Focused Searches

How to use the commonly occurring database fields, Title (TI) and Abstract (AB), to search across different databases using words as well as phrases.

Background

Ovid MEDLINE, Embase on Ovid and more specialist resources like APA PsycInfo, are often described collectively as bibliographic databases. Or, even more specifically, as abstracting and indexing (A&I) databases. This description refers to a common feature they all share, namely that the record in the database is not the full-text equivalent of the original publication but a reflection of it. A reflection composed of specific units of information, called database fields whose own content or values have been gleaned from the original document and mapped to the database record. This process of converting the information from an original publication into a database record is carried out by indexers. The highly trained and qualified indexers at the National Library of Medicine (NLM) carry out this task for PubMed.

Generally speaking, original documents up for inclusion in the database, can be represented by their original title. The majority of these original documents, particularly those from scientific journals, will also have an abstract associated with them. These two pieces of information, taken over by the indexers ‘as is’, constitute the informational foundation of any database record. As such, rendered as the database fields, Title (TI) and Abstract (AB), either alone or in combination, they provide a sure and reliable way of retrieving information from any bibliographic database. Some of which, e.g.  Ovid MEDLINE and Embase on Ovid, are composed of tens of millions of records.

The recommended search mode for a Title (TI) and/or Abstract (AB) based search, either alone or in combination, is Ovid’s Advanced Search. Whereas in the previous Focused Search, the search terms adolescents adhd ritalin adverse effects were entered into the Basic Search search mode altogether in one line, the Advanced search can take and search for each term individually (see the search example, [Version 1: Terms as given], lines #1 to #4). The individual search lines can then be combined to achieve a result set that contains all of them (#5).

In this Focused Search the first iteration i.e. [Version 1: Terms as given] operates on the terms exactly as stated. So, for instance, the search engine only finds the term adolescent in its plural form. The singular version is not captured. Similarly the medication ritalin is only captured via its representation in the form of its brand name. The generic name of the medication, namely ‘methylphenidate’ is not captured. In such cases the term expressions can be enhanced using truncation and the addition of synonyms using the OR operator. See the search example, [Version 2: Terms enhanced], lines #6 to #10. For truncation Ovid employs the characters ‘*’ or ‘$’. Also known as ‘wildcards’, these can be used to search for ending variations of a word. See the Focused Search Targeted Searching using Term Variants (Wildcards) – All Databases for more detailed coverage of wildcards.

Practice Suggestions

1. Line #3 of the Focused Search retrieves a number of results for ritalin.ti,ab. As an alternative, run a truncated version of the search expression, i.e.  ritalin*.ti,ab. Drug trade names do not usually have a plural form, so what were you expecting? What did you find?
2. In Ovid it is possible to set the number of letters (n) that should follow the truncated search term prefix. Run the search expression  ritalin*1.ti,ab. To find ‘ritalin’ plus ‘ritalin plus 1’ additional character. Are the additional results relevant? 
3. Although the NOT operator can only be used via the command line, it is a very useful means of separating out sets of results. Use the NOT operator to isolate the additional results obtained from the previous search. What do the isolated results have in common?

Strategy

adolescents.ti,ab. [Version 1: Terms as given]

adhd.ti,ab.

ritalin.ti,ab.

adverse effects.ti,ab.

1 and 2 and 3 and 4

(adolescent$ or young adult$).ti,ab. [Version 2: Terms enhanced]

(adhd or Attention Deficit Hyperactivity Disorder).ti,ab.

(ritalin or methylphenidate$).ti,ab.

(adverse effect$ or adverse event$ or untoward effect$).ti,ab.

6 and 7 and 8 and 9

How to use the operators OR, AND, NOT, ADJ, ADJn and FREQ to include term variants and build search queries systematically.

