The basic search design strategies that are listed below can be used across academic databases. All of them use Boolean logic (AND, OR) and search syntax. Sometimes the syntax may differ but overall the tips below can be used in any of the databases listed in the "Databases & Sources" tab.
Here is an example of a basic PubMed search using all of the elements below:
Question: What effect does chemotherapy have on diarrhea in the elderly?
P: elderly patients
C: not applicable
Search Strategy: (elderly OR senior OR aged) AND (chemotherapy OR chemo OR "adjuvant chemotherapy") AND (diarrhea OR diarrhoea)
Having an appropriate topic scope is critical to successfully finding and appraising literature. If your scope is too broad you're search will be full of irrelevant hits, but too narrow means you may find nothing - not even similar articles. Before you search you should define your research topic and the scope of your project. Several strategies to do so can be found below.
There are several different mnemonics to help you map your search. Some will be more applicable than others depending on the discipline and subfield your topic fits into.
PICO: Patient/Problem; Intervention; Comparison; Outcome
PPAARE: Problem; Patient; Action; Alternative; Results, Evidence
Concept Breakout: start by breaking your topic into its major concepts. This is similar to what you do when you have a clinical question but you are not confined to thinking about your concepts as either a patient or an intervention or a comparison or an outcome. By doing this you will be able to match concepts to keywords and controlled vocabulary terms and break your topic into discrete, searchable pieces.
Searching for literature is an iterative process. It is rare that the first search you put together will be your first and only attempt. You will want to revise and refine your search as you go along. Always keep your research topic and inclusion and exclusion criteria in mind as you refine the search.
After breaking your search out into concepts you will want to think about keywords to include.
Keywords, synonyms, and related terms are all helpful. Try thinking about how a concept would be described in the literature.
For example, it is rare that academic literature would use the term "baby" instead one could use: infant, newborn, pediatric.
For an elderly population you could use elderly, senior, or aged.
When thinking of keywords try writing them underneath your major concepts, this will help you keep everything organized and straightforward. It will also help you craft a boolean search with consistent internal logic.
Along with Boolean Logic you can also use punctuation and truncation to craft an efficient and effective search strategy.
Punctuation: use quotation marks " " around words to search them as a phrase. For example, search for "rheumatoid arthritis" with quotation marks and most databases will search for it as a phrase instead of two distinct search terms.
Truncation: in many databases the truncation symbol is the asterix * . Put an asterix in word like prevent* and it will search for prevent, prevents, prevention, preventing etc.. It is an easy way to find many forms of a word without typing in all the keywords yourself.
Example: residen* finds: resident, residents, residency
In most academic databases you combine keywords using "Boolean Logic." Boolean Logic is the process of using AND, OR, or NOT to combine search terms into search strategies. An easy way to think about Boolean Logic is that AND narrows your search, OR broadens your search, and NOT excludes results.
Sample Question: Are there studies on education for residents or medical students about ebola?
Sample Search: ("medical education" OR "curriculum") AND (residen* OR "medical student") AND (ebola OR "ebola hemorrhagic fever" OR "ebola fever")
Use AND when you want the search to find all of those words in the results.
Use OR to find any or all of your terms in the results.
Use NOT to exclude terms from your results. Use this sparingly, in particular if you are looking for a specific population because if you want to find a study that includes men if you NOT women than it will exclude studies that have both genders included in the sample. This can remove articles that may be useful.
You can build your search in a Word doc or other text editor and copy and paste it into the PubMed search box or you can use the advanced search builder. The advanced search builder has the advantage of being a method by which you can really see how your search is being effected by the terms you are putting in.
Simply put your terms in, select the 'add to history option' to the right of the query box, and search each concept individually. The 'add to history' button will add it to the search history table below so you don't have to click back and forth between the search results page to add your concepts together, After you have run a search on each concept you can add them together using the search number and the boolean operator AND to get a final result set. You must put a # (hashtag) symbol before the number of the search line or it will not add them together properly.
The tutorial below can give you more information on this process and most other databases have this function as well.
PubMed (and all academic databases) includes limits (sometimes known as filters) to help narrow your results. The most commonly used ones are study/article type, date, and language.
The limits menu is on the left of the results page. When you are done your search choose "clear all" before performing another search or these limits will continue to be applied.
Study Type is called "Article Type" in the limits menu.
Make sure to choose "customize to get the full list of article types. The ones you would commonly use would be Meta-Analysis, Randomized Controlled Trial, and Systematic Review. This would help you get the top level of evidence for the clinical question you are searching but don't be worried if there is not a Meta-Analysis or Systematic Review on your question - perhaps one hasn't been done yet!
Press the "show" button to apply the limit to your search results.
Date is a good limit to use. Often you want only the most recent literature or have a specific date-range in mind for a research project. Under the "Publication Dates" limit simply choose "custom range", put in the dates needed, and select apply.
If you cannot read the language a result is in chances are you should not use that as an example of a good search result - you cannot ascertain if its results and methodology are accurate and applicable. If you find you are seeing a lot of non-english results, shown by article's whose titles are in square brackets [ ], try using the language filter. It is under the show additional filters list. Simple choose "Language" and click the "show" button - this will add language to the limits menu and you will be able to select English.
Sex & Age
Be cautious when using either the Sex or Ages limits in PubMed. They are not always evenly applied and you may be removing studies that examine both sexes or multiple ages by choosing a specific sex or age that could be useful. Use this as a last resort if you are getting a lot of unhelpful hits in your search results based on sex or age.
Documenting your search is essential. Nothing is worse than crafting the perfect only for your browser to crash and you lose it all. It also helps with reproducibility later on and allows you to refine the search in a controlled manner.
You will want to record where you searched, the search strategy for each location, any limits used (date for example), and number of results. If you created a table of key terms that can be really helpful to include and you can refine it overtime as you find more relevant materials.
How your document your search is personal preference but using some sort of text editor like Word or a GoogleDoc for a group project can be very helpful.