AI-assisted literature review

How to manage a systematic literature review from citation import to data export using the core features of the Covidence platform

Covidence is a web-based software designed to streamline the process of conducting systematic and literature reviews. It provides a structured workflow for research teams to collaboratively manage tasks such as screening citations, reviewing full-text articles, extracting data, and assessing quality. This specialized tool helps ensure a rigorous and transparent review process, saving significant time and reducing the administrative burden on researchers.

1. 00:50 Invite collaborators to your review. Navigate to Settings > Reviewers and click Invite another reviewer. Enter their first name and email address to send an invitation to join the project.
2. 01:09 Import references for screening. From the Review Summary page, click the Import button. Select the stage to import into (e.g., Screen), choose your reference file, and click Import. The system will automatically detect and manage duplicates.
3. 01:25 Screen titles and abstracts. In the Title and abstract screening section, vote Yes, No, or Maybe for each reference based on your inclusion criteria. Use the available tools to filter references, highlight keywords, or add notes.
4. 02:00 Resolve screening conflicts. Go to the Resolve conflicts tab to see references where reviewers disagreed. As the conflict resolver, make the final Yes, No, or Maybe decision to move the study to the next stage or exclude it.
5. 02:12 Conduct the full-text review. For studies that passed the initial screen, proceed to the Full text review stage. Read the full article and decide whether to Include or Exclude it. If excluding, you must select a reason.
6. 02:30 Create or customize data extraction and quality assessment templates. Navigate to the Extraction section. Click Data Extraction Template or Quality Assessment Template to design the forms your team will use. You can build a template from scratch or modify the default provided.
7. 02:54 Resolve data extraction and quality assessment conflicts. After two reviewers have completed extraction for a study, a third reviewer must complete Consensus. Review the conflicting data points side-by-side and select or enter the correct final data for each field.
8. 03:34 Export your review data. Click the Export button at the top of the Review Summary page. Select the data you wish to export (e.g., References, Inter-rater reliability, Data extraction) and choose the desired format to download the file.

Related skills

• Apply human judgement to validate AI outputs — The systematic review process in Covidence requires researchers to apply structured human judgment at every stage, from screening titles to resolving conflicts in data extraction.
• Validate AI outputs against sources — The full-text review and data extraction stages are exercises in validating information by critically assessing primary research articles, which serve as the authoritative sources.
• Locate and manage supported GenAI tools — This tutorial shows how to navigate and manage a specialized academic software platform to conduct a complex research task.
• Integrate multiple tools into workflows — This process demonstrates how to integrate Covidence into a larger research workflow by importing references from bibliographic management tools or databases.

Software used in this activity