Topic: Systematic Searching for Evidence Synthesis

Embase is a biomedical and pharmacological database, similar to PubMed (the source of the MEDLINE data set), but with a greater international scope. Embase is particularly good for drug and medical device research and pharmacovigilance, and is considered an important database in systematic searching, as a complement to MEDLINE.

There is some overlap among journals indexed by MEDLINE and Embase. In cases where an article is indexed in both, Embase will only show Emtree indexing. For MEDLINE-only references, MeSH terms are displayed, but are mapped to Emtree, so they won't be entirely consistent with the MeSH terms in PubMed. Most of Embase's journals are also included in Scopus, but Scopus does not include the extremely granular Emtree indexing that is one of Embase's most significant features. Like Scopus, Embase is also a good source of conference abstracts.

Embase, like most subject-based research databases, uses a unique controlled vocabulary, or standardized language, to label its references. In Embase, this language is called Emtree. Emtree was developed based on the MeSH, Medical Subject Headings, used in PubMed/MEDLINE, but has a finer level of detail. More Emtree terms are added to records on average, and the granularity of terms can be helpful for systematic searching. Among the areas in which Emtree excels is detailed indexing of drugs, including both generic and trade names.

You can access Emtree in the navigation at the top of Embase. When searching Emtree directly, Embase makes suggestions of the terms you may be looking for, along with potential synonyms.

Once you select a term, Emtree shows you its location in the hierarchy or "tree," including any narrower terms. By default, searches you build in Emtree include all the narrower terms in the thesaurus when you execute your search; this is called exploding, and is indicated by /exp appearing after the term here and later in your search history.

If you don't wish to explode on a term, because the narrower terms are not relevant to your search, select the down arrow next to /exp and choose Index term /de.

In a set of search results, you can view Emtree terms by tapping Index Terms under a reference. In individual records, Emtree terms display after the abstract, in the Terms sections. This may include Disease Terms, Drug Terms, Device Terms, or Other Terms. These should be differentiated from the Author keywords, which are not standardized.

Embase offers several places in which to search. The Quick search page offers a default Broad search, which maps your search to an Emtree term (similar to PubMed's Automatic Term Mapping), explodes on that term to include all narrower terms in the hierarchy, and also searches the term as a keyword (Embase calls keywords "free text"). You can use the + Add field button to add additional concepts to your search.

The PICO search in Embase displays the various elements of PICO, to help you conceptualize your search more easily, but in fact you can use any line for any concept. As with Quick search, PICO search assists you in determining appropriate Emtree terms and possible synonyms.

Truncation

Truncation is searching for any ending of a word. In Embase, use an * (asterisk) to replace one or more characters at the end of a string of at least 2 letters. For example, mindful* finds mindful or mindfulness.

Wildcards

Wildcards can replace one or more characters. Embase offers three different wildcard options.

* (asterisk): In addition to truncation, the * can be used within a word to find multiple characters. For example, sul*ur finds sulfur or sulphur.

? (question mark): The ? replaces only one unknown character in a word. For example, ne?t finds neat, nest, or next. You can use the ? at the end of a string of characters to represent a single character, so catheter? finds catheters, but not catheter. The ? cannot be used if you're doing field searching, such as specifying your search take place in the title of a reference.

$ (dollar sign): The $ replaces 0 or 1 letter, so it can be used for words where an alternate spelling may have one extra character. For example, catheter$ finds catheter or catheters but not catheterization.

Phrase Searching

Embase requires you to use single quotation marks around phrases to ensure the words are searched next to each other in that exact order (Embase will automatically correct double to single quotation marks as needed). For example, if you're interested in low back pain, you need to search for 'low back pain' if you want that to be the exact phrase found in your results. You can use truncation on any or all words inside the quotation marks, so you can also try "low* back pain" to ensure you get articles about lower back pain.

Proximity Searching

In proximity or adjacency searching, you can tell the database to look for words near each other in a variety of ways. It's essential when you do this type of searching that you use parentheses to group terms accurately!

Near, which uses the all capitals NEAR, is one proximity operator. Near finds words if they are a maximum of a certain number of words apart from one another, regardless of the order in which they appear. You decide the number. For example, teaching NEAR/3 strateg* will find results that have a maximum of three words between the beginning and ending terms. You'll find results that mention teaching strategy, teaching strategies, or strategies often used for teaching. Note that you can use truncation and other syntax when employing proximity searching. That means you could also try something like mindfulness NEAR/5 'hand surgery' or even, using parentheses to properly group your keywords, (mindful* NEAR/5 ('hand surger*' OR operating).

Next, which uses the all capitals NEXT, finds words if they are within a certain number of words of one another, in the order in which you entered them. Otherwise it works the same as NEAR, so needle NEXT/3 program* will find needle sharing programs but not programs such as needle exchange.

A systematic search in Embase should include appropriate Emtree terms as well as keywords, in the same way that a systematic PubMed search would have both MeSH and keywords. In PubMed, [tiab] can be used as a search tag to ensure keywords are searched in title, abstract, and author keywords. In Emtree, the equivalent syntax is (search terms):ti,ab,kw

To construct your search, begin by building an Emtree search for your first concept. You see further guidance for this process in the earlier box in this page that explains how controlled vocabulary works in Embase.

Embase also offers a selection of synonyms for each Emtree term. You can copy the synonyms before you tap the Show results button; if you forget, just recreate your Emtree search to obtain them.

Return to the Embase search page and paste the synonyms in the box. Using the pencil, select the fields you wish to search. These synonyms are a good start, but you may find you wish to remove some of them, or add keywords not included there.

Your search history appears at the top of the search results page. Use OR to combine the results of your Emtree and keyword searches into one set.

Follow the same steps for each of your search concepts. Then, in your search history, use AND to combine them.

To export more than 500 citations, you'll need an Elsevier account. First, tap the blue Sign in button at the top right of the page. If you already have a login for the Scopus database, that's the same account, so just login with your Scopus credentials. If not, enter your email address and follow the instructions to create an account.

To export all results from an Embase search, up to 60,000 citations, tap the drop-down menu labeled "Select number of items" at the top left of your search results list. Select the range of references you want to download. Once you've indicated the number of references to download, tap the Export link just above and to the right of the menu.

Choose your preferred format, such as RIS (for Zotero) or RefWorks Direct Export, then tap Export.

A new tab will open in your browser. Tap the Download button when your RIS export is ready, then import the file into Zotero. If you're using RefWorks Direct Export, tap Submit to RefWorks.

To save your Embase search history, select all lines using the box to the left of the word History (or choose individual lines if you prefer). Then tap Export.

In the Export box, selecting CSV will yield a file that you can open in Excel. The "Check to expand combined queries" box displays any searches that combine numbered sets (e.g. #1 AND #2) in full textual form.

Download the file.

Below is an example of what a search, with all combined queries expanded, would look like in Excel. The columns have been resized so all text is visible.