Evidence-based framework for identifying opioid use disorder in administrative data: A systematic review and methodological development study
Robert W. Hurley, Khadijah T Bland, Mira D Chaskes, Daniel J. Guth, Elaine Hill, Meredith C B Adams
Abstract
OBJECTIVE: To systematically evaluate existing approaches for identifying opioid use disorder (OUD) in administrative data sets and develop evidence-based recommendations for standardized identification methods. DESIGN: Systematic review following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Scoping Review guidelines with comprehensive literature search and evidence synthesis for framework development. SETTING: Administrative data sets including commercial claims, Medicaid, Medicare, and electronic health records. SUBJECTS: In brief, 169 studies using administrative codes to identify OUD, primarily from US healthcare systems (94.7%). METHODS: Systematic search of EMBASE, MEDLINE, Google Scholar, and PubMed through February 2024. Three independent reviewers screened articles and extracted data using standardized tools. Study quality was assessed using modified Newcastle-Ottawa Scale. Framework development employed systematic integration of evidence-based components from high-quality studies. RESULTS: Our analysis of 169 studies revealed four distinct identification approaches: Direct diagnosis codes (36.7%), composite definitions (48.0%), overdose codes (10.1%), and medication-assisted treatment codes (1.2%). Commercial claims data predominated (60.4%), followed by Medicaid claims (10.1%) and electronic health records (7.7%). Multi-modal strategies incorporating both diagnostic and treatment codes showing superior theoretical foundation compared to single-method approaches. Substantial variation existed in reference periods, code requirements, and treatment verification approaches. CONCLUSIONS: An evidence-based framework incorporating diagnosis codes, specific temporal requirements, validated indirect indicators, and treatment evidence provides theoretical foundation for standardized OUD identification protocols. The framework addresses known sources of misclassification while maintaining diagnostic specificity through clinical diagnostic alignment and systematic validation research programs. REGISTRATION: Prospero (CRD42023406173) and OSF (osf.io/ru4j3).