Accountability, AI limits, and methodology

Who is responsible: Methodological content on this site is led by George Burchell (see ORCID). Claims about AI-supported appraisal should be weighed against your protocol, registry entries, and independent methodological oversight.

AI disclaimer: Automated or AI-assisted steps may misread context, omit contradictory text, or reflect training-data gaps. This feature is not a validated medical device and has not been peer-reviewed as a standalone appraiser. Treat outputs as draft support only; two independent human reviewers should confirm domain judgments for publication-grade systematic reviews, with disputes and rationales recorded in your audit trail.

For conceptual guidance on quality assessment in end-to-end workflows (tool choice, GRADE alignment, limitations of rapid and scoping reviews), see our methodology article. The companion SaaS Evidence Table Builder focuses on structured evidence tables and extraction; start there for table-centric workflows, then read its How to Choose the Right Quality Assessment Tool walkthrough for instrument selection mechanics.

Core references (verify current versions)

About this tool

Quality Assessment

Apply RoB 2 or another checklist to uploaded studies and export an Excel summary of judgments. Reviewers remain accountable for each domain.

Watch the tool

Video demo coming soon

A walkthrough of this tool will appear here after it is uploaded to YouTube.

How this page differs from generic “AI PDF” tools

Risk-of-bias work is protocol-driven: the instrument version, eligible outcomes, and adjudication rules should be locked before anyone opens a full text. This tool is meant to accelerate drafting and consistency checks—not to replace two independent human appraisers on publication-grade reviews.

Pair uploads with the methodology article on quality assessment and tool choice.
Run deduplication and screening first so appraisal time is not wasted on records you will never include.
When judgments need to become supplementary evidence tables, many teams export structured rows into Evidence Table Builder for column-locked templates and QC passes that match funders and journals.

Read the amber Accountability, AI limits, and methodology panel above before relying on any automated domain text in a protocol or manuscript.

What the tool looks like

Assessment process

01
Select Tool
Choose the appropriate quality assessment framework
02
Upload Studies
Upload PDF files of studies to be evaluated
03
AI Analysis
Automated quality assessment using AI
04
Get Results
Download comprehensive quality assessment report

Available Assessment Tools

Risk of Bias 2.0 (RoB 2)

For randomized controlled trials, evaluating bias across multiple domains

ROBINS-I

Risk of bias assessment for non-randomized intervention studies

QUADAS-2

Quality assessment tool for diagnostic accuracy studies

Newcastle-Ottawa Scale

For assessing quality of observational studies in meta-analyses

Open quality assessment

Apply RoB 2, ROBINS-I, QUADAS-2, or NOS-style checklists and export judgments to Excel.

The working tool is in the dashboard, not on this page.