Systematic Reviews

Data Extraction in Systematic Reviews

Data extraction is the process of systematically collecting relevant information from included studies to answer the review question. This typically includes study characteristics (e.g., design, population, interventions), outcomes, and results. To ensure accuracy and reduce bias, data extraction should be performed by two independent reviewers, with discrepancies resolved through consensus.

Best practices include:

• Using a standardized, piloted data extraction form
• Extracting data in duplicate
• Documenting decisions and handling missing or unclear data
• Tailoring fields to match the review’s objectives

Cochrane Handbook: Data Collection Standards
JBI Guidance on Data Extraction

Data Extraction in Covidence

Covidence offers two tools for data extraction: Extraction 1 and Extraction 2.

• Extraction 1 is optimized for intervention systematic reviews and includes structured sections for methods, population, interventions, and outcomes. It supports automatic suggestions based on study metadata and full-text PDFs, saving time while maintaining control.
• Extraction 2 is more flexible and suited for scoping, qualitative, or mixed-methods reviews, allowing custom fields, tables, and multiple-choice formats.

Key features:

• Dual or single reviewer workflows
• Consensus review and conflict resolution
• Exportable data for synthesis or use in RevMan
• Customizable templates for different review types

How to Complete Data Extraction in Covidence
Creating a Data Extraction Template
Extraction 1 vs. Extraction 2 Comparison