We thank Finley and Bero for their interest in our work.1 Their study was 1 of 37 randomized comparisons (involving a total of 12 355 participants) that we were able to synthesize. We agree with Finley and Bero about the importance of antidepressant concordance. Contrary to their letter, we pooled both depression outcomes and concordance with medication in 2 separate meta-analyses, both of which were improved in association with collaborative care (depression outcomes at 6 months: standardized mean difference, 0.247 [95% confidence interval, 0.176-0.318]; antidepressant concordance: odds ratio, 1.916 [95% confidence interval, 1.536-2.389]). Their study contributed to both of these analyses, and their results in relation to antidepressant concordance seem entirely consistent with our overall pooled result.
Gilbody S, Bower P, Sutton A, Fletcher J, Richards D. Clarification of Study and Citation—Reply. Arch Intern Med. 2007;167(22):2531-2532. doi:10.1001/archinte.167.22.2531-b