Evidence-based framework for identifying opioid use disorder in administrative data: A systematic review and...

Area of Science:

• Health Services Research
• Public Health
• Data Science

Background:

• Opioid use disorder (OUD) is a significant public health crisis.
• Accurate identification of OUD in administrative datasets is essential for surveillance, research, and intervention.
• Existing methods for identifying OUD in administrative data lack standardization, leading to inconsistencies.

Purpose of the Study:

• To systematically evaluate current approaches for identifying OUD in administrative datasets.
• To develop evidence-based recommendations for standardized OUD identification methods.
• To propose a framework for improving the accuracy and consistency of OUD identification.

Main Methods:

• Systematic review following PRISMA-Scoping Review guidelines.
• Comprehensive literature search of EMBASE, MEDLINE, Google Scholar, and PubMed.
• Evidence synthesis and framework development integrating components from 169 studies.

Main Results:

• Four main approaches for OUD identification were identified: direct diagnosis codes, composite definitions, overdose codes, and medication-assisted treatment codes.
• Commercial claims data were most frequently used, followed by Medicaid claims and electronic health records.
• Multi-modal strategies combining diagnostic and treatment codes demonstrated a stronger theoretical foundation than single-method approaches.

Conclusions:

• An evidence-based framework incorporating diagnosis codes, temporal requirements, indirect indicators, and treatment evidence is proposed.
• This framework aims to standardize OUD identification protocols and address misclassification issues.
• The framework emphasizes clinical diagnostic alignment and systematic validation for enhanced specificity.