Evaluation of inconsistency in networks of interventions

doi:10.1093/ije/dys222 Int J Epidemiol 2013;42:332–45 Following the publication of our manuscript we noticed two minor errors: Page 341, fourth line: The text ‘The REML Wald test indicated 5 (13%), 6 (15%), 7 (17%) and 5 (12%) inconsistent networks’ should have read: ‘The REML Wald test indicated 5 (13%), 6 (15%), 7 (17%) and 5 (13%) inconsistent networks’ Page 341, line 14 The section ‘(12% to 17%, depending on effect measure) than the latter (20% to 28%, depending on effect measure)’ should have read: ‘(13% to 17% depending on effect measure) than the latter (20% to 28% depending on effect measure)’ The authors apologise for these errors.