Focused Searches

How to run a simple yet focused search to quickly identify records connecting a drug and a specific side effect.

Background

In busy and demanding healthcare environments, practitioners may often be called upon to obtain up to date information on a specific drug and its possible or even known side effects. The large A&I biomedical databases are able to provide such information. However, given the differences in the way they index the source material, the search strategy needs to be constructed with the database specific indexing methods in mind, in order to obtain the best results. 

For its part Embase uses extensive and deep indexing, providing targeted access to over 40 million records. The Embase indexing system is based upon the Emtree thesaurus where special focus is placed on the 3 D’s namely, drugs, devices and diseases.

With respect to deep indexing, Emtree’s understanding of “drug terms” is that of any chemical entity. That said, clinical drugs are indexed in greater depth than other drug terms. The indexing of clinical drugs can also be supplemented or modified using drug subheadings. In Embase the subheadings are specific to each of the 3 D’s; diseases, drugs and devices. One of the 5 key subheadings for drugs, is the subheading Adverse Drug Reaction and for diseases Side Effect. See the reference below for more details on how Embase indexes drugs and chemicals.

The search example demonstrates how in 3 lines, users can make good use of Emtree’s very focused and deep indexing, related to clinical drugs to retrieve specific and authoritative information. In the example Ibrance was submitted as the initial entry term. This, Embase maps to palbociclib, its generic name which in turn, as an Emtree term also covers synonyms and variants. Neutropenia, also submitted as an entry term, was mapped to neutropenia, which also being an Emtree term, similarly covers synonyms and variants.   

1. exp palbociclib/ae [Adverse Drug Reaction]
2. exp neutropenia/si [Side Effect]
3. 1 and 2

The lines of the search can always be supplemented. For example, line #1 could also include the key terms in the form of .mp searches. Similarly with line #2 in respect of the side effect condition.

For recall and teaching purposes the search could also be documented in the form of a MODEL SEARCH,

1. exp < DRUG NAME >/ae [Adverse Drug Reaction]
2. exp < ADVERSE EVENT >/si [Side Effect]
3. 1 and 2

The search as listed here is the Embase focused search for running a simple search to quickly identify records connecting a drug and a specific side effect. For the functionally equivalent search in Ovid MEDLINE, please see Adverse Drug Reactions: Quick Search – Ovid MEDLINE.

Strategy

1. exp palbociclib/ae [Adverse Drug Reaction]

2. exp neutropenia/si [Side Effect]

3. 1 and 2

Comparison of similar search strategies between dissimilar databases.

Background

Abstracting and Indexing (A&I) databases are the primary source of information for systematic reviews, be the retrieved information actual systematic reviews themselves, or their constituent units of analysis, e.g. randomized controlled trials, observational studies etc.

A systematic review, irrespective of type, benefits from reviewers being able to consult multiple databases. However, while such databases fit the general label of being Abstracting and Indexing (A&I) database resources, they vary in their scope and coverage, in their editorial policies and processes as well as in how the information may be retrieved from them.

This has the practical consequence that while a search can be designed to generally follow the same sequence of conceptual blocks, the exact formulation of the search terms and syntax needs to be adapted for each database.   

The search discussed here is an Ovid MEDLINE search that was constructed and run to support the production of a qualitative systematic review, the results of which were reported in the paper “Patients’ perspectives on interprofessional collaboration between health care professionals during hospitalization: a qualitative systematic review” and published in the journal JBI Evidence Synthesis (see references below).

The Ovid MEDLINE search was one of six searches, each being constructed separately for the six databases considered by the systematic review. An initial search was constructed and run in 2017 and then updated in 2018. The search considered here is the updated version from 2018.

Strategy

1. Patient Care Team/

2. exp Interprofessional Relations/

3. ((doctor$1 or physician$1) adj1 nurse$1 adj2 (collaboration or communication or cooperation or relation$1 or round$1)).ab,ti,kf.

4. ((interprofessional or inter-professional or interdisciplinary) adj2 (care or collaboration or communication or cooperation or healthcare or management or relation$1 or round$1 or team$1)).ab,ti,kf.

