LAUSR

The authors reply: We thank Cohen for her letter. There is some inconsistency in the used terminology for risk factors. Investigators with different interests use “environmental” with different connotations or different specificity thresholds. Our umbrella review assessed any nongenetic risk factors, in the broadest possible fashion. This included lifestyle, biobehavioural, sociodemographic, mental health and many other factors, and it also included all the factors that Cohen uses the term “environmental” for in a very strict sense. Therefore, the statements of Cohen that our paper “does not, in any way, shape, or form, address or review environmental risk factors” are incorrect. Contrary to what Cohen asserts, strictly speaking environmental factors and interventions were eligible. However, there was a dearth of such eligible evidence as of 1/2017 when we performed our last search. Our eligibility criteria focused only on obesity, meta‐analyses, and longitudinal studies and excluded systematic reviews without meta‐analyses, meta‐analyses considering body mass index and those not distinguishing obesity from overweight. As Cohen admits “true environmental changes related to obesity have been very limited, and none have yet targeted the factors that are really driving overconsumption and physical inactivity.” We cannot exclude that strict‐terminology “environment” factors and interventions may be important. Interest in them is growing, and we welcome more formal meta‐analyses on these factors focused specifically on obese individuals. More randomized trials of policies that modify the environment are particularly needed. Whether physical environment and policy interventions do work for obese individuals needs to be rigorously tested and reviewed. Some evidence to‐date is not very favourable. Subtle biases and differences in definitions may affect the apparent strength of the evidence, for example, in our umbrella we included a systematic review on school‐based interventions where the evidence was very low. Conversely, another umbrella review of school‐based interventions had more favourable conclusions, but it included also overweight and not just obesity and had several other differences. Our paper was extremely extensively peer reviewed. After three months of peer‐review by a high‐impact general medical journal and comments by 10 people, it was rejected because of priority. We replied to all the comments and sent the paper along with our replies to another high‐impact general medical journal; the editor rejected it within 2 hours from submission, apparently without reading either the paper or the replies, because of priority. We then submitted to another high‐impact general medical journal that asked for revisions and initially seemed favourably inclined to publish it. After another 4 months of review and re‐review by three reviewers and by several editors and after satisfactory responses to the reviewers’ and editors’ comments, the journal decided not to accept it because of priority. The editors also believed that observational studies and intervention studies may need to be presented separately, but we do not sympathize with salami publication. Furthermore, the editors also suggested that the paper would be welcome in a specialty journal of the same publishing group with minor final modifications. We preferred to make minor modifications and publish it in EJCI instead.