Statistical reporting and table construction guidelines for publication in the Canadian Journal of Anesthesia

Jones PM, Bryson GL, Backman SB, Mehta S, Grocott HP.
Can J Anaesth. 2017 Dec 5. [Epub ahead of print]

When publishing scientific information, data must be presented accurately, consistently, and transparently for maximally effective communication. At the Canadian Journal of Anesthesia, we commonly see problems in the reporting of results in the text and tables of manuscripts submitted for publication. Some of the common problems we encounter include:

  • Effect sizes, such as relative risks (for binary variables) or risk differences (for binary or continuous variables) are not presented.
  • The uncertainty of effect sizes is not presented with a 95% confidence interval (CI).
  • Information (e.g., weight or height) is presented in either an overly precise (and clinically irrelevant) manner, by using two (or more) decimal places, or a falsely precise manner, by using decimal places that go beyond the precision of the initial measurement.
  • P values are reported using too many decimal places.
  • P values are reported in an imprecise manner (P < 0.05 or P = NS).
  • The specific statistical test(s) that pertain to the primary (and secondary) outcome are not reported.

To address these issues, we herein provide some general guidance for statistical reporting in manuscripts submitted to the Canadian Journal of Anesthesia. 

(Jones, 2017)

The authors have published this article as open access. Read the full article here