For Medicine, Nursing, & Health Sciences
Most databases store their filters on the left-side of your results page. Below are some recommended filters to consider using when looking to further define your search.
Stuck looking for information? Here you will find strategies and techniques to help with database searching. Experiment with these strategies and techniques across multiple databases. What doesn't work in one place might work in another.
Lieberum et al (2025) published a scoping review that examines how large language models (LLMs)—especially tools like ChatGPT—are being used to support the process of conducting systematic reviews in health research. It finds that while LLMs show promise in tasks like literature searching, study selection, and data extraction, most applications are still experimental and require human oversight and further validation before they can be reliably integrated into research workflows. Chen and Feng (2025) explored the use of generative AI in systematic searching. They conclude that AI could expand the scope of search terms, automating repetitive aspects of the systematic-review process, while human expertise remains crucial in refining search terms and ensuring methodological rigor.
In this section, we will focus on how Copilot and ChatGPT can assist with literature searches, highlighting its strengths, limitations, and appropriate use cases. If you are interested in AI use of other tasks in systematic reviews, please check the scoping review for more details.
(Note: This is a rapidly evolving topic, and I'm sharing what I've found so far, based on current resources and my own experience.)
The following slide summarizes AI tools' capabilities in the conduction of a systematic review (Bridgeman, 2025).

From my work experience, I've found that Copilot and ChatGPT can be very helpful for identifying research concepts and finding keywords—such as alternative terms and synonyms—especially when you're stuck or unsure the concepts and keywords you identified. In this process, AI tools may also generate MeSH terms, they don't search in MeSH database. There’s a chance that AI tools will give fabricated or incorrect MeSH terms. Keep in mind, always double check the concepts, keywords and MeSH terms given by an AI tool. Select the keywords that you want to include in the search.
For example, you might ask Copilot or ChatGPT using the following prompt:
"I am doing a literature review for this question: "For athletes with lateral ankle sprains, do surgical ligament repairs reduce sprain recurrence?". What are the concepts and search terms for this question?"
Currently, search queries generated by AI tools are often suboptimal. If you copy and paste them directly in PubMed, you'll likely notice they return fewer results than expected. Once you collect search terms and know your concepts, it's easy to do a search in PubMed. Then you will get the search query from PubMed.
In summary, Copilot and ChatGPT are very helpful for finding keywords and identifying concepts in your research question. Use them to find keywords and concepts when you are unsure. Then go to PubMed and other databases to do literature searches.
If you are interested in AI tools in research, please visit Purdue AI guide. Remember that publishers require you to disclose AI use in your research. See JAMA Instructions for Authors for more details.
References
Bridgeman, M. (2025, September 24). AI, Librarians, and Covidence: Building Trust in Systematic Reviews [Webinar]. Covidence. https://us02web.zoom.us/webinar/register/WN_yrSn4n9zQJi6GjgC_BLV-Q#/registration
There are a number of different sources that can be searched for literature, including:
Adapted from RMIT University's Systematic Reviews in Health Research Guide.
When searching, keywords can be used to...
Databases that use subject headings
![]() |
Examples:
|
![]() |
Examples:
|
![]() |
Examples:
|
A search for attention deficit disorder = attention AND deficit AND disorder -- without phrase searching
A search for "attention deficit disorder" = "attention deficit disorder" -- with phrase searching
Example: Child* = child, child's, childhood, children, children's,
Note: The placement of the truncation symbol can impact the type of results you get back.
Example: If I searched Hum* with the expectation of getting versions of 'Human', I'd also get....
Examples:
Also known as Medical Subject Headings, MeSH is the U.S National Library of Medicine's (NLM) controlled vocabulary used to index journals in MEDLINE/PubMed. These terms provide a consistent way to find articles that have utilize different terminology for the same concepts. MeSH is assigned to each articles by by subject analysts when applicable. It typically can take anywhere from 3 to 6 months for MeSH terms to be applied to an article, creating a delay.
The MeSH Database is accessible directly from PubMed's homepage. In the MeSH database, you can:
In the next tab, I will outline how to search for MeSH terms in PubMed. For an alternative experience, please check out this video from NCBI that walks you through using the MeSH database. Note: this video was made prior to the update of PubMed. Although PubMed looks different now, the MeSH database has not changed.
The easiest way to access the MeSH database is by using the link on PubMed's homepage, located right under the Explore icon.

Search for one term at a time. Although the MeSH database is extensive, it is important to note that not every term is going to have a corresponding MeSH term.
Results will be displayed with brief definitions, ranked by relevance. Select the entry that best fits your needs to view the full record page. If there is only one result available, you will be brought directly into the record

At the top of the MeSH recored page, you will see the term's definition and a list of subheadings. Scroll down to view Entry Terms (if available) and the MeSH Hierarchy.

There are three possible options for refining your search with the MeSH record:
When you are ready, click the "Add to Search Builder" button located right under the PubMed Search Builder box, found on the right side of the screen. If you are using any subheadings, be sure to check their boxes prior to select "Add to Search Builder".
If you are ready to search for your MeSH term in PubMed, select the Search PubMed.

| Keyword | Medical Subject Heading (MESH) | CINAHL Subject Heading |
| Heart attack | Myocardial infarction | Myocardial infarction |
| Distance education | Education, non-traditional | Education, distance |
| EHR | Computerized Patient Record | Electronic Health Records |
PRESS Peer Review of Electronic Search Strategies (2016)
McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. Journal of Clinical Epidemiology, 75, 40-46