5. (team$1 adj1 (care or healthcare)).ab,ti,kf.

6. teamwork.ab,ti,kf.

7. (((collaboration or communication or cooperation or relation$1) adj2 (provider$ 1 or clinician$1)) not (patient$1 adj2 (provider$1 or clinician$1))).ab,ti,kf.

8. or/1-7

9. exp Patient Satisfaction/

10. Patient Participation/

11. ((patient$1 or inpatient$1 or client$1 or user$1 or wom#n$1) adj3 (participation or perspective$1 or view$1 or viewpoint$1 or perception$1 or satisfaction or experience$1 or attitude$1 or role$1 or preference$1 or expectation$1 or involve$1 or involvement or engagement or dissatisfaction$1)).ab,ti,kf.

12. or/9-11

13. (qualitative or interview_ or findings or focus group_ or themes or mixed method$1).mp. or exp qualitative research/

14. 8 and 12 and 13

15. limit 14 to yr=‘‘1980 -Current’’

How to incorporate a search filter from an external, third-party source into your own Ovid search libraries and documents (Part 1).

Background

One of the main drivers behind the OvidGO! / Focused Searches project is to make searches, and especially search filters, discoverable, actionable and useable. Search filters, searches designed to achieve a specific outcome, are particularly challenging. On the one hand they are spread across a large variety of publication vehicles, e.g. journal articles, systematic reviews, websites etc.. A consequence being that many of them are “locked” in analogue formats, most notably print. On the other hand, the role and value of search strategies, filters or otherwise, as the informational foundation of systematic reviews is understated. And as a result, problematic.

An analysis of critically assessed systematic reviews using the KSR Evidence database (also available on Ovid), demonstrates – dramatically – that a large number of currently available systematic reviews have to be marked as having a high risk of bias because they have not at all, or only partially, reported on the search strategies employed for their construction. Ovid’s search strategy construction and sharing functionality, used purposefully and systematically, can help remedy this situation.  

Taking as its reference the Adverse Events (AE) filter published by the European Medicines Agency (EMA), this Focused Search provides an example of what is meant by making search filters, discoverable, actionable and useable.

As part of its pharmacovigilance obligations and activities the EMA is mandated to keep a watching brief over some 300 substance groups and 100 herbal substance groups. Specifically, to monitor medical literature and enter relevant information into the EudraVigilance database. This the EMA does by running search strategies on EMBASE (Embase.com) that combine (a). a comprehensive list of terms describing the target drug or herbal substances, together with (b). an extensive list of terms and phrases that refer to adverse events in a variety of ways. The list of substances and the AE filter used is available from the EMA and is published and updated regularly. A link to the published document is given below.

The current version of the AE filter, displayed below, can be found on page 13.

AND ('adverse drug reaction'/exp OR adverse:de,ab,ti OR 'side effect':de,ab,ti OR 'side effects':de,ab,ti OR 'side effect':lnk OR 'drug overdose'/exp OR overdos*:de,ab,ti OR 'drug misuse'/exp OR misus*:de,ab,ti OR 'drug abuse'/exp OR 'substance abuse'/exp OR abus*:de,ab,ti OR 'pregnancy'/exp OR pregnan*:de,ab,ti OR 'pregnancy complications'/exp OR 'lactation'/exp OR 'breast feeding'/exp OR ('milk human'/exp AND 'secretion'/exp) OR 'fertility'/exp OR 'reproduction'/exp OR 'fetus'/exp OR 'embryo'/exp OR terat*:de,ab,ti OR 'drug efficacy'/exp OR 'drug withdrawal'/exp OR 'medication error'/exp OR 'death'/exp OR death*:de,ab,ti OR fatal*:de,ab,ti OR 'drug interactions'/exp OR 'drug interaction'/exp OR 'carcinogenicity'/exp OR carcinogen*:de,ab,ti OR mutagen*:de,ab,ti OR 'off label drug use'/exp OR 'occupational exposure'/exp OR 'toxicity'/exp OR toxic*:de,ab,ti OR pharmacotox*:de,ab,ti OR neurotox*:de,ab,ti OR cardiotox*:de,ab,ti OR nephrotox*:de,ab,ti OR immunotox*:de,ab,ti OR hepatotox*:de,ab,ti OR cytotox*:de,ab,ti OR immunocytotox*:de,ab,ti OR 'intoxication'/exp OR 'congenital disorder'/exp OR 'drug treatment failure'/exp OR 'case report'/exp OR 'case reports':de,ab,ti OR 'environmental exposure'/exp OR 'treatment contraindication'/exp OR 'drug contraindication'/exp OR 'injury'/exp OR suicide:de,ab,ti OR 'poisoning'/exp OR 'drug tolerance'/exp/mj OR 'treatment failure'/exp/mj OR 'drug resistance'/exp/mj OR 'substance-related disorders'/exp OR 'treatment failure'/exp OR 'drug resistance'/exp OR 'drug resistance':lnk) AND [humans]/lim

