A Meta-Epidemiological Study of Positive Results in Clinical Nutrition Research: The Good, the Bad and the Ugly of Statistically Significant Findings

Abstract

Positive (statistically significant) findings are easily produced in nutrition research when specific aspects of the research design and analysis are not accounted for. To address this issue, recently, a pledge was made to reform nutrition research and improve scientific trust on the science, encompass research transparency and achieve reproducibility. The aim of the present meta-epidemiological study was to evaluate the statistical significance status of research items published in three academic journals, all with a focus on clinical nutrition science and assessing certain methodological/transparency issues. All research items were published between the years 2015 and 2019. Study design, primary and secondary findings, sample size and age group, funding sources, positivist findings, the existence of a published research protocol and the adjustment of nutrients/dietary indexes to the energy intake (EI) of participants, were extracted for each study. Out of 2127 studies in total, those with positive findings consisted of the majority, in all three journals. Most studies had a published research protocol, however, this was mainly due to the randomized controlled trials and not to the evidence-synthesis studies. No differences were found in the distribution of positive findings according to the existence/inexistence of a published research protocol. In the pooled sample of studies, positive findings differed according to study design and more significant findings were reported by researchers failing to report any funding source. The majority of items published in the three journals (65.9%) failed to account for the EI of participants. The present results indicate that there is still room for the improvement of nutrition research in terms of design, analyses and reporting.

Misleading presentations in functional food trials led by contract research organizations were frequently observed in Japan: meta-epidemiological study

“…This aligns with prior studies indicating widespread selective outcome reporting in Japanese FFCrelated RCTs [14,15]. Metaepidemiological studies conducted in various countries have highlighted the problem of selective outcome reporting in nutritional research [13,78,79]. Hence, nutritional researchers should strive to enhance the quality of their studies and reporting.…”

“…Despite the call for rigorous evidence, nutritional research has historically faced significant skepticism [11]. Previous metaepidemiological studies conducted outside Japan have indicated a potential bias owing to the use of sponsored products in nutritional trials, with the reported research quality being low [12,13]. A previous metaepidemiological study highlighted discrepancies between trial protocols and articles related to FFC [14].…”

Variations in the results of nutritional epidemiology studies due to analytic flexibility: Application of specification curve analysis to red meat and all-cause mortality

“…The application of specification curve analysis to nutritional epidemiology: red meat and all-cause mortality A large body of evidence shows inconsistency in the results of nutritional studies, some of which may be explained by analytic flexibility (3,8,20,21). Such inconsistencies have eroded trust in nutritional epidemiology and subjected the field to criticism (22,23).…”