Skip to main content

Systematic Reviews

A guide for researchers beginning a medical systematic review or meta-analysis.

Team Members Involved

The Librarian or information specialist will conduct the primary searches in major literature databases and record details of the search strategies and results.

The team's content experts perform any hand searching of journals or conference programs, or reference lists from relevant articles.

When clinical trial data will be used, the librarian can local relevant trials, but the team's content experts should contact trial investigators to request any unpublished results.

Databases

The Mulford Librarians will use the following databases for the identification of studies for systematic reviews as relevant to your topic. Other databases specific to the topic or the discipline may also be searched.

Beyond Databases

Any formalized database search can be enriched by two activities.

1) Hand Searching: This is the activity of mining the cited references of your selected studies for any additional relevant works.  This should be performed by a content expert on the team (not typically the librarian).

2) Web Searching: Search engines like Google Scholar can be great discovery tools. Unfortunately, since they generally do not have the advanced search features necessary to create a focused, reproducible search they are not typically a part of the systematic review approach.

Grey Literature

Grey Literature refers to literature which is not published through traditional publishers. Though a broad term, it usually means content released directly to the web. The Cochrane Collaborative discusses the utility of grey literature in preparing systematic reviews: Cochrane: (6.2.1.8  Grey literature databases)

Clinical Trials

In order to be comprehensive, you may want to expand your search to include unpublished clinical trial results.  These databases can help.  Content experts on the team may wish to contact the principle investigators of ongoing trials to see if there are results that could be included in the meta-analysis.  Likewise, trials that are completed but have no published results may be important sources to reduce bias from only analyzing published studies.  These 'negative results' may be very useful to your analysis.