So that you can also use this search filter on Ovid, we have translated the Embase.com syntax into Embase on Ovid syntax. Availing of the ABC format options behind Focused Searches, you can display the Ovid rendered version by clicking on ACTIONS / Display. Alternatively, ACTIONS / Launch will run the filter in Embase on Ovid. If you would like to talk to a member of the Ovid Training Team about the filter, select ACTIONS / Ovid Coaching. Once in the database you can combine the filter with substances relevant to your investigation. You can also save the filter in My Workspace for re-use as part of your own or institutional library of searches and search blocks. 

The next Focused Search, Search Blocks in Line Format (EMA AE Filter) – Embase (Part 2), will consider the relative merits of this filter in its paragraph format vis-à-vis a line-by-line format.  

Strategy

exp adverse drug reaction/ OR adverse.hw,ab,ti. OR side effect.hw,ab,ti. OR side effects.hw,ab,ti. OR side effect.fs. OR exp drug overdose/ OR overdos*.hw,ab,ti. OR exp drug misuse/ OR misus*.hw,ab,ti. OR exp drug abuse/ OR exp substance abuse/ OR abus*.hw,ab,ti. OR exp pregnancy/ OR pregnan*.hw,ab,ti. OR exp pregnancy complications/ OR exp lactation/ OR exp breast feeding/ OR (exp milk human/ and exp secretion/) OR exp fertility/ OR exp reproduction/ OR exp fetus/ OR exp embryo/ OR terat*.hw,ab,ti. OR exp drug efficacy/ OR exp drug withdrawal/ OR exp medication error/ OR exp death/ OR death*.hw,ab,ti. OR fatal*.hw,ab,ti. OR exp drug interaction/ OR exp drug interaction/ OR exp carcinogenicity/ OR carcinogen*.hw,ab,ti. OR mutagen*.hw,ab,ti. OR exp 'off label drug use'/ OR exp occupational exposure/ OR exp toxicity/ OR toxic*.hw,ab,ti. OR pharmacotox*.hw,ab,ti. OR neurotox*.hw,ab,ti. OR cardiotox*.hw,ab,ti. OR nephrotox*.hw,ab,ti. OR immunotox*.hw,ab,ti. OR hepatotox*.hw,ab,ti. OR cytotox*.hw,ab,ti. OR immunocytotox*.hw,ab,ti. OR exp intoxication/ OR exp congenital disorder/ OR exp drug treatment failure/ OR exp case report/ OR case reports.hw,ab,ti. OR exp environmental exposure/ OR exp treatment contraindication/ OR exp drug contraindication/ OR exp injury/ OR suicide.hw,ab,ti. OR exp poisoning/ OR exp *drug tolerance/ OR exp *treatment failure/ OR exp *drug resistance/ OR exp substance-related disorders/ OR exp treatment failure/ OR exp drug resistance/ OR drug resistance.fs.

A look at the relative merits of search filters in a paragraph format vis-à-vis a line-by-line format, using as an example the EMA AE Filter (Part 2).

