Several errors occurred in table 2 of this paper by Jadad et al (26 February, pp 537-40). Under the heading “Was bias in the selection of the studies avoided?” the number for “all” should be 14/30, not 15/30; under “Were the findings of the relevant studies combined appropriately?” the number for “all” should be 22/26, not 24/26; and under: “Were the conclusions made by the author(s) supported by the data provided?” the number for “peer reviewed journals” should be 14/38, not 14/14. The P values are the same.