Protocol Text Review & AI Performance Analysis
EasySLR is designed to help reviewers streamline their workflows and make evidence-based decisions faster. Two powerful features - Protocol Text Review and AI Performance Analysis
ensure your review protocol is aligned, AI-ready, and easy to evaluate.
The Protocol Text Review tool evaluates your protocol for completeness, clarity, and consistency before AI-assisted screening begins. It provides focused, actionable feedback on information that may be unclear or difficult to interpret.
Rather than rewriting the protocol, the tool highlights only the areas that may require clarification, such as undefined abbreviations, incomplete information, ambiguous wording, or conflicting criteria.
The tool reviews the criteria defined for Population, Intervention, Comparator, Outcomes, and Study Design. It can identify:
- Vague or undefined terms, such as “elderly” or “high PD-L1 expression”
- Abbreviations that have not been explained
- Missing or incomplete PICOS information
- Conflicting inclusion and exclusion criteria
- Wording that may be open to multiple interpretations
- Criteria that may be difficult to apply consistently
When potential confusion is detected, the tool recommends clearer wording or alternative phrasing. These suggestions help reviewers and AI-assisted screening tools interpret and apply the protocol more consistently.
Review all recommendations before updating the protocol. The suggestions are intended to support clearer and more reproducible screening decisions, but the review team remains responsible for defining and approving the final criteria.
How to use it:
- Go to the Tools section
- Select 'Protocol Text Review'
- Click on 'New Text Review'
- Select the Stage: Title/Abstract or Full Text
- Click Analyze Protocol
- Review suggestions:
- Apply recommended changes manually.
Tip: Use the Protocol Optimiser before launching large-scale screening or rerunning AI to ensure optimal AI performance.
2. AI Performance Analysis
The AI Performance Analysis provides a transparent way to evaluate how well the AI is performing on your dataset. It helps build confidence in the AI by letting you validate its decisions before scaling.
What it does:
- Compares AI suggestions with reviewer decisions on a screened sample
- Generates recall, and agreement metrics
How to use it:
- Go to the Tools section
- Select ‘AI Performance Analysis’
- Click ‘New Performance Analysis’
- Select the Stage and Scope
- Click 'Run Analysis'
- Review key metrics like:
- Recall (AI’s ability to identify relevant studies)
- Reviewer Agreement (how often AI and humans agree)
Note:
- For best results, run the analysis on a representative sample of articles
- If AI is disabled, performance analysis cannot be initiated
Use both together for optimal AI-guided screening:
- Start with the Protocol Review to refine your protocol
- Then run the AI Performance Analysis to evaluate how well AI applies your refined criteria
Need Help?