Background

Generally speaking, search filters, whether displayed on screen or reported in scientific literature come in three formats. These are, firstly the “line-by-line” format where search expressions are entered or displayed as a list, one on top of the other, literally line by line. The second format, here referred to as the “paragraph” format has all the search expressions bundled into a single line that appears as a block of text. This provides for a more compact display, and can in turn, be also run as a single line search. Thirdly, the “hybrid” format describes search strategies that contain a mixture of line-by-line and paragraph format search expressions.

Opting for one format over another depends upon many factors. These vary from available screen space to the capacity of the reporting vehicle, e.g. a printed journal, to simply personal choice.

The purpose of this Focused Search is to raise a note of caution when using the paragraph format without firstly verifying whether or not the search expressions (a). run in the target database, and (b). are relevant to the task at hand. The AE Filter as shown above, unpacked, extends to 60 lines listed line-by-line. This way the relative presence of the terms and phrases can be more easily assessed. As too can be the balance between free term and thesaurus-based searches, and with respect to the latter,  whether they have been exploded. Using Ovid’s Edit function (Actions / More / Edit) a thesaurus expression can have the “exp” command function deleted or added to see what a difference this makes.

It is recommended that your key search filters be saved and stored in both paragraph and line-by-line formats. The former for ease of execution and display. The latter for verification, updating and overall knowledge asset management.  

Ovid Trainer Tips

• Command line Syntax: Lists of search expressions can be neatly combined using the OR operator as displayed in line #61 above, i.e. or/1-60.
• Ovid Search History Launcher Tool: Having unpacked a paragraph formatted search using for example, Notepad, the resulting list of search expressions can be easily imported into Ovid using the Ovid Search History Launcher Tool. See https://tools.ovid.com/ovidtools/launcher.html .
• Share Search History: Conversely, searches regardless of format can be exported from Ovid into other applications for eventual re-use, for peer review and reporting purposes. Located in the Search History section of the Ovid interface, the three Share Search History options are, Email All Search History, Copy Search History Link and Copy Search History Details.

Strategy

exp adverse drug reaction/

adverse.hw,ab,ti.

side effect.hw,ab,ti.

side effects.hw,ab,ti.

side effect.fs.

exp drug overdose/

overdos*.hw,ab,ti.

exp drug misuse/

misus*.hw,ab,ti.

exp drug abuse/

exp substance abuse/

abus*.hw,ab,ti.

exp pregnancy/

pregnan*.hw,ab,ti.

exp pregnancy complications/

exp lactation/

exp breast feeding/

(exp milk human/ and exp secretion/)

exp fertility/

exp reproduction/

exp fetus/

exp embryo/

terat*.hw,ab,ti.

exp drug efficacy/

exp drug withdrawal/

exp medication error/

exp death/

death*.hw,ab,ti.

fatal*.hw,ab,ti.

exp drug interaction/

exp drug interaction/

exp carcinogenicity/

carcinogen*.hw,ab,ti.

mutagen*.hw,ab,ti.

exp 'off label drug use'/

exp occupational exposure/

exp toxicity/

toxic*.hw,ab,ti.

pharmacotox*.hw,ab,ti.

neurotox*.hw,ab,ti.

cardiotox*.hw,ab,ti.

nephrotox*.hw,ab,ti.

immunotox*.hw,ab,ti.

hepatotox*.hw,ab,ti.

cytotox*.hw,ab,ti.

immunocytotox*.hw,ab,ti.

exp intoxication/

exp congenital disorder/

exp drug treatment failure/

exp case report/

case reports.hw,ab,ti.

exp environmental exposure/

exp treatment contraindication/

exp drug contraindication/

exp injury/

suicide.hw,ab,ti.

exp poisoning/

exp *drug tolerance/

exp *treatment failure/

exp *drug resistance/

exp substance-related disorders/

exp treatment failure/

exp drug resistance/

drug resistance.fs.

or/1-64

How run a search history to locate references to information retrieval topics and set it up as an AutoAlert.

Background

Sponsored by the European Commission, the European Network for Health Technology Assessment (EUnetHTA) was set up to bring together the Health Technology Assessment (HTA) Agencies across the EU, to exchange experience and develop practice as well as methodological guidelines.

