Rabassa et al (1) contend that the use of GRADE to adjudicate the “veracity” of claims about diet and health directly to the general public will enhance understanding. We disagree.

The severe limitations of GRADE for application to diet and health have been described (2,3). These are vividly illustrated by its inability to differentiate even a salad of mixed greens from a bowl of sugar; the reported conclusion based on GRADE is that “the quality of evidence is low, so it depends almost entirely on…preferences” (4).

A method for scoring evidence derived for pharmacotherapy trials applies poorly to a behavior that cannot be blinded; for which there is no “placebo;” and for which randomization with high adherence for long periods (years or decades) is often impossible or unethical.

Important contributions of observational epidemiology include studying outcomes of morbidity and mortality and the ability to explore “instead of what.” Only rarely will there be RCTs that recruit and retain tens of thousands of people to eat “lots” vs. “none” of a specific food item until they die or are diagnosed with a debilitating illness. Among the small number of such RCTs, trials would even more rarely be able to examine the “instead of what” options for the “none” group (e.g., instead of “lots” of red meat, more cheese pizza vs. poultry, vs. fish vs. beans and nuts). For hard clinical outcomes and for modeling the many possible “instead of what” options, observational epidemiology is uniquely well suited.

Randomized controlled trials (RCTs) contribute importantly to knowledge about diet and health, but so do in conjunction with other varieties of evidence, including randomized studies with risk factors as outcomes, observational epidemiology, and mechanistic studies. The extreme emphasis on RCTs in GRADE scoring skews nearly all conclusions about diet and nutrition toward uncertainty or, even worse, false negative conclusions based on studies of insufficient duration, adherence, or statistical power to examine vulnerable subgroups such as those with low intake of the dietary factor being investigated (5).

Of note, much knowledge that is of fundamental importance in epidemiology, public health, and medicine has been established without reliance on RCTs, or even by overcoming misleading RCT results. Examples include but are not limited to: the harms of tobacco; the hazards of occupational chemical exposures; the harms of environmental contaminants; the health impacts of climate change; and long term side effects of pharmaceuticals.

Further, there is much humans know reliably solely on the basis of observation and pattern consistency, including such self-evident propositions as the value of looking both ways before crossing a busy street, or the advantages of water over gasoline for dousing a campfire. These are admittedly homely examples, but they illustrate the liabilities of excessive reliance on RCTs as the arbiter of “veracity.” Alternatives to GRADE have been developed expressly to capture the important contributions to understanding made by diverse sources of evidence (2,3).

The Nutrimedia platform seems to represent marketing collateral for GRADE, and is far more likely to undermine public understanding of nutrition and health than to advance it.

Sincerely,

David L. Katz, MD, MPH
True Health Initiative

Walter Willett, MD, DrPH
Professor of Epidemiology and Nutrition, Harvard T.H. Chan School of Public Health

Christopher Gardner, PhD
Nutrition Scientist & Professor of Medicine, Stanford University


References
1) Rabassa M, Alonso-Coello P, Casino G. Nutrimedia: A novel web-based resource for the general public that evaluates the veracity of nutrition claims using the GRADE approach. PLoS One. 2020;15(4):e0232393. Published 2020 Apr 30. doi:10.1371/journal.pone.0232393.
2) Qian F, Riddle MC, Wylie-Rosett J, Hu FB. Red and Processed Meats and Health Risks: How Strong Is the Evidence? Diabetes Care. 2020;43(2):265-271. doi:10.2337/dci19-0063.
3) Katz DL, Karlsen MC, Chung M, et al. Hierarchies of evidence applied to lifestyle Medicine (HEALM): introduction of a strength-of-evidence approach based on a methodological systematic review. BMC Med Res Methodol. 2019;19(1):178. Published 2019 Aug 20. doi:10.1186/s12874-019-0811-z.
4) Hamblin J. The Actual Reason Meat Is Not Healthy Nutrition studies leave out a crucial factor. The Atlantic. October 10, 2019. https://www.theatlantic.c...
5) Morris MC, Tangney CC. A Potential Design Flaw of Randomized Trials of Vitamin Supplements. JAMA. 2011;305(13):1348–1349. doi:10.1001/jama.2011.383).