Assessment of heterogeneity
Heterogeneity will be explored by Q test (Chi2 and I2 tests). We
will interpret a Q test resulting in a P value less than 0.10 as indicating
significant statistical heterogeneity. In order to assess and
quantify the possible magnitude of inconsistency (i.e. heterogeneity)
across studies, we will also use the I2 statistic as a rough guide
for interpretation as follows (Higgins 2003): an I2 of 30% or under
will indicate a low level of heterogeneity and justify use of a
fixed-effectmodel for meta-analysis; values from31%to 59%will
be considered to indicate moderate heterogeneity and a randomeffects
model will be used; values of 60% or above will indicate
considerable heterogeneity and meta-analysis will not be undertaken.