Overview

Abstrackr is a free, web-based tool designed to streamline the title and abstract screening phase of systematic reviews. It uses machine learning to predict which references are most likely to be relevant as you screen.

Abstrackr is especially useful when working with large citation datasets, allowing reviewers to focus effort where it matters most.

Website: https://abstrackr.cebm.brown.edu

Key Features

  • Semi-automated abstract screening
  • Learns from your inclusion/exclusion decisions
  • Can prioritize articles for manual review
  • Supports collaborative screening
  • Works best with datasets of 200+ citations

When to Use Abstrackr

Use Abstrackr when:

  • You need to screen hundreds or thousands of references
  • You want to speed up the process while retaining control
  • You have multiple reviewers
  • You want to prioritize screening for likely-includes

It is not used for full-text review or post-inclusion data extraction.

How to Use Abstrackr

1. Create an Account

Go to https://abstrackr.cebm.brown.edu and sign up for a free account.

2. Create a New Project

  • Name your review project
  • Choose single or dual-screening mode
  • Optionally, invite collaborators

3. Upload Citations

Supported formats include:

  • .RIS
  • .TXT (tagged citations)
  • .XML (e.g., from EndNote or PubMed)

4. Begin Screening

  • Mark each abstract as Include, Exclude, or Unsure
  • You may also add reasons for exclusion if desired

5. Machine Learning in Action

  • Abstrackr observes your screening decisions
  • It updates its predictions after ~50–100 abstracts
  • The more you screen, the more accurate the model becomes

6. Prioritize Screening

Once predictions are available, you can:

  • Focus on high-likelihood Includes
  • Stop early (if using active learning strategies)
  • Export the final screened dataset

Export Options

When finished, export your decisions as a .CSV file for:

  • Documentation in PRISMA diagrams
  • Loading into full-text screening tools (e.g., Covidence, CADIMA)
  • Integration with other tools in your review pipeline

Strengths & Limitations

Strengths Limitations
Free and easy to use Web-only; no API or offline use
Speeds up large screenings Occasional bugs or downtime
Learns from your inputs Predictions may vary by topic
Prioritizes likely includes Export format can be tricky