Close
Close
This guide © 2024 by UCL - Library Skills is licensed under CC BY-NC-SA 4.0 
For systematic, comprehensive and / or complex search strategies, we recommend you check your strategy against standard criteria. Originally published in 2008, by CADTH (Canadian Agency for Drugs and Technologies in Health, now CAD - Canada's Drug Agency) the Peer Review of Electronic Search Strategies (PRESS) was designed to help librarians or information professionals check search strategies for systematic reviews. However, the principles can be used to help anyone with a knowledge of literature searching techniques as a way to check your strategies or as a list of things which you should have considered, including in disciplines other than health.
Use the abridged checklist below, or view the full checklist on the CAD website.
Please refer to relevant pages in the Library Skills guides, linked in context below, for information about the points mentioned in the checklist if you are unsure what they mean or what they refer to.
Don't introduce extra concepts into the search strategy which are not indicated by your research question.
If you have adopted a search method which is not usually recommended (e.g. restricting to the title field or using a NOT operator) have you justified your reasoning?
There are various tools available to help you formulate a research question and identify search concepts, e.g. PICO. Make sure that your concepts are not ambiguous or unclear - this is important for you as well as for anyone reading your work.
You don't always need to search for every concept you identify from your research question. If using PICO, for example, it often doesn't help to search for 'outcomes' and you might not have a comparison.
Avoid searching for something too specific, when the sought information might be found in articles which cover the topic more broadly. Conversely, do not mistake a broader term for a synonym.
There is no such thing as the ‘right’ number of results. However if there are a high number of results and most of them are not relevant, the search strategy needs to be more focused. If there are a very small number of records and they are almost all relevant, it’s likely that your strategy is not comprehensive enough.
Make sure you understand the difference between AND and OR. Do you know whether the database you are using needs to have Booleans entered in upper case?
Try to avoid nesting unless necessary. If you do nest terms, do so in the simplest way possible and make use of multi-line searching instead where possible.
We recommend avoiding the use of NOT.
Use this technique as an alternative to phrase searching to find more results, if applicable.
Subject headings are recommended for a systematic review search where available. Check that the subject headings you have applied are appropriate by consulting the scope note in the database thesaurus. Also, look at the headings applied to highly relevant articles which you have previously identified.
Check the scope note for the headings you have chosen and check the headings which are broader and narrower terms, to ensure your selected term is sufficiently focused.
Check that you understand how exploding works and which narrower terms will be selected if you have exploded a heading.
We recommend that you do not use the ‘focus’ option / restrict to ‘major headings’.
Not every search or database will include a search for subject headings, but if they are included, free text search terms must still be used as well.
Search for both UK and US spelling where necessary e.g foetus / fetus. Also think about hyphenated versions of words or phrases, e.g.: “covid 19” / covid19 / covid-19 and possessive forms e.g. Alzheimer’s / Alzheimer / Alzheimers
For example, if your search question was ‘does chocolate make you happy?’ you would not restrict your search to only words about happiness, as this would not pick up any articles talking about how chocolate makes you sad.
Truncation placed at the end of a word inappropriately reveals a lack of understanding, e.g. biology*. Be aware that many databases do not support truncating the beginning of a word e.g. *phage. In many databases a minimum number of characters is required before the truncation symbol.
Truncation placed too early can retrieve irrelevant results; e.g. transp* would find transplant, transplantation, transplants but also transpiration, transparent. Sometimes truncating earlier can get more relevant results, e.g. adolescen* rather than adolescent*.
Beware of acronyms which can have multiple meanings. AIDS can also be a word, e.g. hearing aids, walking aids.
For example, if you are looking for information about postnatal depression, a search for depression will retrieve results about other types of depression too, e.g. major depression.
The more keywords you combine together with OR, the more results you will get. Using many synonyms is fine as long as they are all relevant and focused on the concept being sought. If you do not use sufficient synonyms for each of your concepts, this reduces the chances of finding all relevant results. Be aware also of redundancy. For example, if you search for diabetes, you don’t also need to search for diabetes mellitus because the word diabetes by itself will pick up records which include the phrase diabetes mellitus anyway.
A stop word is a word that is ignored in a search string because it is assumed to be irrelevant to the search, e.g. an, is, the. A command word is a specific term that is part of the search syntax used to define a search query. e.g. AND, TITLE. Be careful about using stop words or command words as search terms. For example, ‘substance use’ (use is a command word in Ovid databases), ‘root canal’ (root is a command word in Ovid databases) ‘foot and mouth disease’ (AND is an operator). If you need to use stop or command words consult the database help for how to do this.
Make sure you are aware of the fields which have been searched. If you have not selected a search field and used the default field, check on the database help pages what that will include. Often an ‘all fields’ search will include unwanted fields. For textword searches you should include at least the title and abstract fields. In most databases the default field selection will include these.
Keep search strings easy to read. Do not use more than one type of operator in the same search string. See the Box on this page - 'long search strings'.
If you enter each search term on its own line this will help to spot errors, i.e. if there are 0 results.
Be very careful about the use of syntax from one database to the next.
Check your combinations thoroughly as mistakes can easily slip in especially with complex searches. Check all search lines are accounted for.
Use of any limits or filters will restrict results and must be justified. For example, if you restrict to RCTs this should be used only if you are sure that no other study design would be useful or relevant.
Note that in some databases limits or filters will remove recent, non-indexed content. Check whether this applies to the specific limit and database you are using (it will apply to most limits in, for example, Medline and Cinahl).
Potentially, helpful limits might be language or publication date range depending on the aims of the search.
A search filter (or ‘hedge’) taken from another source must be cited just as you would for any other source you have used in your work.
Keep search strings easy to read. Do not use more than one type of operator in the same search string.
For example:
(((computer OR laptop OR PC) ADJ screen) OR ((UV OR blue) ADJ light) AND (“dry eye” OR blepharitis))
would be better presented as:
