Brown rice swaps: a cooked-weight model, not a feeding trial
Three US cohorts linked modeled rice substitutions with diabetes risk, without assigning anyone a rice diet.

Illustrative food image, not a photograph of the study intervention.
Three cohorts, repeatedly updated
The researchers followed diet and health in the Health Professionals Follow-up Study and the two Nurses' Health Studies. Together the cohorts included 39,765 men and 157,463 women. Diet, lifestyle practices and disease status were prospectively assessed and updated, allowing the authors to examine rice intake before subsequent diabetes outcomes. That design is different from assigning participants to eat brown or white rice. It can identify associations across long-running cohorts, but it cannot by itself establish that the rice choice caused the difference. The study addressed type 2 diabetes risk rather than glucose treatment in people already living with diabetes.
Sun Q, Spiegelman D, van Dam RM, Holmes MD et al.. Arch Intern Med, 2010. PMID 20548009 · DOI
The intake categories were different
For white rice, the comparison was at least five servings weekly versus fewer than one serving monthly. The adjusted pooled relative risk was 1.17, with a 95% confidence interval of 1.02 to 1.36. For brown rice, the high-intake category began at two servings weekly, again compared with fewer than one monthly. Its relative risk was 0.89, with a confidence interval of 0.81 to 0.97. Those category thresholds are not interchangeable. The results compare people in different intake groups; they do not mean that five white-rice servings and two brown-rice servings were directly tested against one another in a controlled experiment.
Sun Q, Spiegelman D, van Dam RM, Holmes MD et al.. Arch Intern Med, 2010. PMID 20548009 · DOI
What the substitution calculation meant
The modeled substitution replaced 50 grams per day of cooked white rice with the same cooked weight of brown rice. The abstract describes that quantity as about one-third of a daily serving. This substitution was associated with 16% lower type 2 diabetes risk, with a 95% confidence interval of 9% to 21% lower risk. These are relative estimates from a statistical model, not the proportion of participants who avoided diabetes after an assigned dietary change. Keeping both the cooked-weight denominator and the modeled nature of the comparison visible prevents a practical-looking food quantity from being mistaken for a proven prescription.
Sun Q, Spiegelman D, van Dam RM, Holmes MD et al.. Arch Intern Med, 2010. PMID 20548009 · DOI
Rice was not the only modeled replacement
Replacing the same amount of white rice with whole grains as a group was associated with 36% lower diabetes risk, with a confidence interval of 30% to 42% in the corrected abstract. That estimate is not identical to the brown-rice estimate and should not be presented as though brown rice alone produced it. The analyses adjusted for age and other dietary and lifestyle risk factors, but adjustment does not turn a cohort into a randomized trial. The useful conclusion is specific: these cohorts supported an association between replacing refined grain intake and lower diabetes risk, without proving an individual prevention effect.
Sun Q, Spiegelman D, van Dam RM, Holmes MD et al.. Arch Intern Med, 2010. PMID 20548009 · DOI