Advanced Literature Searching for Higher Degree Students of Public Health
Last update: 19 May 2011,
by Maureen Bell
When you embark on a higher degree the dreaded literature review will loom large in your early preparations. It can seem rather daunting, but if you follow a few simple rules you'll save time, and avoid a lot of confusion and frustration. Essentially you are dealing with two simple components - language, and logic. This page is to help you with both.
Language - trawling PubMed for terms and field codes
Compiling your list of terms to search
Logic - constructing your search from the terms you've collected
Useful tips - to save you time
Other databases - moving on from PubMed to Scopus (my favourite) and other sources
Systematic reviews - what they are, and some guidance for those undertaking one
Grey literature - essential searching for a systematic review
Language
One of the problems we all have is that we expect other people to call things by the same name that we do. Even if they don't, in normal conversation this is not a problem, as we rapidly process the alternatives as equivalents. So if I talk about sickness, and you talk about illness it really doesn't matter, as we still understand one another. Databases (with very few exceptions) don't work like this. Instead they search only the exact words we use (including misspellings!). This means that for comprehensive searching you need to provide them with as many possible alternatives as you can. Some of them (like PubMed and PsycINFO) even have their own indexing language, or thesaurus which needs to be considered as well. PubMed has a thesaurus called MeSH - which is simply an acronym for Medical Subject Headings.
We'll presume you're starting with PubMed, and the topic I've chosen is - health policy and the economics of obesity.
Most postgraduate topics will have more components than this, but I'm using a simple example to demonstrate general principles.
At this stage we will just explore the language you'll need to use in your search, without worrying about how many, or how few articles your retrieve, or even how relevant most of them are.
Field Codes
When searching eitherther PubMed Embase, PsycINFO, or CINAHL it's important to tell the database where to look for words - titles of articles, abstracts, the indexing language of the database. To do this you need to use field codes.
PubMed has a wide range of field codes, but those you're most likely to use are:-
[mh] to search for Mesh terms
eg. obesity[mh]
[mh:noexp] if you don't want to explode the search term
eg. economics[mh:noexp] will search only for the term Economics, not those in the list below
[majr] to restrict your search to major indexing terms only
eg. obesity[majr]
[tiab] to search for words in titles and abstracts - this is essential for searching the most recent literature
eg. obesity[tiab]
[ta] to search for (Medline) journal title abbreviations
eg. Am J Bioeth[ta]
Method
At first we won't worry about field codes - they'll come later. For a start just use your own words
- Open a Word document and set up a logic grid - you'll need a column for each concept.
Obesity
|
Economic Aspects
|
Health Policy
|
| obesity |
economics |
health policy |
- Perform a simple search, using your own words - obesity AND economics AND health policy
- Scroll through the results to see if there are any articles that look as if they're on your topic. What you'll need is one that includes
(PubMed - indexed for MEDLINE)

Very new material will be labelled either
(PubMed - as supplied by publisher)

or
(PubMed - in process)

and will not have indexing (MeSH) terms added to it.
- Once you have found a suitable article click on the title. This will display the article details, plus its abstract, and a link to enable you to view MeSH Terms.

- Clicking on the + sign will allow you to see the indexing (MeSH) terms which have been used to describe the content of the article.
The article I will use here is
Paying for obesity: a changing landscape.
Simpson LA, Cooper J.
Pediatrics. 2009 Jun;123 Suppl 5:S301-7. Review.
and the indexing is

Terms with an asterisk after them are Major terms, and represent the most important subject aspects of the article. Those without an asterisk are Minor terms, and represent less important aspects of the article's content. NOTE the asterisk here is NOT a wildcard.
Many of the indexing terms in this example have /economics as a subheading.
When searching, using subheadings is very risky, as indexing at this level is rather variable. It is preferable to use economics[mh] OR economics[sh]. This will cover both options, and will retrieve Economics as a MeSH term, as well as Economics as a subheading.
- Check it out in MeSH - there's a link to the MeSH Database in the drop down box from the Search box