Background

Overview

The Ovid platform provides a comprehensive selection of features and functions to help build robust search strategies that are able to cope with the complexities of both language and document composition. First and foremost among these features and functions are the so-called ‘operators’. Essentially, these are commands that instruct the search engine to combine search terms in a certain way. These commands, used individually but more usually in their combination, will enable you to carry out targeted searching. The operators available on Ovid are OR, AND, NOT, ADJ, ADJn and FREQ. An overview is given in the graphic below. Fuller details are provided in the Ovid Online Help as well as the Database Field Guide for the individual database you are searching in.

A note on the NOT Operator (See line #7):
When constructing searches using thesaurus terms in combination with free text terms, caution is advised with the NOT operator, as using it may run the risk of unwittingly excluding relevant documents. However, this operator is very effective for isolating document sets. 

A note on the ADJ and ADJn Operators (See line #4):
There are two variations of the Ovid proximity operator, ADJ and ADJn. The first, ADJ, carries out a regular phrase search finding the terms together in the order in which they have been entered. By contrast, the second variant, ADJn where ‘n’ is a number, carries out a search for the adjacent terms separated by n-1 words apart, the terms occurring in any order.

A note on the FREQ Operator (See line #12):
When used in Ovid, the FREQ or frequency operator shown in [Example 3], takes the form ‘x.ab./FREQ=n’ where ‘x’ represents a search term, ‘.ab’ the field in which the search should be conducted, here the Abstract  (AB) field, ‘freq’, the Ovid command for this function and ‘n’ the number of occurrences of term x being sought after. This operator can be used to great effect in full-text databases, such as the Cochrane Database of Systematic Reviews or YourJournals@Ovid. It can also be used meaningfully in text-rich fields such as Abstract (AB). In this Focused Search example, the terms being search for, i.e. truncated forms of the terms ‘migraine’ and ‘acupuncture therapy’ will return documents where these two ideas, occur together, at least 3 times each in the abstract field. The implication being that the returned documents are on point, are useful in their own right but can also be examined further to see how their subject headings and keywords treat the subject matter.

The 12 search expressions listed in this Focused Search also run on other bibliographic databases, for example, Embase on Ovid. Use the Change option on the Ovid platform to bring up the All Resources menu and select the database or the databases in combination, across which you want to search. 

Lastly, note too, the user created annotations, e.g. [Example 1]. These can be added to search lines when placed between block brackets [ ]. The annotations can also be added to, amended, or deleted using the Edit option available via the search screen under Actions / More / Edit.

Practice Suggestions

1. Using the ADJ operator renders line #4 more succinct. How would further increasing the number ‘n’ impact your results? How can you test that the ADJn-based search expression is indeed delivering the search term words in the phrase combinations you expect?
2. In [Example 2], in specifically search line #10, the US spelling of ‘randomized’ as well as the British spelling, ‘randomised’ are used. How could you use the ADJ operator to reduce the number of terms in the search expression, yet yield the same number of results for the search?
3. The search in line #12 yields only three results. Name three ways with which you could derive additional benefits from using this search expression. 

Strategy

(adolescents or young adults or teenagers).ti,ab. [Example 1]

(adhd or Attention Deficit Hyperactivity Disorder).ti,ab.

(ritalin or concerta or daytrana or equasym or methylphenidate).ti,ab.

(adverse events or ((adverse or untoward or side) adj3 effects)).ti,ab.

1 and 2 and 3 and 4

(infant or infants or child or children or preschool or pre school).ti,ab.

5 not 6

(migraine or migraines or sick headache or sick headaches or status migrainosus).ti,ab,kf. [Example 2]

(acupuncture therapy or acupuncture treatment or acupuncture analgesia or ear acupuncture or electroacupuncture or meridians or acupuncture points or moxibustion).ti,ab,kf.

(randomized controlled trial or randomized controlled trials or randomised controlled trial or randomised controlled trials).ti,ab,kf.

8 and 9 and 10

migraine$.ab. /freq=3 and acupuncture therap$.ab. /freq=3 [Example 3]      

How to use Ovid’s character replacement options, so-called ‘wildcards’, to include singular and plural versions of search terms as well as other possible variations.

