Thread by @dr_asadnaveed

ChatGPT can act as your librarian. Here is how to use it to do a search for your systematic review (and meta-analysis):🧵

#AcademicTwitter

A systematic search is different from a normal literature search.

It is a comprehensive, transparent, and reproducible search process to answer a specific research question.

It is a cornerstone of all systematic reviews and meta-analyses.

Let's carry it out on ChatGPT:

Type in your query as a PubMed prompt.

In this example, there are three main concepts based on the boolean operators:

Junctional Hemorrhage AND Trauma AND Treatment Strategies

The results are awe-inspiring!

I am 90% satisfied with these results. ChatGPT could've added a similar term, "junctional bleeding," in addition to hemorrhage.

It could've added a MeSH terms as well. We can also do full text search.

However, further prompts can easily fix this.

Pasting this search strategy on PubMed yields 23 results.

Adding similar terms to our main concepts with an OR can produce more papers.

You can find these similar terms using a new prompt.

Let's quickly recap of the boolean operators before moving ahead:

AND = both terms must be present

OR = either must be present

NOT = shouldn't return a result for that term

It's also good to learn the usage of parentheses, asterisks, and quotation marks.

We are now satisfied with a basic search

The next step is to translate this search to other databases for search in Embase, CENTRAL, CINAHL, and PsycINFO.

ChatGPT can easily do this. See the example below:

Here are the results.

Pasting these to CINAHL could help you discover additional papers.

You can repeat this process for all other databases.

In conclusion, ChatGPT can give you an impressive headstart on your systematic search and simplify your librarian's life.

It is ~90% correct. Always make sure to double-check your results.

Follow me on @dr_asadnaveed for my next ChatGPT hack!