Bias in nutrition-health associations is not eliminated by excluding extreme reporters in empirical or simulation studies

Reviewer #1 (Recommendations for the authors):

Yamamoto et al. analyzed the IDATA data to assess the bias of nutrition intake and the reduction of the bias using the Goldberg cutoffs. They successfully illustrated that the bias reduction is not large in some situations. The data used contain biomarker-based nutrition intake information like the DLW method used as a gold standard in the present study. Although the analyses were precise, the message of the study is quite complicated. A more understandable message would raise the value of the present study.

We appreciate Reviewer #1's positive comments on the precision of our analysis. We also agree that the message of the study may be complicated, and we apologize for any confusion caused by our presentation of the results.

To address this concern, we have revised the manuscript to make the key findings more accessible to a wider audience. Specifically, we have simplified the language used throughout the manuscript and provided more context for our analyses. Additionally, we have emphasized the main takeaways of the study conclusion section of the paper, which we hope will make our message more understandable to readers.

Once again, we appreciate the constructive feedback from Reviewer #1 and believe that our revisions will improve the clarity and accessibility of the manuscript.

1. In the abstract, IDATA should be spelled out. Also, please show the name of "IDATA" and it’s spelling out in the introduction section.

Thank you for this comment. As suggested by the reviewer, we have spelled out IDATA in the abstract and introduced IDATA in the introduction section.

2. There are many figures, although the result section of the manuscript does not contain sufficient numerical information; readers are hard to follow the discussion alongside the table and figure. For example, please specify in the text in what situation the bias is especially large or is reduced by the modification method and how large it is.

We appreciate the reviewer’s comment. We revised the manuscript to provide more clarity in the presentation of results for the empirical study (Page 9, Lines 172-177):

“In the two cases of underestimation from self-report (EI and BW, and EI and WC), there were also significant associations between NI_BIO and the outcomes. In all cases, the nutrition-health outcome pairs with significant associations between NI_BIO and HO had smaller confidence intervals for the percent bias than those pairs without significant associations between NI_BIO and HO (EI and BW, EI and WC, SI and BW, SI and WC, PoI and VO2, PrI and BW, and PrI and WC, shown in italic bold in Figure 2).”

We also revised the manuscript of the Results section for the simulation study (Page 10, Lines 198-202):

“After application of the Goldberg cutoffs, the bias was reduced in 14 among 24 combinations of NI and HO, which can be seen in Figure 3 where the green circles are closer to 0 than the blue triangles. Specifically, ${\displaystyle |{\beta }_{N{I}_G,HO}|<|{\beta }_{N{I}_{\text{SR}},HO}|}$. Notably, the bias was reduced in all six outcomes and reduction was large when EI is used as a predictor. When SI, PoI, and PrI are used as predictors, the bias was not much changed by applying the Goldberg cutoffs (green circles and blue triangles are close to each other in Figure 3).”

Similarly, we emphasized the qualitative assessment of the figures by drawing readers to compare the green circles and blue triangles later in that text section and in the figure legend. We hope that this, combined with our original design of keeping the symbols and colors consistent across main and supplementary figures, will help the reader follow the results.

3. The study concludes that "there is a situation that the Goldberg cutoffs did not solve the problem." However, readers would like to know in what situation the Goldberg cutoffs are not useful and which alternatives are available.

We would like to thank the reviewer for this comment, which allowed us to clarify our discussion. Goldberg cutoffs are useful in reducing and in some cases completely eliminating the bias in mean nutrition intake; however, their performance when considering the association between nutrition intake and health outcome is more nuanced (Page 13, Lines 279-282):

“In conclusion, our study demonstrated that the use of Goldberg cutoffs can effectively reduce or eliminate bias in mean nutrition intake measurements in the IDATA dataset. However, our analysis also revealed that these cutoffs have limited efficacy in addressing bias in the associations between nutrition intake and health outcomes, as both empirical and simulated results showed.”

In addition to the conclusion, we also emphasized these in the introduction section (Page 7, Lines 120-122):

“While reducing bias in the mean nutrition intake is useful for determining the true value of population nutrition intake from self-reported data, it is also important to reduce bias in the associations between nutrition intake and health outcomes.”