One of the project outcomes has been the publication of a methodological guideline entitled, Process of information retrieval for systematic reviews and health technology assessments on clinical effectiveness. Comprehensive, systematic and helpfully readable, this publication is highly recommended to information specialists and medical librarians working on systematic review. The authoring team comes from IQWIG in Germany, NIPH in Norway and AETSA in Spain.

This Focused Search draws upon a search strategy, described in Annex 2 of the EUnetHTA methodological guideline, which retrieves references that discuss information retrieval topics. These can range from accounts of experience gained by varying user groups using search techniques to reports of interactions with specific resources, for example PubMed and Google Scholar.  

Of use to librarians and information specialists keen to monitor developments in the field of information retrieval, the search can be stored in an Ovid My Workspace account and be run manually at intervals determined by the user. Alternatively, the search strategy can be set up as an AutoAlert and run daily, weekly, monthly, or following a database update, as you please. If you would like some assistance from an Ovid Trainer on how to set up this search strategy as an AutoAlert, please request an Ovid Coaching Session.

Strategy

(Semi-automated or data mining or (Capture adj3 recapture) or machine learning).ab,ti.

(search adj3 (strateg* or term* or filter*)).ab,ti.

(query or queries).ab,ti.

search*.ti.

or/1-4

exp "Information Storage and Retrieval"/

"Medical Subject Headings"/

"Abstracting and Indexing as Topic"/

Documentation/

reporting.ab,ti.

(bibliographic databas* or Pubmed).ab,ti.

(MeSH or controlled vocabulary or indexing).ab,ti.

or/6-12

"Review Literature as Topic"/

exp "Evidence-Based Practice"/

"Technology Assessment, Biomedical"/

or/14-16

and/5,13,17

exp *"Information Storage and Retrieval"/

4 and (16 or 19)

(Medline or PubMed).ti.

(Clinical Queries or Haynes or hedge or search).ti,ab.

and/21-22

or/18,20,23

How to check for thesaurus term compatibility. An example using the entry term adverse drug event on Ovid MEDLINE and then on Embase on Ovid.

Background

When running search strategies in Embase on Ovid, that were originally designed in Ovid MEDLINE, care needs to be taken to ensure they are delivering the required information as expected. Although there is clear kinship between the two thesauri, MeSH and Emtree, they are nevertheless different in terms of their structure, editorial processing and subsequent operation.

This Focused Search demonstrates how the phrase adverse drug event submitted as an entry term into the Advance Search on Ovid MEDLINE is mapped onto the MeSH term "Drug-Related Side Effects and Adverse Reactions". As adverse drug event is a “used for” term for the MeSH term "Drug-Related Side Effects and Adverse Reactions", entering adverse drug event/ , i.e. with a backslash which forces Ovid to perform a thesaurus search, this search expression yields the same result. See lines #1 and #2 in the screen shot below.

Ovid MEDLINE < 1946 to Present >

However, when we re-run the same search expressions in Embase on Ovid, they execute seemingly without a problem and yield results. The MeSH term "Drug-Related Side Effects and Adverse Reactions" is “understood” by Embase. But, from a returned results point of view, not in the same way as entering the phrase adverse drug event when in Embase on Ovid and have Embase map the entered expression onto the Emtree thesaurus. Compare lines #1 and #2 below.

Embase <1974 to Present>

Although Ovid MEDLINE to Embase on Ovid searches appear to function seamlessly in an operational sense, in terms of being informationally useful, further assumption checks and testing may be needed.  

Strategy

"Drug-Related Side Effects and Adverse Reactions"/

[ Entry term is 'adverse drug event', term is mapped to MeSH ]

adverse drug event/

[ Entry term is 'adverse drug event/' ,i.e. ends with a backslash ]

adverse drug event.mp.

[ Entry term is 'adverse drug event.mp.' ,i.e. runs an multi-purpose fields' search ]

docz.dz. [ Now go to Change below, select Embase and re-run the above search in Embase on Ovid ]