HK 710: Meta-Analysis

Library resources to support the theory and application of meta-analytic techniques for quantitative analysis and review of scientific literature.

Evidence Based Practice

Types of Reviews

There are several different kinds of articles frequently found in the literature for medical and health sciences. Here are the best:

  • Meta-analysis (PLATINUM STANDARD!) - A method of synthesizing the data from more than one study, in order to produce a summary statistic. 
  • Systematic review – An approach that involves capturing and assessing the evidence by some systematic method, where all the components of the approach and the assessment are made explicit and documented.  Some systematic reviews include a meta-analysis (see above).
  • Scoping review - A type of knowledge synthesis that uses a systematic and iterative approach to identify and synthesize an existing or emerging body of literature on a given topic. The key differences between systematic reviews and scoping reviews are differing purposes and aims. The purpose of a scoping review is to map the body of literature on a topic area. The purpose of a systematic review is to sum up the best available research on a specific question.
  • Randomized Controlled Trial – An experimental study in which users are randomly allocated to one of two or more options, where some get the option of interest and others get another option (e.g. a standard service).

Content Contributor: Booth A & Brice A (2004) Evidence Based Practice for Information Professionals: A Handbook. London: Facet Publishing.

Citation Help Videos Using EndNote

PICO Method

PICO Method

P: Patient, Problem or Population – What are the most important characteristics of the patient and their health status?

I: Intervention – What main intervention are you considering (medical, surgical, preventative)

C: Comparison – What are the alternative benchmark or goldmark standards being considered, if any?

O: Outcome – What is the estimated likelihood of a clinical outcome attributable to a specific disease, condition or injury?

 

Databases to find Articles

Meta-Analysis Examples

Literature Searches And PRISMA

When conducting literature searches, you should know that the average time to complete is about 44 hours!
UNC Health Sciences Library Guide provided the following helpful information:
When you design a search strategy to find all of the articles related to your research question. You will:
  1. Define the main concepts of your topic
  2. Choose which databases you want to search
  3. List terms to describe each concept
  4. Add terms from controlled vocabulary like MeSH
  5. Use field tags to tell the database where to search for terms
  6. Combine terms and concepts with Boolean operators AND and OR
  7. Translate your search strategy to match the format standards for each database
  8. Save a copy of your search strategy and details about your search

Step 1: Preparation To complete the the PRISMA diagram, save a copy of the diagram to use alongside your searches. It can be downloaded from the PRISMA website

Step 2: Doing the Database Search Run the search for each database individually, including ALL your search terms, any MeSH or other subject headings, truncation (like hemipleg*), and/or wildcards (like sul?ur). Apply all your limits (such as years of search, English language only, and so on). Once all search terms have been combined and you have applied all relevant limits, you should have a final number of records or articles for each database. Enter this information in the top left box of the PRISMA flow chart. You should add the total number of combined results from all databases (including duplicates) after the equal sign where it says Databases (n=). Many researchers also add notations in the box for the number of results from each database search, for example, Pubmed (n=335), Embase (n= 600), and so on.  If you search trial registers, such as ClinicalTrials.gov, you should enter that number after the equal sign in Registers (n=).

NOTE: Some citation managers automatically remove duplicates with each file you import.  Be sure to capture the number of articles from your database searches before any duplicates are removed.

 

Records identified from databases or registers

 

Step 3: Remove All Duplicates To avoid reviewing duplicate articles, you need to remove any articles that appear more than once in your results. You may want to export the entire list of articles from each database to a citation manager such as EndNote, Zotero, or Mendeley (including both citation and abstract in your file) and remove the duplicates there. If you are using Covidence for your review, you should also add the duplicate articles identified in Covidence to the citation manager number.  Enter the number of records removed as duplicates in the second box on your PRISMA template.  If you are using automation tools to help evaluate the relevance of citations in your results, you would also enter that number here.

Records removed before screening: duplicates, automation tool exclusions, or other reasons

NOTE: If you are using Covidence to screen your articles, you can copy the numbers from the PRISMA diagram in your Covidence review into the boxes mentioned below.  Covidence does not include the number of results from each database, so you will need to keep track of that  number yourself.

Step 4: Records Screened- Title/Abstract Screening The next step is to add the number of articles that you will screen. This should be the number of records identified minus the number from the duplicates removed box.

Number of records screened in Title/Abstract level

 

Step 5: Records Excluded- Title/Abstract Screening You will need to screen the titles and abstracts for articles which are relevant to your research question. Any articles that appear to help you provide an answer to your research question should be included. Record the number of articles excluded through title/abstract screening in the box to the right titled "Records excluded."  You can optionally add exclusion reasons at this level, but they are not required until full text screening.

Records excluded after title & abstract screening

 

Step 6: Reports Sought for Retrieval This is the number of articles you obtain in preparation for full text screening.  Subtract the number of excluded records (Step 5) from the total number screened (Step 4) and this will be your number sought for retrieval.

Reports sought for retrieval

Step 7: Reports Not Retrieved List the number of articles for which you are unable to find the full text.  (Note: You can use Interlibrary Loan to request articles to see if we can order them from other libraries before automatically excluding them.)

Reports not retrived

Step 8: Reports Assessed for Eligibility- Full Text Screening  This should be the number of reports sought for retrieval (Step 6) minus the number of reports not retrieved (Step 7). Review the full text for these articles to assess their eligibility for inclusion in your systematic review. 

Reports assessed for eligibility

 

Step 9: Reports Excluded After reviewing all articles in the full-text screening stage for eligibility, enter the total number of articles you exclude in the box titled "Reports excluded," and then list your reasons for excluding the articles as well as the number of records excluded for each reason.  Examples include wrong setting, wrong patient population, wrong intervention, wrong dosage, etc.  You should only count an excluded article once in your list even if if meets multiple exclusion criteria.

Reports excluded, including reason for exclusion and number

 

Step 10: Included Studies The final step is to subtract the number of records excluded during the eligibility review of full-texts (Step 9) from the total number of articles reviewed for eligibility (Step 8). Enter this number in the box labeled "Studies included in review," combining numbers with your grey literature search results in this box if needed.  You have now completed your PRISMA flow diagram, unless you have also performed searches in non-database sources.

Studies included in review

 

*Thanks to UNC Health Sciences Library for this helpful walkthrough!*

 

When designing and conducting literature searches, a librarian can help advise you on:

  • How to create a search strategy with Boolean operators, database-specific syntax, subject headings, and appropriate keywords 
  • How to apply previously published systematic review search strategies to your current search
  • How to translate a search strategy from one database's preferred structure to another

 

EXAMPLE SEARCH STRATEGY USED IN A SYSTEMATIC REVIEW PAPER:

Systematic Review Help

Need more help with your Systematic Reviews? Librarians are available to assist you! Find more tips and resources on the Systematic Review library research guide.