*5b. Correcting the MeSH Lab and Additional MeSH Practice

Objectives:

Before proceeding to the correction and practice portion of today’s session, I want to give you some important MeSH search tips.  These tips are based on the way MEDLINE is indexed.  These may not make sense immediately, but don’t despair.  I will explain and provide examples.

  1. Consider using a floating subheading
  2.  Remember that MeSH indexing is disease-focused
  3. Use broad subheadings
  4. Try searches with and without subheadings.
  5. Only use a subheading if it’s important to focus on a specific aspect of a topic

 

I’ll use stories of searches that were requested during my early career to illustrate these points.

Consider using a floating subheading

Example 1:

During the early years of my career as a librarian, I was asked for articles that focused on endoscopic pituitary surgery methods.   I started with the following MeSH search:

“Pituitary Gland/surgery”[MeSH] AND “Endoscopy”[Mesh] AND “methods” [Subheading]

“methods” is a floating subheading.

A floating subheading  will retrieve any record with that subheading regardless of the heading the indexer actually used with that subheading.   I didn’t know what type of methods would be mentioned — endoscopy methods, anesthesia methods, hemorrhage control methods , etc. — I just knew that I wanted articles that focused on methods.

The process of adding a floating subheading to a search is no different than adding a heading to a search.  Both can be located by searching the MeSH database, selected, and added to the search using the “Add to Search Builder” button.

The “methods” floating subheading was helpful.  My results were focused on pituitary surgery methods.  However, when I compared my search results to the results of a quick keyword search, I found that I had missed many relevant MEDLINE records.  Then I remembered —

 MeSH indexing is disease-centered

When an article can be indexed as being about a disease or group of diseases, it will be.  Most articles about pituitary surgery methods are indexed as being about surgical treatment of a specific pituitary disease/s.    Only the articles  about surgical methods that are not targeted at a particular disease (but pituitary surgery in general) are likely to get the “Pitutiary Gland/surgery” heading/subheading combination.  Those focused on a specific pituitary condition will be indexed with that condition’s headings and the surgery subheading (for example: “Pituitary Apoplexy/surgery”).  All of the specific pituitary disease headings will be retrieved by a search for  “Pituitary Diseases”[MeSH].    I revised my search strategy:

(“Pituitary Diseases/surgery”[MeSH] OR “Pituitary Gland/surgery”[MeSH] ) AND “Endoscopy”[Mesh] AND “methods” [Subheading]

Example 2:

At another time I needed to search for articles about weight loss diets.  There’s a good heading for weight loss diets —

“Diet, Reducing”[Mesh]

I again found that a search using this heading was missing a lot of relevant MEDLINE records.  Why?  It wasn’t a heading introduction date problem.  The “Diet Reducing”[MeSH] heading was introduced into the MeSH vocabulary before MEDLINE existed in online form.

Eventually, I found that most of the other relevant articles were indexed as articles about treatment  of obesity using either

“Obesity/diet therapy”[MeSH]

or

“Obesity/therapy”[MeSH]

The history table below shows that the sets of records retrieved by “Diet Reducing”[MeSH]  and “Obesity/diet therapy”[MeSH] overlap but are not identical.

A screenshot shows that the number of results of a search for "Diet, Reducing"[MeSH] OR "Obesity/diet therapy"[MeSH] is much greater than the number of results retrieved by each of the terms separately.

I mentioned that some of the relevant records retrieved by the keyword search were indexed with “Obesity/therapy”[MeSH] rather than “Obesity/diet therapy”[MeSH] .   Can you guess why this is true?  An explanation brings me to my next point —

 

Use broad subheadings when applicable

The “therapy”, “therapeutic use”, “adverse effects” , and “etiology” subheadings recommended by two of the search templates are ‘broad’ subheadings.   All four have narrower subheadings that may be applicable to your search and will be sitting there in the list of subheadings, saying, “pick me, pick me”  every time you build a MeSH search.    The reason you should remember the broader subheadings is explained below:

“Therapy” subheading

Here’s the MeSH tree for the “therapy” subheading:

The therapy subheading with narrower subheadings -- diet therapy, drug therapy, nursing, prevention and control, radiotherapy, rehabilitation, surgery, transplantation -- is shown. Arrows show that if two narrower subheading apply the broader subheading should be used.

