Reference management software
For a systematic review we recommend that you export all search results to a reference management program such as EndNote or Zotero and then remove any duplicates.
Deduplication
Due to overlap between the bibliographic databases, you will get a number of duplicate records when searching multiple databases. Both EndNote and Zotero can be used to remove duplicates. If you use our search service for systematic reviews, the librarian will remove duplicates in EndNote. We can also deliver a RIS file that can be imported into Zotero and screening programs such as Covidence and Rayyan. The total number of hits from bibliographic databases and number of remaining records after deduplication should be reported in a PRISMA flow diagram.
Study selection (screening)
The step following the literature search is study selection or screening. It is important to have clear inclusion and exclusion criteria to know which records to include and exclude.
Screening is usually done in two steps and should be done by (at least) two people independently. In step 1, you screen titles and abstracts. If at this stage you are in doubt as to whether to include or exclude a record, you should include it. When both have finished, you go through the decisions made by each reviewer and resolve any conflicts. To be sure that both of you interpret the inclusion and exclusion criteria in the same way, you may want to start by screening a sample of 200 records.
Records included in step 1 are retrieved in full-text and carried forward to step 2. The inclusion and exclusion criteria are used again to assess whether the full-text articles should be included or excluded. When excluding on the basis of full text, a reason for exclusion must always be stated, e.g. wrong population, wrong intervention or wrong study design. This step should also be performed by two people independently of each other.
We recommend using a screening tool such as Covidence.
Covidence
Covidence is an online tool that allows two or more people to screen references independently. In addition to screening of titles and abstracts (step 1), and full-text screening (step 2), Covidence can be used for data extraction and critical assessment of methodological quality (risk of bias assessment). In the video below you get an overview of the functions in Covidence.
The University of Oslo (UiO) has a Covidence institutional license. Students and staff at UiO and Oslo University Hospital (OUS) can create an unlimited number of reviews in Covidence. On the Covidence support pages, you will find more detailed information, as well as instructions on how to create an account or link your existing account to UiO's license.
Get full text articles
Via the University Library, you have access to a large collection of electronic journals. The full text of many articles are most easily retrieved via EndNote or by searching Oria as described on the support pages for EndNote. You can also search for full text via Google Scholar or ResearchGate. If there are any articles you can not find using these methods, you can send a request by e-mail to your library.
Critical appraisal
Critical appraisal of the methodological quality, also called or risk of bias assessment, in the included studies should be done for all systematic reviews. More about critical appraisal and links to checklists can be found on the methods and standards page.