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 databases 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 databases. 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 (#4).

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.

Practice Suggestions

1. Line #3 of the Focused Seach 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

A practical overview with examples of Ovid’s searching features and functionality, usable across databases on the Ovid platform

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. Drawing upon six examples of common questions that are often put to Ovid Customer Support, this Focused Search looks at how operators, wildcards and commands can be used to effect across databases on the Ovid platform. Fuller details are provided in the Ovid Online Help as well as the Database Field Guide for the individual database you are searching in.

Example 1:
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.

Example 2:
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 as the example demonstrates.

Example 3:
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.

Example 4:
When used in Ovid, the FREQ or frequency operator, 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. PHQ-9, PHQ-15 and GAD-7 refer to established forms of reporting patient outcomes (PROs). Used at times individually, they are also deployed in combination. The searches as depicted in Example 4 yield records indicating that the mixed use of PROs is a dominant theme in the retrieved records as opposed to being simply mentioned.

Example 5:
Results from a search in a single database can be deduplicated. The Ovid command for this is ..dedup n where ‘n’ is the number of the search line whose records are to be deduplicated, e.g. line 8. In the search history, however, the completed function is displayed as remove duplicates from 8. Ovid can also deduplicate records obtained from searches across multiple databases.

Example 6:
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 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.

The 12 search expressions listed in this Focused Search also run on other bibliographic databases, for example Embase on Ovid and APA PsycInfo. 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. [Ovid Trainer: Example 6]. 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.

Strategy

acupuncture therapy.tw. [Example 1]

(acupuncture adj therapy).tw.

(acupuncture adj2 therapy).tw.

3 not 2 [Example 2]

randomized controlled trial.mp. [Example 3]

randomi#ed controlled trial.mp.

randomi#ed controlled trial$.mp.

PHQ-9.ab. /freq=3 [Example 4]

PHQ-9.ab. /freq=3 and PHQ-15.ab. /freq=3

PHQ-9.ab. /freq=3 and GAD-7.ab. /freq=3

remove duplicates from 8 [Example 5]

"substance use".mp. [Example 6]

Example C-01-005 - Reserved

Background

Example C-01-005 - Reserved

Strategy

(adolescents or young adults or teenagers).ti,ab. [Example 1: Variations added and qualified]
(adhd or Attention Deficit Hyperactivity Disorder).ti,ab.
(ritalin or concerta or daytrana or equasym or methylphenidate).ti,ab.
(adverse effects or adverse events or untoward effects or side 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: What is the question?]
(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

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

Example C-01-007 - Fine Tuning Searches (Limits) – All Databases UNDER CONSTRUCTION

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 "systematic review"

limit 6 to full text

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. Broader terms are those for which your chosen term is a narrower term or sub-concept. Narrower terms on the other hand are the narrower terms or sub-concepts of your chosen term.

We have already seen how, on Ovid, you can use the Advanced Search search mode to locate a term in a specific field (i.e. migraine.ti., see line #1 in this Focused Search) or in a group of fields (i.e. migraine.mp., see line #2). In the Advanced Search, with mapping turned on (which for most users is a default setting) you can search for a thing not just in terms of the characters that describe it i.e. migraine, but also in a conceptual sense. That is you are able to automatically include additional terms and phrases that more fully represent the concept that best describes the thing you are looking for. In this Focused Search the entry term for all 10 search lines, using the Advanced Search with mapping turned on, is ‘migraine’. With mapping turned on, our search journey starts and continues through the MeSH thesaurus where at various stages we are offered opportunities to expand upon or narrow the scope of our search (see lines #3 to #10).

To investigate this further, let’s consider the following research question and subsequent task.

Is acupuncture effective in the treatment of migraines? Find randomised controlled trials.

From the research question, let us describe the population group as ‘migraine sufferers’ and that by proceeding with the search term ‘migraine’ we are likely to capture a relatively high proportion of relevant documents.

Upon entering the term ‘migraine’ into the Advanced Search search mode with mapping turned on, Ovid maps our entered term onto a MeSH term, in this case Migraine Disorders/ (see line #3).  Clicking on the MeSH terms displays it in the context of its occurrence and positioning within the controlled vocabulary  as displayed below. Here we can see that Migraine Disorders is a sub-term of the broader term, Headache Disorders, Primary. But Migraine Disorders is itself the broader term for its narrower terms, namely Alice in Wonderland Syndrome, Migraine with Aura, Migraine without Aura and Ophthalmoplegic Migraine.

The returned display, which highlights the mapped to term Migraine Disorders, also tells us that a specific number of documents, indicated by the search results have been indexed with that specific term. In other words, the concept or idea represented by the term ‘migraine’ within a constellation of scope defining broader and narrower terms, has been identified in the database. This, as opposed to ‘just’ occurrences of the word ‘migraine’ as is the case in lines #1 and #2.

Explode and Focus Options:
The MeSH thesaurus, both in terms of its structure and direct integration in Ovid MEDLINE, also facilitates systematic searching. The scope of a selected MeSH term can be expanded upon by selecting the Explode option next to the selected term. See the green highlighting in the screen shot above. MeSH terms displayed in a search strategy with the ‘exp’, as shown in line #5 of the displayed Focused Search, tells us that the total number of records returned for that search line includes the MeSH term along with its narrower terms. By contrast, selecting the Focus option, highlighted in purple above, reduces the scope of the term, whether exploded or not, to records where the selected MeSH term represents one of the main topics as described in the article. The ‘*’ character placed before the MeSH term, as displayed in line #4 , indicates that the search has been run with this more specific requirement in mind. 

MeSH Subheadings:
Each MeSH term has a collection of subheadings associated with it. The list of subheadings for the MeSH term Migraine Disorders is displayed in the screen shot below.

As part of MeSH thesaurus construction and maintenance, the indexers at NLM allocate subheadings to MeSH terms. Subheadings provide an additional means of further fine-tuning the context within which the MeSH term, i.e. the concept, is being sought.  In the displayed Focused Search, lines #6, #7 and #8 return records that cover Migraine Disorders but specifically those related to the etiology, diagnosis and prevention and control of the condition respectively. It is also possible to select multiple subheadings as shown with respect to therapy, in line #9. 

Practice Suggestions:

1. Does the subheading /dh – Diet Therapy include vitamin or mineral supplements? Use the ‘i’ icon next to the subheading to find out.   
2. How many PubMed-not-MEDLINE records would you expect to find in line #5 of the displayed Focused Search? Look at the Focused Search Database Size and Shape (docz.dz) – Ovid MEDLINE to help you find the answer.
3. Describe what Ovid id doing in line #10 of the displayed Focused Search. In terms of status, what type or types of database record are being returned with this search?

Strategy

migraine.ti.

migraine.mp.

Migraine Disorders/

*Migraine Disorders/

exp Migraine Disorders/             

exp Migraine Disorders/et [Etiology]

exp Migraine Disorders/di [Diagnosis]

exp Migraine Disorders/pc [Prevention & Control]

exp Migraine Disorders/dh, dt, th [Diet Therapy, Drug Therapy, Therapy]           

exp Migraine Disorders/ or migraine.mp.