Covidence will check for duplicates in the first set of references that are imported and also in the order in which they are imported. The first unique occurrence of a reference becomes a primary reference and all subsequent instances are duplicates. Covidence then checks within each subsequent import and also against all previous imports for duplicates.

It checks on the following fields:


The titles must match exactly after basic normalisation, ignoring punctuation and special characters, but maintaining Unicode. 

Where there is a non-English character, there is no conversion.

For example, the citation:

Dabigatran-related coagulopathy: When can we assume the effect has "worn off"?

Would be normalised to:


As would the citation:

Dabigatran related coagulopathy (When can we assume the effect has worn-off?)

Meaning that it would be detected as a duplicate.

However, the citation:

Dabigatran-related coagulopathy

Would not be detected as a duplicate.


   The four digit field for year must match exactly.


   The volume must match exactly, ignoring special characters, such as brackets.


   The authors must be sufficiently similar in order to be detected as a duplicate. They do not have to be exact matches.