PubMed and Google have much in common: 1) both are freely-available search interfaces that help people locate digital information on the Internet; 2) both were launched in the late 1990’s (PubMed in 1996 and Google in 1998), just as access to the World Wide Web was becoming more widely available; 3) and both are high-traffic tools on which millions of searches are performed daily.
But while PubMed averages about 3.5 million searches per day, Google handles about 2.3 million searches per minute. Because it has many valuable uses with its great breadth and wealth of information, Google has become essential to most in our world today. But we must all beware to not become an “Homo unius machina inquirendi” (“person” of one search engine). Despite Google’s many efforts to replicate the PubMed experience for its users by creating and continuing to develop tools like Google Scholar, PubMed is still far superior when it comes to comprehensive searching of the life sciences/biomedical literature.
PubMed was originally developed to provide free access to the MEDLINE database, which still makes up the primary component of PubMed. The MEDLINE database, first launched in its electronic format in the 1960s, now generally goes back as far as the 1940s in its online version, with the original print version (Index Medicus) going back as far as 1879. The most important and distinguishing feature of MEDLINE: its records are indexed with NLM Medical Subject Headings (MeSH®). Google’s content is not – which is what makes these two search tools hugely different when it comes to the search results that they return on a topic.
A simple search example illustrates their essential difference:
Say you were interested in reading some recent scholarly articles on the disease shingles. If you go to Google to search the Web using the keyword “shingles”, on the first page you will be presented on the most past with links to consumer health information resources aimed at patients, not to mention links (or even ads) to roof shingles. If instead you tried searching Google Scholar for “shingles” with the hopes of having more scholarly content returned, you won’t be disappointed on first glance. However, you will soon notice that the first page of results will be listing mostly academic papers authored by individuals with the surname “Shingles” and not actually discussing shingles, the disease, as you intended.
Conduct this same search by typing “shingles” in the PubMed search box, however, and you will exclusively be presented with scholarly articles on shingles the disease. In fact, you will even get articles in which the authors were French-speaking (or otherwise) and referred to the concept of shingles as “zona” because these citations records will have been indexed with the MeSH term “Herpes Zoster”. “Herpes Zoster” is the subject heading/controlled vocabulary term (ie. umbrella term) used in the MEDLINE database for this concept, and when applied to the records on this topic, will help the search engine find all relevant articles regardless of how each author chose to refer to this disease in their manuscript.
When you type in “shingles” in Google, Google searches for word matches to that string of characters that you typed into the search box. The same applies to Google Scholar, except that Google Scholar ranks the matches made to author names higher in the results list.
When you type in “shingles” in PubMed, however, the search engine translates your search query into this more comprehensive search strategy:
“herpes zoster”[MeSH Terms] OR (“herpes”[All Fields] AND “zoster”[All Fields]) OR “herpes zoster”[All Fields] OR “shingles”[All Fields]
Because the structured records of MEDLINE are indexed with the appropriate MeSH terms and the search engine then maps the keyword inputted by the searcher to the relevant MeSH term, PubMed has the ability to return a very comprehensive list of scholarly articles on “shingles”, regardless if the author or searcher referred to it as shingles, zona, or zoster.