In Reply We thank Faltinsen and colleagues for their judicious comments on our meta-analysis¹ and will hereby address each point. The authors underscore the lack of a registered protocol, which we fully acknowledge as a limitation. However, as meta-analyses deal with secondary observational data, the potential pernicious influence of investigator biases might be lessened. The authors maintain selective reporting in systematic reviews is widespread, but cite evidence pertaining primarily to inadequate harm reporting in primary trials included in systematic reviews, rather than in reviews themselves. For these, the evidence for selective outcome reporting and its association with outcomes is mixed.² Nevertheless, for full transparency, we made all extracted data publicly available (https://figshare.com/articles/Data_for_Cristea_et_al_2017_JAMA_Psychiatry/4892111).

作为回复，我们感谢 Faltinsen 及其同事对我们的荟萃分析¹的明智评论，并将在此解决每一点。作者强调缺乏注册协议，我们完全承认这是一个限制。然而，由于荟萃分析处理的是二次观察数据，研究者偏见的潜在有害影响可能会减少。作者认为系统评价中的选择性报告很普遍，但引用的证据主要与系统评价中包括的主要试验中的危害报告不足有关，而不是评价本身。对于这些，选择性结果报告及其与结果的关联的证据好坏参半。²尽管如此，为了完全透明，我们公开了所有提取的数据（https://figshare.com/articles/Data_for_Cristea_et_al_2017_JAMA_Psychiatry/4892111）。