Screening Guide

Title & Abstract Screening

Best practices for the most time-consuming phase of your systematic review — and how to do it faster without sacrificing quality.

What Is Title & Abstract Screening?

Title and abstract screening is the first filter in a systematic review. You read each paper's title and abstract and decide whether it could be relevant based on your inclusion criteria. Papers that pass move to full-text review.

🎯 Key Principle: At this stage, be liberal. When in doubt, include the paper. You'll make the final decision during full-text review. It's better to read 20 extra full texts than to miss a relevant study.

How to Screen Effectively

1

Pilot Test Your Criteria

Screen 50–100 papers together as a team first. Discuss disagreements and refine your criteria before starting the full screening.

2

Read Title First, Then Abstract

Many papers can be excluded on title alone (e.g. clearly wrong population or topic). Only read the abstract when the title is ambiguous.

3

Use a Three-Decision System

Include (clearly relevant), Exclude (clearly irrelevant), Maybe (needs discussion or full text). The "Maybe" category reduces premature exclusions.

4

Screen in Batches

Screen 100–200 papers per session. Fatigue leads to inconsistent decisions. Take breaks and maintain focus.

5

Document Everything

Record reasons for exclusion. You'll need these for your PRISMA flow diagram.

Single vs. Dual Screening

Single Screening Dual Screening
How it works One reviewer screens all papers Two reviewers screen independently
Speed Fast Slower
Reliability Lower Higher
Best for Rapid reviews, scoping reviews Cochrane reviews, journal publications

📖 Read our detailed Dual Screening guide →

Using AI to Screen Faster

Active learning tools can dramatically reduce screening workload:

50–80%

Less screening workload

95%+

Recall of relevant papers

10x

Faster than manual screening

Screen Smarter, Not Harder

Lumina's AI prioritizes the most relevant papers first, so you find what matters without reading thousands of irrelevant abstracts.