Gaming the peer review system: Evidence for a review mill in medicine highlights the need to ensure reviewer integrity.

BACKGROUND: Review mills are recognized when individuals generate numerous generic review reports, typically containing suggestions for citations to their own work. Here, we report a network with characteristics of a review mill in the field of gynecological oncology. METHODS: Our search started with a review that contained "boilerplate" comments as well as suggestions that specific PubMed IDs be cited. We searched the internet using Google for review reports using the same boilerplate comments. We coded text to quantify similarities between reviews and compiled citations suggested by reviewers. For comparison, we analyzed 59 reviews of the same articles by other peer reviewers. RESULTS: We identified a network of 195 review reports that shared boilerplate text from 170 articles. One hundred and eighty-six reports suggested citing articles coauthored by a member of the network. Five members of the network had editorial roles. Authors of 142 articles complied with suggestions for citation. Boilerplate text and citation recommendations were rare in the comparison reports. CONCLUSIONS: Review mills lead to articles being published without proper peer review. This is of particular concern in medical research. Open peer review and transparent reporting of the editors responsible for handling papers will make it easier to detect review mills.