 [0.77–0.87]; accuracy: 80 %). The classification performance of the cutoffs was modest in the holdout sample (62 %), but consistent cutoffs were found across sexes/genders, age groups (children and adults), and participants from Europe and the United States.

Conclusion

These BDD-YBOCS severity cutoffs can be used for clinical and research purposes across different populations but should not be used as the sole basis for important clinical decisions affecting individual patients. The boundary between subclinical and clinical BDD will require further study.