Poor statistical reporting: do we have a reason for concern?

"Statistical reviews in journals of the World Association of Medical Editors"

Introduction: In recent years, there has been a decline in the quality of statistical reporting. Objectives: The aim of this survey was to find out the opinions of members of the World Association of Medical Editors (WAME) on the statistical reviews conducted in their journals and the related recommendations that should be implemented. Methods: A questionnaire containing 25 questions on a range of statistical aspects was distributed to WAME members and their journals. Results: The survey was completed by 141 respondents, the largest proportion of whom were editors-in-chief (36.9%). According to 40 per cent of them, only 31–50 per cent of accepted manuscripts are statistically correct. The higher their assessment of their own statistical knowledge, the lower they believe this percentage to be (P = 0.02). The frequency of statistical peer review was estimated by most respondents to be only 1–10%. These mainly include the difficulty in finding people with the right skills and the lack of funding in this area. Among respondents without a statistical editor on the editorial board, 49% believe that statistical reviews enhance the quality of published manuscripts, whereas those confirming such a presence evaluated the percentage as high as 84% (P <0.001). Only 5% of respondents said that their journal uses SAMPL (Statistical Analyses and Methods in the Published Literature) recommendations. Conclusions: Nowadays, members of editorial boards face significant problems in conducting reviews in their journals. For this reason, it is imperative to start implementing statistical peer review for biomedical journals.