Response to Letter to the editor from Dr. Kawada regarding the article Dietary glycaemic load and cognitive performance in elderly subjects

Abstract

In his offered opinion piece, (Dietary glycaemic load and cognitive performance in elderly subjects) Dr. Kawada comments upon the statistical analysis and suggests that the conclusions of the study should be interpreted with caution. Having closely examined these comments, we believe that they are over-stated and we draw different conclusions. At first viewing, the statistical arguments put forward by Dr. Kawada look complicated, but one may summarize that he believes the analysis lacked statistical power. This argument is directed towards two sets of regression analyses, a Poisson analysis on which one of the messages of the paper hinges, and a second logistic analysis that was acknowledged as statistically underpowered in our publication. No statistical argument is provided as to why the Poisson regression model is underpowered; the critique contains no new scientific content but relies on a technical re-iteration of the limitations of the study (that were highlighted in the original manuscript) combined with quasi philosophical arguments on data set size and the need for biochemical markers in observational dietary studies.