Abstract Screening Tool
Load the records from your database search, read each title and abstract on its own card, and press I, M or E to include, mark as maybe or exclude. Record an exclusion reason with one key, highlight your own keywords, and export the decisions as a CSV and the included records as RIS, with the counts ready for a PRISMA flow diagram. It runs in your browser; nothing is uploaded. Free, no sign-up.
Paste the text or upload a file. A CSV needs a Title column and may have Abstract, Authors, Year, Journal, DOI and Keywords columns; a CSV saved by this tool also brings back your decisions. Records are read in your browser — nothing is uploaded.
Paste or upload some records first.
How to Use the Abstract Screening Tool
- 1. Load your records. Paste or upload RIS, CSV, a PubMed .nbib file or an EndNote XML export. Records with neither a title nor an abstract are skipped.
- 2. Screen one record at a time. Read the title and abstract, then press I to include, M for maybe or E to exclude. Pressing 1 to 7 excludes with that reason. The screen advances itself.
- 3. Add highlight words if they help. List words that suggest a record fits and words that suggest it does not. They are marked in green and red in the title and abstract.
- 4. Export. Download the decisions as a CSV — it is also your save file — and the included records as RIS to carry on to full-text screening. Copy the counts for your flow diagram.
What Title and Abstract Screening Is
In a systematic review, the first screening stage reads the title and abstract of every record from the searches and discards the ones that clearly do not meet your eligibility criteria. Records that pass go on to full-text assessment. Because the stage handles hundreds or thousands of records, it needs a quick, consistent routine.
- Include — the record appears to meet the criteria and goes to full-text assessment.
- Maybe — the abstract is too short or unclear to decide. It also goes to full text.
- Exclude — the record clearly fails at least one criterion. Record which one.
At this stage, err on the side of keeping a record when in doubt: a wrongly kept record costs a few minutes at full text, while a wrongly excluded one is lost from the review for good. Most protocols also ask for two reviewers working independently, as PRISMA 2020 item 8 expects the report to say.
Keyboard Shortcuts
| Key | Action |
|---|---|
| I | Include |
| M | Maybe |
| E | Exclude |
| 1–7 | Exclude with the reason of that number: population, intervention or exposure, outcome, study design, publication type, language or date, duplicate |
| U | Clear the decision on this record |
| Z | Undo the last decision and return to that record |
| ← and → | Previous and next record in the current view |
The shortcuts work whenever no text box or menu has the cursor.
Loading and Saving
- RIS — title (TI or T1), abstract (AB or N2), authors, year, journal, DOI and keywords are read.
- CSV — the first row must name the columns. A column headed Title or Article Title is required; Abstract, Authors, Year, Journal or Source Title, DOI and Keywords are used when present.
- PubMed .nbib and EndNote XML — read through the same converters as the NBIB Converter and EndNote XML Converter.
- Resuming. The decisions CSV has Decision, Reason and Note columns first. Load that file again and your decisions come back, so you can stop and carry on. Nothing is saved on the page itself.
A Worked Example
The sample button loads five invented records on exam anxiety. The fifth repeats the first — the same title and year — and the card says so. One reasonable way to screen them:
| Record | Decision | Reason |
|---|---|---|
| 1. Mindfulness training for exam anxiety in undergraduates: a randomised trial | Include | — |
| 2. Mindfulness and stress responses in laboratory rats | Exclude | Wrong population |
| 3. Teachers' views on supporting anxious students: a qualitative interview study | Exclude | Wrong study design |
| 4. Interventions for test anxiety: a systematic review | Maybe | — |
| 5. Repeat of record 1 | Exclude | Duplicate |
The tool then reports: Records screened: 5. Records excluded: 3. Reports sought for retrieval: 2 (1 included, 1 maybe). The exclusion reasons table shows one each for wrong population, wrong study design and duplicate. A systematic review found at title and abstract stage is usually not itself included, but its reference list is worth checking — see the Citation Chaser.
What It Does Not Do
- It does not rank or predict relevance. Every decision is yours. The highlight words are a plain keyword match, not a model.
- It is one reviewer at a time. For independent double screening, each reviewer loads the same file, screens, and exports a CSV; compare the two in a spreadsheet and resolve the conflicts by discussion.
- It is not a duplicate remover. It flags records sharing a DOI, or a title and year, but removing duplicates properly is best done in a reference manager before screening.
- It does not retrieve full texts or assess risk of bias. For the latter see the Risk of Bias Plot Generator.
What Is Sent, and Where
Nothing. Your records are read and screened by code running in your browser; no part of them goes to CitationEasy or any other service. Because nothing is kept after you leave, download the decisions CSV before you close the page.
Frequently Asked Questions
- What is title and abstract screening?
- It is the first selection stage of a systematic review, when you read each record's title and abstract and decide whether it could meet your eligibility criteria. Records that might are retrieved in full for a closer second stage.
- How do I screen abstracts for a systematic review for free?
- Export your search results from each database as RIS or CSV, combine them, load them here, and screen with the keyboard. Export the decisions at the end and use the counts in your PRISMA flow diagram.
- Should I choose maybe or include when I am unsure?
- Either keeps the record for full-text assessment. Use Maybe to mark the ones where the abstract was too thin to judge, so you can look at them first when you retrieve full texts.
- How many reviewers should screen the abstracts?
- Two independent reviewers is the usual standard for a full systematic review, and PRISMA 2020 asks you to say how many screened each record and whether they worked independently. Check your protocol or supervisor, since some projects allow one reviewer with a check on a sample.
- Can I stop and carry on later?
- Yes. Download the decisions CSV, and when you return load that file in place of the original. Your decisions, reasons and notes come back, and unscreened records are still waiting.
- How do I record exclusion reasons for PRISMA?
- Choose a reason with the number keys or the menu when you exclude. At full-text stage PRISMA asks for reasons for excluded reports; recording them at abstract stage as well makes the final tally easier.
- Is my search data uploaded?
- No. The records are read and screened entirely in your browser.