The Peer Review Process at Annals
Most readers highly value the fact that articles in a journal like ours have undergone formal peer review. Many readers also have a relatively simple understanding of that term as describing a single well-defined process of review by expert reviewers, but it is a lot more complicated and nuanced than that. We therefore provide a very brief summary of our procedures to provide appropriate levels of review for most (but not entirely all) the journal content.
Although we try to be a model among journals for the rigor of our peer review process, like most of them (including the most prestigious) this does not mean that all content is peer reviewed in the same way. All original content (particularly research content) in the journal is peer reviewed by one of the many experts on our editorial board, but additional peer review of every submission by members of our reviewer panel is not always necessary or appropriate. Many submissions are not appropriate for the journal for one fairly obvious reason or another (eg, target audience), so like most other journals we reject many manuscripts after review by an editor. For those which are not obviously inappropriate, however, we receive far more submissions than we can publish, so our further process seeks to identify the best of the best.
The vast majority of scientific content that we publish is critically reviewed first by members of our editorial board with specific expertise, and then gets additional scrutiny from our expert reviewers. Our most stringent level of review is reserved for original research, which will form the basis of the scientific record in the future. These submissions are reviewed by at least two of our expert reviewers who are blinded to the identity of the authors. Quite a few papers are reviewed more than once, and sometimes in particularly complex cases 5 or 6 reviewers and editors may be involved, including deputy editors. During this process there is much consultation and discussion between editors, reviewers, and authors and recommendations are made to the authors. Sometimes that discussion exceeds the length of the original paper itself, and it certainly is a laborious and time-consuming process. Editors and reviewers must disclose potential conflicts of interest which are managed as per a rigorous policy [EMBED LINK]. Virtually no original research is accepted with no revisions whatsoever, and our authors strongly agree that in general the process improves the quality of the final manuscript. Once it has been discussed, revised, and received the final stamp of approval from the supervising editor (whose name is always published with the manuscript for transparency), all original science content in the journal undergoes a final review by the editor in chief before acceptance.
None of this means the final article is irrefutable truth; such a thing does not exist in science where our state of knowledge is (we hope) constantly evolving and no study should be judged in isolation. But it does mean that we’ve asked all the appropriate questions we could think of, made suggestions, and required revisions to make the paper as complete and transparent to replication as possible.
This process for original research is the most rigorous and is probably what most readers think of as “formal peer review,”, but the journal contains much other content of a factual and scientific nature which does not lend itself to this approach. For example, we have a number of regular journal features (like News & Perspective, CDC Update, NHTSA Notes, etc) that are updates written by selected topic experts on a routine basis. These are also reviewed by an editor but not sent out for additional review. A very few items, such as ACEP Clinical Policies, are published verbatim from the experts that develop them and are not revised (for obvious reasons); this fact is published along with each.
There are always some exceptions to the above processes as we develop new types of content or relatively unique contributions occur. We try to describe the particular variants of peer review that were used for each of these, or if there was none (as for example in the EM:RAP commentaries), that is made clear as well. Our goal is to provide as much oversight as is needed and logistically practical, and to enable readers to determine what that level of oversight was as conveniently as possible.