It's a good practice to start by looking for reviews. Include the word "review" or the phrase "systematic review" with your other search words. PubMed has a "Reviews" button on the upper left.
Choosing search words is a challenge for researchers. Sometimes, there is not a good fit between the possible words and the concepts. Other times, you just don't know what the best words are at first.
1. Keep it simple. Choose a word or short phrase (in quotation marks, if it helps) for each concept. Which words distinguish your concept? Many words you might use could show up in many articles on lots of topics. Don't bother with those, if you can avoid them.
2. Before you start, think. What words would researchers use? Think about narrower words, broader words, and words that are related. Consider jotting them down or making a document that you can copy and paste from.
3. Watch for alternative words as you go. Look at the subject headings.
4. As you go, think about adding more words to get fewer results that are more focused on your topic. Or try taking words out or substituting in words with broader meanings, in order to get more results. You can search on authors, methodologies, data sources, outcomes, or almost anything else of interest.
5. Consider searching in the CINAHL Headings and PubMed MeSH databases before you start really searching for articles.
6. You can ask for help also.
7. Above, I suggested using just one word or short phrase for each concept. If you have synonyms or related terms, you can search on them at the same time using OR.
examples: (teenagers OR adolescents OR youths)
("eating disorders" OR bulimia OR anorexia)
(carolina OR virginia OR appalach*)
You can also use the asterisk, to search for different variations of a word.
Example: theor* will find all these: theory, theories, theoretical
Helps with creating an answerable, useful question. Helps with choice of search words.
Problem or population -- soccer players
Intervention -- preventive training (or more specific type of preventive training)
Comparison treatment (or placebo) -- alternate type of preventive training? what's being used up to now.
Outcome -- count frequency and severity of injuries. Missed games or training.
Most of the time, our systems for delivering full text work smoothly. Occasionally they do not, sometimes we do not subscribe, and sometimes you're starting with a paper citation. What do you do?