Background

The Ovid platform provides a comprehensive selection of features and functions to help build robust search strategies that are able to cope with the complexities of both language and document composition. For instance, spelling variations are a challenge when formulating search expressions. Not only are there regional spelling differences, e.g. British versus American spelling, but also very commonly, singular and plural forms of words.

Ovid provides a comprehensive selection of so-called ‘wildcards’ to accommodate these differences. Generally speaking, wildcards are symbols that can be used as character replacement holders. Some wildcard characters can be entered with numbers to indicate the number of characters for which a place or places should be held. An overview is given in the graphic below.

Fuller details are provided in the Ovid Online Help as well as the Database Field Guide for the individual database you are searching in.

A note on searching for text strings “n” (See line #12):
Although Ovid performs phrase searching by default, quotation marks “ ” , can be used to require Ovid to search for the expression as a literal string. This is particularly helpful when the string includes special characters, such as a forward slash (/) or reserved words, such as the word ‘use’. On the Ovid platform the word ‘use’ is a reserved word in that it is interpreted as a command. In combination with a database’s abbreviation, e.g. medall for Ovid MEDLINE ALL, Ovid understands this expression as a command, i.e. use medall to change from the resource you are in, to the database, Ovid MEDLINE ALL. This being the case, the search substance use.ti,ab. would need to be entered as ”substance use”.ti,ab.

Practice Suggestions

1. Drug trade names, like ‘Ritalin’ are usually clear cut terms and do not warrant truncation with wildcard characters. That said, would there be any benefit in truncating ritalin in line #3 even in a limited way, e.g. using the search expression  ritalin*1.ti,ab. for example. How would you establish the usefulness or even a possible need for this.
2. Drawing upon the example in line #6, run the searches child$.ti,ab. and child$3.ti,ab., the latter in a separate line. Use the NOT operator to establish the difference in the numbers of results between them. What terms have you obtained with child$.ti,ab. that were excluded by child$3.ti,ab.?
3. Previous test searches using the ADJ operator for line #10 established that it could be rendered efficiently as ((randomized or randomised) adj (controlled trial or controlled trials)).ti,ab,kf.  With specific wildcards we can now further reduce this to (randomi#ed controlled trial$).ti,ab,kf.. On the face of it, these two expressions should return the same number of results. Do they? Were there to be a difference, how could you explain it?

Strategy

(adolescent$ or young adult$ or teenager$).ti,ab. [Example 1]

(adhd or Attention Deficit Hyperactivity Disorder).ti,ab.

(ritalin or concerta or daytrana or equasym or methylphenidate).ti,ab.

(adverse event$ or ((adverse or untoward or side) adj3 effect$)).ti,ab.

1 and 2 and 3 and 4

(infant$ or child$3 or preschool or pre school).ti,ab.

5 not 6

(migraine$ or sick headache$ or status migrainosus).ti,ab,kf. [Example 2]

(acupuncture therap$ or acupuncture treatment$ or acupuncture analgesia or ear acupuncture or electroacupuncture or meridian$ or acupuncture point$ or moxibustion).ti,ab,kf.

(randomi#ed controlled trial$).ti,ab,kf.

8 and 9 and 10

("wom#n with migraine" or "m#n with migraine").ti,ab. [Example 3]

How to construct searches using either a user defined group of fields or the database specific ‘multi-purpose’ (.mp) set of fields as provided by Ovid.

Background

When using the Advanced Search mode in a bibliographic database like Ovid MEDLINE, it is common practice to enter a search term followed by the field or fields you want Ovid to search in. For example, the search expression (adolescent$ or young adult$).ti,ab,kf. (see this Focused Search, line #1) tells Ovid to search for the terms adolescent or adolescents or young adult or young adults in the Title (TI) field as well as the Abstract (AB) and Keyword Heading (KF) fields. This kind of search is referred to as a qualified search in the sense that the instruction to the database is explicit, namely “go and look for these specific terms in these specific fields, that I, the user have selected.”

