A literature search was conducted in October 2014 usingfive electronic databases: Ovid Medline (since 1946), Cumulative Index to Nursing and Allied Health Literature Plus (CINAHL Plus, since
1937), Scopus (since 1823), PsycINFO (since 1806) and Pubmed. A combination of keywords was used to identify relevant articles (Fig. 1). The articles identified were assessed against inclusion and exclusion criteria (Table 1). A secondary search was made by reviewing included studies' reference lists. Two reviewers independently extracted the data from the studies, whose quality was assessed by the critical appraisal tools for randomised controlled trials ofThe Joanna Briggs Institute (2011)(Table 2). Two reviewers independently rated the articles. Any disagreement arising over the rating was resolved through discussion andfinal consensus. The
quality rating ranged fromfive to seven out of ten (Table 3).