Skip to main content


Stony Brook University

Systematic Reviews: Getting Started

A guide to conducting systematic reviews.

Software Tools

Consider using the following software/websites to facilitate the process. There are a number of alternatives out there, but these are among the best and will save you a lot of time.

  1. EndNote - this is freely available to Stony Brook faculty, students and staff with a NetID through SBU's Division of Information Technology (http://it.stonybrook.edu) - Login to Softweb and search for EndNote. Download and install the software. Recommended: a video on the use of Endote for Systematic Reviews.
  2. Covidence (http://www.covidence.org) - Covidence offers one free trial review per individual before requiring payment for further use. The Health Sciences Library is investigating licensing of this or one of its competitors (DistillerSR or Eppi-Reviewer), but in the meantime take advantage of your one free review. This is a big time-saver and very easy to use. Covidence also offers some resources for bias assessment and data extraction, but its best feature is the easy screening process.
  3. RevMan (Review Manager) from the Cochrane Collaboration - A very useful program designed for Cochrane authors to write their systematic reviews.It is free for download and a web version may soon be available. When installing, unless you're actually doing a Cochrane review, make sure to select "Non-Cochrane" as the type of review. This provides you with a template for writing the review, extracting data, judging bias, creating tables, and running some basic statistics. 
  4. GradePro (http://www.guidelinedevelopment.org/) - Create a free account. This is synergistic with RevMan and will create Summary of Findings tables for your review, including GRADE evidence evaluation of studies. See also http://tech.cochrane.org/gradepro for more information.

Planning

  1. Form a research team.
    1. One person can't do a systematic review. You will need two people at the absolute minimum, preferably more. At least one team of two individuals is required to do independent screening of abstracts and full text as well as data extraction; be prepared to calculate inter-rater reliability statistics.
    2. You will need specialized expertise in the topic area, literature searching, and statistics, especially if you wish to do a meta-analysis.
      1. It is strongly recommended to involve a librarian to help you select resources, create the search protocol, and manage references. The Institute of Medicine strongly recommends librarian involvement. It decreases bias and will make writing up your methodology easier.
      2. Have someone who can do statistics.
  2. Do a bit of preliminary, unofficial searching to get a sense of how much is out there on your topic and what kinds of studies are available. You may need to narrow or broaden your topic and having a couple of pre-identified key studies can help with finding search terms and developing search strategies.