How to use the PICO Method in an Ovid Database (Source: Allan Finn, Ovid)
The Table Method
One way of planning to do a search in a database, is to place each of the subjects from your clinical question in a box, in a spreadsheet or table, and then use it to think about the subjects that are similar or related to your question. The words and phrases in a row become the similar items that you OR together in your search. Then the columns are ANDed to find where they occur together in a record. It’s a way of gathering up your ideas before you begin a search to answer a question. Often when we perform a database search we are hoping to answer a specific clinical question.
Q: Does Echinacea help children with Winter Colds ?
|
Question |
Synonyms |
|
Flu |
Influenza, Common Cold |
|
Echinacea |
|
|
Children |
Child, Adolescent |
|
Winter |
Seasons |
The PICO Method
The PICO method adds to the table method by placing your subjects in a structure. This helps you to focus on the roles of your subjects (and so where they are located in the Ovid interface), and helps to make sure you have included all the items you are interested in in your search. In our example above, are we asking about prevention or treatment ? PICO is a method of creating a strategy to find relevant information, to answer a clinical question.
How can I use the PICO method in a database ? Where are each of the parts of my question ? PICO(T) stands for Population, Intervention, Comparison (comparison can be with placebo or with no treatment), Outcome and Time frame (which is optional).
Where do I find these items in the Ovid interface ?
|
|
Item |
Where in Ovid |
|
P |
Patient or Population |
Subject Mapping or Additional Limits Page (Age group or population) |
|
I |
Intervention or Issue |
Subject (might be more than one) |
|
C |
Comparison |
Subject (alternative or control – or no comparison) |
|
O |
Outcome |
Subject or Subheading as part of the Mapping process or Limit |
|
T |
Time |
Often optional, time can be a difficult subject |
Example 1 – Prevention of Whooping Cough in adults using vaccination, during an epidemic.
|
Item |
Example |
|
Patient or Population |
Adults (aged over 18 years) |
|
Intervention or Issue |
Whooping Cough, Vaccination |
|
Comparison |
no vaccination ? |
|
Outcome |
Prevention (Whooping Cough prevention) |
|
Time |
Epidemic ? |
Example 1 Ovid Search
|
# |
Searches |
|
1 |
Exp Whooping Cough/pc [Prevention & Control] |
|
2 |
exp Vaccination/ |
|
3 |
1 and 2 |
|
4 |
limit 3 to "all adult (19 plus years)" {could also be a subject} |
|
5 |
exp Epidemics/ |
|
6 |
4 and 5 |
Note that the PICO search entries aren’t in the same order as in the PICO table. Sometimes time is a difficult concept, we often use “published in the last five years” as a criteria, but what is “early treatment” for example, is it an hour, a day or a week later ?
Example 2 – Treatment of men’s ankle injuries after playing football.
|
Item |
Example |
|
Patient or Population |
Men (Male) |
|
Intervention or Issue |
Ankle Injuries during Football |
|
Comparison |
No comparison |
|
Outcome |
Treatment (more than one type) |
|
Time |
|
Example 2 Ovid Search
|
# |
Searches |
|
1 |
exp Football/ |
|
2 |
exp Ankle Injuries/dt, rh, su, th [Drug Therapy, Rehabilitation, Surgery, Therapy] |
|
3 |
1 and 2 |
|
4 |
limit 3 to male |
The subheadings (in orange above) are a very specific way of inserting the outcomes, I could have also used a Limit. Not all the parts of PICO need to be present all the time. But are Men’s and women’s ankles all that different ? Do we really need to use Male ?
Example 3
A middle aged male patient has just been diagnosed with Type 2 diabetes mellitus. Diabetes can be associated with an increased risk of morbidity and mortality due to cardiovascular diseases, such as stroke or diabetic kidney disease. Although the patient’s blood pressure is normal, you wonder if there is any evidence to support blood pressure lowering medication to reduce the risk of these diseases, in addition to his anti-diabetic medication.
