Introducing Keyword Search

Yesterday, we released keyword search in Elicit Systematic Reviews for Pro, Teams, and Enterprise users. Now, you can search for papers across PubMed, ClinicalTrials.gov, and the Elicit paper bank in a reproducible, transparent way.

For serious research, keyword search is essential. But keyword search can also be fraying to work with. You need to find synonyms, add MeSH terms, remove confounding terms, and so on.

We want to make that easier.

How Elicit keyword search works

Elicit keyword search takes your plain language research question and automatically builds a strong first keyword query for you. No more staring at a blank search box trying to figure out all the synonyms for “GLP-1 agonists.”

The Elicit-generated keyword query for “Do GLP-1/GIP agonists like tirzepatide reduce cardiovascular risk in patients with obesity?”

You can then adjust and iterate the query freely, see your search history, and compare results until you land on the best version of your review.

You can run keyword searches over precisely the corpuses available on Elicit: all research papers on Elicit (including PubMed), only papers from PubMed, and only trials from ClinicalTrials.gov. Rest assured that any query you run on Elicit will exactly match the results you’d get on PubMed or ClinicalTrials.gov.

Finally, when you generate a Report summarizing your Systematic Review, we also include the exact search string in the Methods section so anyone else who reads it could replicate it.

Keyword search is available to Pro, Teams, and Enterprise users today. To access keyword search, start a systematic review, and then click to add papers from Elicit search. For more information on how keyword search works, check out our help center.

If you’d like to upgrade to one of the eligible plans, you can do so at this link.

—

What’s next

This is just the beginning of keyword search in Elicit. We have some exciting features in the pipeline, including:

• Using semantic search results to refine your keyword searches

• Using filters to remove less relevant papers from keyword search

• Doing multiple searches and combining the results into one Review

• Analyzing query results and suggesting tweaks to improve the search