Using Subheadings in Your Literature Search

The aim of a search strategy is to find a manageable number of articles that are likely to be relevant to the research question. MeSH subheadings are used to make the search more specific, reducing the number of articles retrieved:

MeSH heading	Subheading	Number of articles
Hemophilia		27,197
	complications	5,877
	diagnosis	2,078
	drug therapy	3,518
	nursing	88

Not every article found will be useful – some will not be in English (unless only English language articles have been searched for), others will mostly be about something else even though they have been indexed with the relevant MeSH headings.

A simple search may be all that is needed. Clearly, if we’re interested in transition arrangements for young people with haemophilia, we can manage to read 18 articles. If it’s nursing we’re interested in, 88 is still do-able but approaching the limit of time spent efficiently. When numbers are in the hundreds or thousands, the search strategy needs to be refined while remaining true to the research question.

Consider ‘drug therapy for haemophilia’ – no-one would attempt this as a research topic because it’s too broad. More realistically, we would want to look at a very specific aspect of drug therapy. Let’s take patient satisfaction with factor replacement therapy as an example. In general, the more specific the question, the more complicated the search strategy and here we need to use several MeSH terms and subheadings:

MeSH heading	Relevant subheadings
hemophilia	drug therapy
blood coagulation factors	therapeutic use adverse effects administration and dosage complications
patient satisfaction	(some are available but none are relevant to this search)

Using more subheadings greatly limits the scope of your search. This is a good thing if you initially retrieve a lot of articles but use subheadings carefully or you will exclude too many articles and may miss something important. You can see from the examples in the next table that the number of articles retrieved depends on how you combine headings and subheadings.

MeSH heading	Number of articles retrieved
hemophilia	27,197
hemophilia/drug therapy	3,518
hemophilia AND patient satisfaction	110
hemophilia/drug therapy AND patient satisfaction	24
blood coagulation factors	438,006
blood coagulation factors/therapeutic use	39,491
blood coagulation factors/therapeutic use AND patient satisfaction	54
hemophilia AND blood coagulation factors AND patient satisfaction	28
blood coagulation factors/therapeutic use AND patient satisfaction AND hemophilia/drug therapy	19

It is important to use a subheading with the best MeSH heading – for example:

‘hemophilia AND blood coagulation factors/therapeutic use’ retrieves 610 articles

but

‘haemophilia/drug therapy AND blood coagulation factors’ retrieves 3,034

This is because an article about haemophilia will often mention blood coagulation factors but not always in the context of their use as a treatment.

Another caution about being too specific: what do we want to know about patient satisfaction? The MeSH heading is defined as:

The degree to which the individual regards the health care service or product or the manner in which it is delivered by the provider as useful, effective, or beneficial.

Patient satisfaction is an end in itself, as a measure of the effectiveness of a service in meeting user expectations. But might we also be interested in ‘quality of life’ or ‘treatment adherence and compliance’ as related outcomes? These examples illustrate the fundamental importance of having a research question that is clear about what we are looking for and taking great care in formulating a search strategy.