How to Remove Duplicate References
Combine records from all databases in one reference manager and run its duplicate-detection tool. Then check borderline matches manually, merge or remove true duplicates, and record how many were removed so the PRISMA flow diagram can report accurate counts.
Steps
- Export each database's results in a standard format and import them into one library (keep a copy of the raw exports).
- Run automatic de-duplication in your reference manager. Tools match on fields such as title, authors, year and DOI; settings differ between programs.
- Review the proposed duplicate sets. Be careful with similar titles from different papers, conference abstracts and their full papers, and records with missing fields.
- Merge or remove true duplicates and keep the more complete record.
- Run a second pass with sorting by title or DOI to catch records the tool missed.
- Record the numbers: records identified, duplicates removed and records screened.
Published guidance describes a stepwise manual approach for EndNote that is also a good model for other tools.
Why it matters
Duplicates inflate the number of records to screen and, if not caught, can cause the same study to be counted twice in a review. Different reports of the same study (for example a trial and its follow-up paper) are different issue: link them as one study during data extraction.
Illustrative example
Invented for teaching. A team searches four databases and gets 4,300 records. After automatic de-duplication, 3,100 remain. A manual pass removes 85 more, leaving 3,015 records to screen. They record each number for the flow diagram.
Common pitfalls
- Deleting records instead of marking them, which makes counts hard to audit.
- Treating different papers about the same study as duplicates and losing data.
Key takeaways
- De-duplicate, check manually and record counts.
- Keep the raw exports.
References and further reading
- Bramer WM, Giustini D, de Jonge GB, Holland L, Bekhuis T. De-duplication of database search results for systematic reviews in EndNote. J Med Libr Assoc. 2016;104(3):240-243. doi:10.3163/1536-5050.104.3.014
- Zotero documentation. Link
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71
- Mendeley Reference Manager (Elsevier). Link
Related reading
This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.
