Results of the search
Our search strategy retrieved 1220 potentially relevant references
of which 183 were duplicate references. We screened a total of
1037 references (see Figure 1).We assessed the abstracts and where
required obtained the full texts of the journal articles. One article
needed to be translated in its entirety, from Portuguese to English,
because the abstract in English did not provide enough information.
At this stage 1009 abstracts were excluded as they were
clearly not relevant for the review. This left 28 full-text articles on
randomised controlled trials (RCTs) that were reviewed. One of
the articles was obtained ahead of press from the primary author.
A further 22 RCTs were excluded. The reasons for exclusion are
summarised in the Characteristics of excluded studies table.The
six remaining RCTs (reported in seven separate articles) met our
inclusion criteria and are described in detail in the Characteristics
of included studies table.
Brief