How Do You Find Research Published Since Your Last Search?

How can you find only the research that appeared after your previous search?

You conducted a literature search three months ago, screened the results, and thought you were finished. Now you need to check what has appeared since then. Rerunning the original search may return hundreds or thousands of records you have already seen, while simply filtering for the current year can miss relevant studies.

The task sounds straightforward: search for everything published after the date of your previous search. In practice, publication dates and database availability do not always line up neatly. A paper can be published before your cutoff but added to, indexed in, or fully processed by a database afterward.

A reliable update therefore depends on something researchers occasionally discover only after wishing they had written it down: exactly what was searched, where it was searched, and when.

Start from your previous search, then identify what is genuinely new

A literature search has more than one relevant date

Your last search date is the starting point

Before updating anything, establish the date on which you last ran the search in each database. Ideally, your search documentation should also preserve the database or platform, exact search string, filters and limits, and number of results retrieved.

That information lets you reproduce the earlier search rather than reconstructing it from memory. It also gives you a defensible boundary for determining what counts as potentially new.

If you are working on a long project, maintaining this record makes it much easier to decide how often the literature search should be updated without repeatedly rebuilding the search process.

Publication date and database discovery date are not the same thing

A common approach is to rerun the original query with a publication-date limit beginning after the previous search. That can be efficient, but the logic has an important weakness.

Suppose you searched a database on March 1. A paper published on February 20 was not yet represented in that database when you searched. The database adds the record on March 10. If your update searches only for papers with publication dates after March 1, that paper can fall outside the update even though it is new to your search results.

This is why an update should distinguish between newly published and newly retrievable research. Those categories often overlap, but they are not identical.

Rerun the same search before redesigning it

If your research question and eligibility criteria have not changed, the cleanest update usually begins with the search strategy you previously used. Changing keywords, Boolean logic, fields, or filters at the same time as changing the date range makes it difficult to know whether additional records are genuinely new or simply products of a different search.

This does not mean a search strategy can never improve. You may discover a missing synonym, a new term may emerge in the field, or the scope of the project may change. But those are search-strategy revisions, not merely updates. Document them separately.

Saved searches can remove much of the repetitive work

Major scholarly databases may provide saved-search and alert functions. Scopus, for example, allows signed-in users to save searches and create search alerts; its documentation also describes running a saved search to see results since it was last run. ScienceDirect search alerts can notify users when a stored search retrieves new results.

These functions are particularly useful when you expect to return to the same query repeatedly. Instead of reconstructing the search each time, you preserve the underlying retrieval logic and examine the records added later.

If the project is still active, converting important searches into database search alerts can make subsequent updating considerably easier.

Use a small overlap when missing late-indexed records would matter

For a rigorous search update, an overlapping date range can be safer than starting exactly one day after the previous search. The overlap deliberately retrieves some records you have already screened, but it also reduces reliance on the assumption that publication and database indexing happened in perfect chronological order.

The appropriate overlap is not universal. Database behavior, discipline, publication practices, and the purpose of the search matter. For an informal scan of recent developments, some duplicate retrieval may simply be inconvenient. For a systematic evidence synthesis, avoiding missed eligible studies is more consequential.

Deduplication is part of updating the search

An overlapping search, or even a full rerun, will naturally retrieve records you have seen before. That is not necessarily a problem if you have a reliable way to identify them.

Export the new retrieval into your reference manager, review software, spreadsheet, or other screening system and compare it with the previous set. DOI, database accession number, title, author, and year can help identify duplicate records, although no single field is perfect in every dataset.

The practical objective is not to make the database return zero old records. It is to distinguish efficiently between previously processed records and records that require new screening.

Do not confuse an update with a search for only the newest papers

If your goal is to update a previous literature search, the relevant interval begins with the previous search, not with an arbitrary definition of “recent.” A study published eight months ago could still be new to your evidence set if your last search occurred nine months ago.

Similarly, deciding whether research counts as current is a different question from determining whether it appeared since a previous search. One concerns substantive currency; the other concerns retrieval history.

Updating a literature search after four months

Notice that the workflow does not depend on manually remembering which article titles looked familiar. Your previous search record becomes the checkpoint against which the update is performed.

Why seemingly simple search updates can miss studies

Choose the update method according to what is at stake

You do not need the same level of search control for every project. Checking recent papers for a lecture is different from updating the evidence search for a systematic review before submission.

If you expect several updates during the project, the next efficiency gain is to design the search so that you do not have to rebuild and rerun everything manually each time.

Before running your next search update

Questions about finding research since a previous search

Make every search leave a usable checkpoint

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.