- Look up Economics as a MeSH
Economics has a large range of more specific terms in the hierarchy beneath it, and PubbMed's default is to search all of these. This is called "exploding" a term, and PubMed explodes all terms with more specific sub-categories listed.
- Look up Obesity as a MeSH
Obesity also has a list of more specific subcategories
- Look up Health policy as a MeSH
Health policy also has a list of more specific subcategories
- Now update your logic grid
Obesity
|
Economic Aspects
|
Health Policy |
| obesity[mh] |
economics[mh] |
health policy[mh]
|
- Repeat this process with other articles which look relevant to see whether consistent indexing patterns emerge, and to extend your range of MeSH.
As new material in PubMed is not indexed, and as indexers, being human, are prone to error, or variation in the way they interpret the content of an article, words in titles and abstract need to be searched too.
- Add terms to the logic grid to allow for this - the list below is not intended to be comprehensive, but will give you an idea of how to build a search. Notice that when phrases are to be searched in titles and abstracts, inverted commas are not needed in PubMed.
Obesity
|
Economic Aspects
|
Health Policy
|
obesity[mh]
obesity[tiab]
obese[tiab] |
economics[mh]
economics[sh]
models, economic[mh]
economics[tiab]
economic[tiab]
health care rationing[tiab]
healthcare rationing[tiab]
cost[tiab] costs[tiab]
costing[tiab]
health expenditure[tiab]
burden of disease[tiab] |
health policy[mh]
health policy[tiab]
health policies[tiab]
health care reform[tiab]
healthcare reform[tiab]
nutrition policy[tiab] |
Logic
When you have finished collecting terms you are ready to search. PubMed requires that Boolean operators are in upper case, so it's important when searching PubMed to use AND, OR, NOT to connect your search terms. PubMed's tutorial gives a clear explanation of how Boolean logic works.
You will need to OR the individual words in each column, and then AND the groups of words together - remembering to use parentheses around the groups of terms from each column to preserve the logic of your search.
For an entertaining demonstration of the use of parentheses in searching try viewing What the heck is Boolean searching? You will never look at a peanut butter sandwich the same way again!
Alternatively, you can search each column separately, then AND the groups of words together.
Here is what your record of your search strategy should look like. It also includes notes on how to "translate" the search into Scopus.
You will need to OR the individual words in each column, and then AND the groups of words together.
This is how the search will appear :-
obesity[mh] OR obesity[tiab] OR obese[tiab]
AND
economics[mh] OR economics[sh] OR models, economic[mh] OR economics[tiab] OR economic[tiab] OR health care rationing[tiab] OR healthcare rationing[tiab] OR cost[tiab] OR cost-effective[tiab] OR costs[tiab] OR costing[tiab] OR health expenditure[tiab] OR burden of disease[tiab]
AND
health policy[mh] OR health policy[tiab] OR health policies[tiab] OR health care reform[tiab] OR healthcare reform[tiab] OR nutrition policy[tiab]
When you have long lists of terms it's much easier to search each column separately, then combine them. When searching each column separately you don't need to use parentheses around the groups of terms.
- Copy and paste the first set of terms into PubMed, and search
- Clear the search box
- Now copy and paste the second set of terms into PubMed, and search
- Clear the search box
- Now copy and paste the third set of terms into PubMed, and search
- Once you have performed the three searches click on the Advanced search option - top right of the PubMed search screen. This will display your search history, and allow you to combine the results - using AND.

Summary:
• Create a logic grid with one column for each concept
• Look for appropriate MeSH terms – remember there may be multiple MeSH terms you could use for a concept, or there may not be a MeSH term to describe a concept you are looking for.
• Look for words to search in titles and abstracts – check the Entry terms listed in MeSH for suggestions, as well as words appearing in titles and abstracts of articles you find
• Add these to your logic grid
• Perform the search
• Add limits
Useful Tips
- Open two tabs or browser windows when searching PubMed - this will allow you to have MeSH open in one, and PubMed open in the other. That way it's easy to move between looking up terms and testing searches.
- Look in the Details box (right hand side of the search screen) to see how PubMed has interpreted your search
- Click on the Title of an article to see MeSH terms used to index it
- Use the Entry Terms lists in MeSH to find words to search in titles and abstracts - remember MeSH alone will not produce a comprehensive search
- Expert Searching in PubMed
This is a single page guide produced by the Countway Library at Harvard University. Don't be put off by the "expert" - it's a very useful and easy to follow guide with lots of examples to help you to take advantage of PubMed's incredible search power.
Other databases
There may well be other databases that are useful for your topic, and I've included a list of additional databases I think will be of use for public health.
The biggest of these is Scopus, and although it doesn't have the sophisticated subject search capability of PubMed it has some extremely useful features which will allow you to extend your search with very little effort. Its greatest strength is its citation searching capacity. You can easily "translate" a PubMed search for Scopus, and while it may produce many of the same articles, there will be lots of additional material, and you will have some extra options available.
To see these additional features:-
- Try copying the title of a the following article from PubMed, and pasting it (enclosed in inverted commas) into the Scopus Search box.
"Paying for obesity: a changing landscape"

- Once the result is displayed click on the article title

This will allow you to see
- the abstract for the article
- articles which have cited this article,
- the list of references from the bibliography at the end of the article, along with a link to other articles which have cited each reference

Each reference in the list also has links to full text, as well as its own abstract and references.
Systematic Reviews
What is a systematic review?
By Pippa Hemingway and Nic Brereton. 2nd. ed. Hayward Medical Communications, 2009.
Systematic Reviews: CRD's guidance for undertaking reviews in healthcare
Guidelines developed and published by the NHS Centre for Reviews and Dissemination, which can be used as a framework for carrying out systematic reviews or used for information by organisations commissioning reviews.
|