Firstly it is necessary to construct an answerable question in a way that helps focus a search for patient specific information.
The analysis could look like this:
|
Item |
Example |
|
Patient, Population |
A middle aged man (Population) |
|
Intervention or Issue |
blood pressure lowering medication, Type 2 diabetes and normal blood pressure (?) |
|
Comparison |
No comparison in this case |
|
Outcomes |
Prevention of morbidity and mortality from cardiovascular diseases, such as stroke or diabetic kidney disease. |
The question could be: In middle aged men with type 2 diabetes and normal blood pressure, does blood pressure lowering medication help prevent the risk of morbidity and mortality from cardiovascular diseases, such as stroke or diabetic kidney disease?
Let’s work through this example in Ovid, looking at where each of the items could be in the interface.
Example 3 Ovid Search
|
# |
Searches |
|
1 |
exp Diabetes Mellitus, Type 2/ |
|
2 |
exp Antihypertensive Agents/ |
|
3 |
exp Stroke/pc [Prevention & Control] |
|
4 |
exp Diabetic Nephropathies/pc [Prevention & Control] |
|
5 |
1 and 2 and (3 or 4) |
|
6 |
limit 5 to (male and "middle aged (45 plus years)") |
Try your own PICO example…
Question:
|
|
|
P |
|
|
|
I |
|
|
|
C |
|
|
|
O |
|
|
|
T |
|
|
Preparing Research for Systematic Reviews, or broadening your search
If you are planning to create a systematic review, it is usually best to start with a PICO style question, it provides structure and boundaries to the task. Let’s try:
Can nicotine replacement therapy cause insomnia in adults
|
|
Item |
Where in Ovid |
|
|
P |
Patient or Population |
Subject Mapping or Additional Limits Page (Age group or population) |
Adult |
|
I |
Intervention or Issue |
Subject (might be more than one) |
Nicotine Replacement Therapy |
|
C |
Comparison |
Subject (alternative or control – or no comparison) |
|
|
O |
Outcome |
Subject or Subheading as part of the Mapping process or Limit |
Cause Insomnia |
|
T |
Time |
Often optional, time can be a difficult subject |
|
To search for the core topics discussed in your review, you are probably going to want to use a combination of approaches. You are probably going to need to use the subject structure of a database: MeSH, Emtree etc., which allows you to quickly do a precision search which groups together subjects. But you probably will also want to include keyword searches which include the authors own words, and will allow you to find early references and more marginal references for your subject. For some special subjects you may also need to include other searching strategies.
Keywords
Adding keywords to a search allows you to use more of the authors own word, or words outside the database vocabulary which also contain subject terms. The keyword fields are set for each database and they take the abbreviation .mp. When we use this method of searching, we need to allow for the variable way in which a subject might be entered. To do this we can use search operators and truncation and wildcard symbols. When I enter a phrase into Ovid, even if I am searching in a group of keyword fields, I am searching that specific phrase, each word in that specific order. The only exceptions are a small number of stopwords which are not available for searching. Look below in the table to see how we might change and develop a list of keywords…
|
Phrase |
Truncation |
Operator |
|
Nicotine patch |
Nicotine patch* (includes nicotine patches) |
|
|
Nicotine transdermal |
Nicotine trans* (for transdermal and transcutaneous) |
Nicotine adj2 trans* (for transdermal administration of nicotine) |
|
Nicotine gum |
Nicotine gum? (for nicotine gums) |
|
|
Nicotine replacement therapy |
Nicotine replacement therap* |
Nicotine adj replacement adj therap* |
Finding lists of keywords
Keywords can be found in the titles and abstracts of the results that you have found in your subject heading search, scan some likely choices to see how authors refer to your topic of interest, and cut and paste into a list.
Keywords can also come from the scope notes that are attached to subjects, they are lists of frequently used synonyms, but for best results, add truncation and wildcards where necessary.