(OR30-01-22) Rigor, Reproducibility, and Transparency in Nutrition Research Using R Dashboards

Objectives: To provide all analyses conducted, not just abbreviated analysis used for the main findings, authors currently submit supplemental information as PDF files. For example, if outliers are removed to arrive at the final results, supplemental materials may contain the analysis for all the data so readers can evaluate the findings with the outliers include. Extensive supplemental information that includes extra analyses, code, and model formulations may be challenging for readers to navigate. This method of dissemination requires authors to balance rigor with reproducibility and transparency. We propose a new framework for publishing supplemental information using interactive dashboards for hosting supplemental information which increases transparency and reproducibility in nutrition research.

Methods: A recent article on the analysis of a weight bias survey was used to develop an example dashboard that would hold all supplemental information. The statistical programing software R and the flexdashboard package were used to develop a dashboard that intuitively displays supplemental information and allows the reader to interactively explore the data and results of published work.

Results: Dashboards allow for full transparency by providing the reader a user-friendly method for exploring supplemental information. Additionally, dashboards allow the ability to include downloadable data, interactive tables as well as embedding all source code used in the analysis. The dashboard goes beyond traditional supplemental files and encourages any reader to interact with the results and data providing clarity and transparency of findings.

Conclusions: Interactive dashboards for publishing supplemental information provide accessible, transparent, and reproducible research in nutrition, leading to greater public confidence in published results.

Funding Sources: None