Systematic Review (SR) searching adopts both systematic and comprehensive approaches with the goal of retrieving, ideally, all the literature relevant to the focused question at the base of your Systematic Review. Typically, an expert searcher, such as an information professional, uses a combination of keywords qualified with field tags (e.g. [tiab] field tag in PubMed related to title and abstract fields of PubMed records) and subject headings (e.g. MeSH in PubMed and EmTree in Embase) for SR searching.
When selecting the terms for an SR, it is best to focus strictly on the terms directly related to the subject or clinical question being addressed. Occasionally it can be appropriate to expand the search to a slightly broader focus to retrieve literature where the exact subject matter may be discussed within the context of other subjects within a broader question.
An example of this expanded search: If the SR is focused on breast cancer surgery, a broader focus would be to look at any/all cancer surgeries wherein breast cancer specific surgery may be discussed.
All search approaches, whether broad or narrow, must be reflected in transparent and reproducible documented search strategies. It is important to remember that there is no “perfect” comprehensive search strategy that will only retrieve relevant citations. It is to be expected that any search, especially a comprehensive SR search, will retrieve many more citations than are actually relevant to the question being asked. Part of the SR process is excluding these irrelevant citations through multiple steps, explained in PRISMA.
However, there is another category besides relevant and irrelevant results, that is typically retrieved – these citations are related to aspects of your topic you did not consider when asking your clinical question and devising your search strategies. These “by-products” might appear important enough that they should be included in your review, but this will be deterring from your original question.
Example: Your SR is on cancer patients’ attitude to health. You devise a comprehensive search strategy and include relevant search terms. As you begin screening the retrieved citations you realize that many of the articles actually focus on health education as it relates to attitudes. You may want to simply add health education as an additional aspect of your SR since it appears to be a valuable aspect of cancer patients’ health attitudes.
The issue with this approach is that unless you backtrack and revise your clinical question and search strategies (and thus essentially starting over from the beginning), your results and conclusions would deviate from the actual question that was proposed initially. If health education was not addressed in your original clinical question and reflected in your search strategies, it would be improper to include it in the final SR as there is likely an entire body of literature that was missed and thus any systematic conclusions could not be made regarding it.
Instead, these “by-product” citations (health education articles that came up in search results for health attitudes) should be treated as irrelevant to the systematic review you are conducting. A potential solution could be mentioning in the discussion section that from this review it was discovered that education is strongly tied to cancer patients’ attitudes toward their disease and their health and that it would be worthwhile to conduct a future review looking at how education can impact these views.
Takeaway: Try not to include “by-product” topics in your final review and analysis.