In general, indexers only apply one subheading from a particular subheading tree.  If an article covers two different types of therapy — drug therapy plus massage, drug therapy plus diet therapy,  etc — the indexer will use the broad “therapy” subheading.   If you want to retrieve these  multi-modality therapy articles you will need to use the “therapy” subheading rather than the “drug therapy” subheading.

Think twice before using a “_________/drug therapy” heading/subheading combo alone! 

In PubMed, subheadings are always ‘exploded’ a search for a heading in combination with the “therapy” subheading will retrieve any records indexed with the chosen heading  (or narrower headings) with either the therapy subheading or any narrower subheading ( including the “drug therapy” subheading).    Of course, when you use the “therapy” subheading you could retrieve articles about “diet therapy” that don’t mention drug therapy.   However, your searches will usually have a separate drug concept that will help focus your search.

“Therapeutic use” subheading

Here’s the “therapeutic use” tree:

A screenshot of the "Therapeutic use" tree shows that "If an article concerns "administration and dosage" and "adverse effects", it will be indexed with "therapeutic use"

In general, avoid using  a “______/administration and dosage”[MeSH] combination alone.  Instead use the broader “_________/therapeutic use”[MeSH] (or use both heading/subheading combinations OR’d together).

Remember, PubMed will explode the “therapeutic use” subheading.   A heading in combination with the “therapeutic use” subheading will retrieve any record indexed with the chosen heading (or a narrower heading) in combination with the “therapeutic use” subheading or a narrower subheading.

“Etiology” subheading

Here’s the “etiology” subheading tree:

If an article contains concerns a drug-induced condition ("chemically induced") and also involves another narrower subheading in the "etiology" tree ( like "immunology") the article will be indexed with "etiology"

In general,  avoid using a “____________/chemically-induced”[MeSH}  combination alone.  Instead use the “_____________/etiology”{MeSH} combination or use both heading/subheading combinations OR’d together.

 

“Adverse effects” subheading

Here’s the “adverse effects” subheading tree:

A screenshot of the "Adverse Effects" tree shows that Adverse effects" will retrieve records indexed with "adverse effects", "poisoning" OR "toxicity

In general, do not use the “_____________/toxicity”[MeSH] combination alone.  Instead use the “_______________/adverse effects”[MeSH] combination or both heading/subheading combinations OR’d together.

Try searches with and without subheadings.

There are several reasons to try removing the subheadings form a MeSH search.  All you need to remember is that it’s worth trying MeSH searches with and without subheadings.

For those who want the reasons it is worth trying a search without subheadings:

  1.  Some headings have not always been headings, but have a spent a portion of their ‘lives’ in the MeSH database as “Supplementary Concepts” .  Supplementary Concepts are unofficial headings that don’t have subheadings.   When a “Supplementary Concept” becomes an official MeSH heading, all instances of the “Supplementary Concept” are changed to the new MeSH heading but subheadings are not added.  This leaves a lot of indexed records with no subheadings.  New drugs are usually represented by “Supplementary Concepts” for years before that ‘powers that be’ decide an official MeSH is needed.   It’s especially important to try removing subheadings you’ve used in combination with drug headings.
  2. MEDLINE indexers are only allowed a limited number of subheadings (~3) with each heading.  If more subheadings apply to an article’s subject matter, the indexer must use the heading without subheadings.
  3.  An indexer may use an expected subheading with a heading different than the expected heading.