But as noted in a previous Focused Search, databases vary not only in size but also in terms of their “shape”. You may, for instance, be uncertain about the field composition or layout of the database. In such cases Ovid can help out with the multi-purpose or .mp search. This is described as an unqualified search across a default set of fields which are specific to the database being used.

In this case, the example as used above would be entered as (adolescent$ or young adult$).mp. For Ovid MEDLINE the default set of fields across which Ovid will then search are:

title, book title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms, population supplementary concept word, anatomy supplementary concept word.

By contrast, for Embase the default set of fields across which Ovid will search are:

title, abstract, heading word, drug trade name, original title, device manufacturer, drug manufacturer, device trade name, keyword heading word, floating subheading word, candidate term word.

Because the default fields to be covered by an .mp search are pre-determined, .mp based searches lend themselves well to cross database searching. In this sense the .mp search option is a helpful supplement to Ovid’s Basic Search which is also able to carry search expressions easily across multiple databases.

Practice Suggestions:

1. Combine and compare this Focused Search with the Basic Search example, Knowledge Discovery with NLP (Basic Search) – All Databases. Note that in the Basic Search the search line adolescents adhd ritalin adverse effects {Including Related Terms} has as its search type Basic. But when you limit this line to Five Stars, the search type changes to Advanced. This now enables you to combine this result set with the results in this .mp example. Do the numbers differ greatly?
2. Use the Change option to run this Focused Search through Embase on Ovid. Do both versions of the search, i.e. the lines using .ti,ab,kf. as opposed to .mp, run properly? How could you easily compare the results obtained in Ovid MEDLINE with those from Embase on Ovid?
3. Use the Change option to run this Focused Search through APA PsycInfo. Do both versions of the search, i.e. the lines using .ti,ab,kf. as opposed to .mp, run properly? How can you tell? Identify two markers that tell you.

Strategy

(adolescent$ or young adult$).ti,ab,kf. [Version 1: User defined group of fields]

(adhd or Attention Deficit Hyperactivity Disorder).ti,ab,kf.

(ritalin or methylphenidate$).ti,ab,kf.

(adverse effect$ or adverse event$ or untoward effect$).ti,ab,kf.

1 and 2 and 3 and 4

(adolescent$ or young adult$).mp. [Version 2: Ovid .mp defined group of fields]

(adhd or Attention Deficit Hyperactivity Disorder).mp.

(ritalin or methylphenidate$).mp.

(adverse effect$ or adverse event$ or untoward effect$).mp.

6 and 7 and 8 and 9

How to fine tune individual search lines as well as whole search queries using user-selected criteria, called ‘limits’.

Background

When constructing a search question it is helpful to separate out the search terms and phrases from additional criteria that can be used to systematically reduce your database results down to a manageable number of references. Such additional criteria could be the type of publication; a journal article, a review, a specific type of review, e.g. Cochrane or ACP, a case report or even a letter or  editorial. Other additional criteria could include narrowing down the result set by language or by time. In the context of searching for information on databases, imposing such additional criteria upon a search is referred to as applying ‘limits’. Whilst the idea of employing limits is common to A&I databases, the variety and scope of the ‘limits’ options available to searchers can differ between databases.    

This Focused Search introduces an initial selection of limits available for Ovid MEDLINE. These are displayed in the image below.

The limits can be selected and applied sequentially (see lines #6,#7 and #8) or imposed in a single step (see line #9). The full range of limits available for the database you are searching in, can be viewed and selected by clicking on the Additional Limits option at the bottom of the initial selection display. You can also place the limits, most useful to you, on the initial selection display by selecting the Edit Limits option. Your selection will remain as your initial display selection for the duration of your database session. After which they will return to and re-display the default selection.

For database administrators, please note  however, that the default selection can be customized for you by Ovid Technical Support so that the default selection of limits better addresses your users’ needs. 

Practice Suggestions

1. From the results shown in line #5, how many of them are from preprint servers?
2. Line #6 removes from the total number of records, those that do not have abstracts. Use the NOT operator to determine how many of these records there are? Could they nevertheless still be relevant to the search query? How could you assess this?
3. Name four of the common structural elements seen in records with structured abstracts. Tip: Display the records involved using Ovid’s Complete Reference display option.

