Skip to main content

Systematic Reviews

A guide for finding standards, procedures, resources, and assistance in supporting systematic searches of the biomedical literature.

Steps in a Systematic Review

1. Develop a research question

   • Often in a PICO format

   • Should run a search for prior reviews on the topic area; use this information to create an evidence table documenting what is known about the issue, where there are gaps in knowledge, and what is still open for further exploration

2. Define inclusion and exclusion criteria

   • Current standards request reviewers to register their review in PROSPERO

3. Locate studies

   • Devise reproducible search methods and launch an exhaustive search

   • Search in multiple databases to reduce bias

   • Record strategies and track results for reporting

4. Select studies for your review based on inclusion/exclusion criteria

   • Assess study quality

5. Extract data into a standardized format

6. Analyze and present results

   • Synthesize data using statistical methods for a meta-analysis

7. Interpret results

   • Write about what you find

8. Update the review as needed

Timeline

The average systematic review requires at least 12 months of work. 

Box 2.3.b: Timeline for a Cochrane Review
Month Activity
1 - 2  Preparation of protocol
3 - 8  Searches for published and unpublished studies
2 - 3  Pilot test of eligibility criteria
3 - 8 Inclusion assessments
Pilot test of 'Risk of bias' assessment
3 - 10  Validity assessments
Pilot test of data collection
3 - 10 Data collection
3 - 10 Data entry
5 - 11 Follow up of missing information
8 - 10 Analysis
1 - 11 Preparation of review report
12 -  Keeping the review up-tp-date

 

 From: Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available from handbook.cochrane.org.

What You Need to Conduct a Systematic Review

Team Members

 

  • Content expert
  • 2 reviewers
  • 1 tie breaker
  • 1 statistician (meta-analysis)
  • *1 information professional (librarian trained in systematic reviews)
  • 1 administrative support person

“Expert searchers are an important part of the systematic review team, crucial throughout the review process-from the development of the proposal and research question to publication.” (McGowan & Sampson, 2005)

*The information professional should write the methods section regarding the search methods.You may also want to consider providing a copy of one of the search strategies used in an appendix.

Searching

The search strategy the information professional develops should be guided by a detailed PICO question, inclusion and exclusion criteria for the study, and two-four articles that match the criteria of the review.

What databases the information professional searches depends on the topic of the review, but should include:

  • At least three standard medical databases like PubMed/Medline, CINAHL, Embase, etc..
  • At least three grey literature resources like Clinicaltrials.gov, SIGLE, ProQuest Dissertations & Theses, etc...

Citation Management System

Employing a citation management system is essential in handling the large number of citations generated in a systematic review. Many research teams have success using EndNote, but teams should also consider Mendeley, RefWorks, or another system depending on what the team members are most comfortable using. The information professional can help with the selection of a citation management system. For more support on using citation management programs, consult: http://hslguides.med.nyu.edu/citationmanagement

Note: NYU faculty and staff have access to EndNote and RefWorks provided by the Libraries. Mendeley is a freely-available option.

Articles

The administrative support person can focus on obtaining the physical articles. The information professional can help train this individual to use the tools and services provided by the Health Sciences Libraries.. If your team is using EndNote, you can automatically download available full-text articles.

The Libraries will not have immediate access to every article selected as part of the review. The Libraries' document delivery service can obtain articles that are not part of the at-hand collection. This service is provided to NYU faculty and staff, and requests may be submitted at: http://hsl.med.nyu.edu/form/request-article-or-book