Example of a record that was not indexed in the way I expected:

Title:

    • “A randomized controlled study for Yuanhu Zhitong dropping pills in the treatment of knee osteoarthritis”

Indexing I expected:

    • “Osteoarthritis/therapy”

Actual indexing:

    • “Arthralgia/drug therapy” (joint pain therapy)
    • “Osteoarthritis, Knee / complications”

 

Only use a subheading if it’s important to focus on a specific aspect of a topic

Let’s say you need to do a search for articles about using methotrexate to treat rheumatoid arthritis in patients with kidney failure.  Part of the search is a “using a drug to treat a disease” search.   I’d be planning

methotrexate heading plus therapeutic use subheading

AND

rheumatoid arthritis heading plus therapy subheading

AND

kidney failure

Do you want a subheading with kidney failure?  Students tend to look at subheadings every time they add a heading.  Remember, the heading without a subheading will retrieve anything indexed with that heading with or without subheadings.  If you don’t need to focus on an aspect of a topic, don’t add a subheading!

Practice and Correction

You’ve now done a bit of MeSH searching.   You’ve used templates to help you plan MeSH searches for :

  1. a drug used to treat a disease
  2. a drug used to prevent a  disease
  3. a disease caused by drug adverse effects

You will be confronted with these three types of searches often during your careers, and I would prefer that the heading/subheading combinations suggested by the templates make sense to you, that they become second nature.  During this session, you’ll get some more practice with the these three types of searches.    You’ll be provided the templates on the exam, but shouldn’t really need to consult them.

 

Drug causes a disease =  adverse effect of a drug is the cause (etiology) of a disease =

drug/adverse effects AND disease/etiology

 

Drug treats a disease = therapeutic use of a drug is the treatment for a disease =

drug/therapeutic use AND disease/therapy

 

Drug prevents a disease = drug is used for prevention of the disease =

Drug AND disease/prevention and control

 

If you were sitting in class with me, I would make you all say recite these several times as a class.  I’ve heard that speaking can serve to reinforce memory.  I suggest you whisper  or speak the bold-faced information above.

 

To practice, use the MeSH database to construct searches for the topics listed below.  It’s difficult to focus on these topics if you don’t use heading/subheading combinations.

You may remember that the final topic above was the topic of the first assignment you completed last week.  Check your work on that assignment to be sure you understand the correct approach.

The process of completing last week’s second assignment (the MeSH search for your final search assignment) is demonstrated in the video below.  Please,  review your work on that assignment while watching the video.

Watch the linked video (required) and check your work on FSA 3.

 

Required video focused on correcting FSA3 thumbnail of the nilotinib, cml, resistance MeSH search video.

Assignment

  1. Watch the linked video above.
  2. Either;

a) Correct your own FSA 3 assignment and b) a classmate’s FSA 3 assignment

or

a) Correct your own search ,  b)  list the errors in the sample search below , and  c) correct the sample search:

 

sample search:

“Leukemia, Myelogenous, Chronic, BCR-ABL Positive/etiology”[Mesh] AND “nilotinib” [Supplementary Concept] OR “Multidrug Resistance-Associated Proteins”[Mesh] 

Filters: Randomized Controlled Trial, Humans, English

 

  1.   Submit your own revised assignment with your own corrections and corrections based on your classmate’s advice (if you agree with their advice) as “FSA 3 corrected”.    Include:
    • Your name
    • Your classmate’s name or the list of errors in the sample search and  corrected sample search.
    • Your  corrected search strategy.
      • Please, check to be sure the entire search strategy is visible.
      •  Be sure your submission includes a copy of your corrected search that is copyable so that you can copy, revise, and rerun it in the future.  You will need to work on your search strategy again, during the session on refining searches.
    • The earliest publication year your search can be trusted to search (explained in video).  Please label this “earliest year.”

 

Questions, Problems, Text Errors?

Before you leave, …

  • Do you have any questions or do you feel that clarification of some aspect of the materials would be helpful?
  • Have you noticed any errors or problems with course materials that you’d like to report?
  • Do you have any other comments?

If so, you can submit questions, comments, corrections, and concerns anonymously — or with your e-mail (your choice) — through this online form .  Alternatively, you’re always welcome to contact Cindy Schmidt directly 402-650-5056, cmschmidt@unmc.edu, or by making an appointment to meet with Cindy via Zoom.

Answers to questions or requests for clarification that are submitted anonymously will be answered in Canvas on the “Discussions” page for this course.

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Introduction to Drug Information Copyright © by Cynthia M. Schmidt is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book