The formula for the pooled effect estimate on page 5 contains an incorrect subscript. The formula should read pooled inline image. This error does not affect any of the calculations that follow.

We also wish to clarify that the Cochran's Q statistic and Wald's Z-statistic reported on page 5 are equivalent tests, and will therefore always provide the same p-values when assessing homogeneity of stratum specific effects.

Neither of these errors alters the substantive recommendations of our commentary.

The authors regret these errors.