Systematic Review and Meta-Analysis: An Empirical Approach to Defining Treatment Response and Remission in Pediatric Obsessive-Compulsive Disorder.

Abstract

OBJECTIVE\nA lack of universal definitions for response and remission in pediatric obsessive-compulsive disorder (OCD) has hampered the comparability of results across trials. To address this problem, we conducted an individual participant data diagnostic test accuracy meta-analysis to evaluate the discriminative ability of the Children s Yale-Brown Obsessive-Compulsive Scale (CY-BOCS) in determining response and remission. We also aimed to generate empirically derived cutoffs on the CY-BOCS for these outcomes.\n\n\nMETHOD\nA systematic review of PubMed, PsycINFO, Embase and CENTRAL identified 5,401 references, 42 randomized controlled clinical trials (RCTs) were considered eligible and 21 provided data for inclusion (N 1,234). A score ≤ 2 in the Clinical Global Impressions Improvement and Severity scales were chosen to define response and remission, respectively. A two-stage random-effects meta-analysis model was established. The area under the curve (AUC) and the Youden Index were computed to indicate the discriminative ability of the CY-BOCS and to guide for the optimal cutoff, respectively.\n\n\nRESULTS\nThe CY-BOCS had sufficient discriminative ability to determine response (AUC 0.89) and remission (AUC 0.92). The optimal cutoff for response was a ≥ 35% reduction from baseline to posttreatment (sensitivity [95% CI] 83.9 [83.7, 84.1]; specificity [95% CI] 81.7 [81.5, 81.9]). The optimal cutoff for remission was a posttreatment raw score ≤ 12 (sensitivity [95% CI] 82.0 [81.8, 82.2]; specificity [95% CI] 84.6 [84.4, 84.8]).\n\n\nCONCLUSION\nMeta-analysis identified empirically optimal cutoffs on the CY-BOCS to determine response and remission in pediatric OCD RCTs. Systematic adoption of standardized operational definitions for response and remission will improve comparability across trials for pediatric OCD.