Strategy

(adolescent$ or young adult$).mp. [Version 2: Ovid .mp defined group of fields]

(adhd or Attention Deficit Hyperactivity Disorder).mp.

(ritalin or methylphenidate$).mp.

(adverse effect$ or adverse event$ or untoward effect$).mp.

1 and 2 and 3 and 4

limit 5 to abstracts

limit 6 to full text

limit 7 to yr=”2020 -Current”

limit 5 to (abstracts and full text and yr=”2020 -Current”)

An introduction to the system and structure of the controlled vocabulary, based upon medical subject headings, that lies behind the bibliographic database Ovid MEDLINE.

Background

Drawing again on the idea of the ‘size and shape’ of databases, this Focused Search provides an introduction to the structure and system of the MeSH thesaurus. The MeSH thesaurus is a controlled vocabulary, based upon medical subject headings. It has been created and is maintained by the National Library of Medicine (NLM) in the USA. The MeSH thesaurus is a domain specific knowledge base curated by humans, that aids both document classification as well as discoverability and retrieval.

MeSH Structure and System

Ovid MEDLINE is Ovid’s rendering of the PubMed database, it too created and maintained by NLM. The content of this kind of bibliographic database consists, for the most part, of records of documents that have been indexed using a controlled vocabulary. A controlled vocabulary is a hierarchical term classification system where terms, representing a thing or concept, can be located within the controlled vocabulary system, along with its broader and narrower terms. For an informative optical rendering of this enter the search term ‘migraine disorders’ into the MeSH database search box as shown below.

https://www.ncbi.nlm.nih.gov/mesh/?term=migraine+disorders

The returned screen provides detailed background information on the target term. Amid the detail is also an indication of where the target term, here ‘migraine disorders’ appears in the MeSH term hierarchy. In the enhanced diagram below the blue bordered terms represent the broader terms associated with ‘migraine disorders’. By contrast the thinner green bordered expressions are its narrower terms. Which is to say that the idea or concept represented by the MeSH term ‘migraine disorders’ can itself be further broken down into finer ideas or sub-concepts, specifically ‘Alice in Wonderland Syndrome’, ‘Migraine with Aura’, ‘Migraine without Aura’ and ‘Ophthalmoplegic Migraine’.

Understanding the MeSH term constellation for your target terms or terms, and being able to navigate through the thesaurus hierarchy effectively, will greatly assist you in the systematic construction of high quality search queries.

In PubMed the MeSH database is a stand-alone, albeit connected, knowledge resource. By contrast the MeSH thesaurus is fully integrated into the Ovid MEDLINE search experience. MeSH terms can be easily located in the thesaurus hierarchy and narrower terms selected (as well as deselected) either as a group or individually as you decide. When you choose to include all of the MeSH terms narrower terms, you are said to have “exploded” the term. The Explode function and a related function, called Focus will be covered in the next Focused Search, Knowledge Discovery with MeSH (Explode & Focus) – Ovid MEDLINE.

Practice Suggestions:

1. Go to the Advanced Search. Check that the option Map term to Subject Heading is activated, that is ticked, and enter as a Keyword the expression migraine disorders. Looking at the screen that is subsequently displayed, list at least 4 options you could act on from there.
2. Go again to the Advanced Search. Deselect the option Map term to Subject Heading and as before enter as a Keyword the expression migraine disorders. Looking at the screen that is subsequently displayed, what kind of search has Ovid executed? Where do you think Ovid has searched for your term?  
3. In the Focused Search used in this example, how can you verify that the results in line #2 are included in the .mp search, run in line #3?

Strategy

migraine disorders.ti.

migraine disorders.ti,ab.

migraine disorders.mp.

Migraine Disorders/

exp Migraine Disorders/             

exp Migraine Disorders/ or migraine disorders.mp.