Nut frequency and survival: a strong association without random assignment
Two health-professional cohorts linked more frequent nut intake with lower mortality.

Illustrative food image, not a photograph of the study intervention.
Long follow-up with repeated dietary reports
Researchers examined nut intake in 76,464 women from the Nurses’ Health Study and 42,498 men from the Health Professionals Follow-up Study. People with a history of cancer, heart disease, or stroke were excluded at baseline. Nut consumption was assessed initially and updated every two to four years. That repeated assessment distinguishes the analysis from one based only on a single dietary question. It remains observational, however: participants chose their diets. The investigators followed existing eating patterns rather than assigning a daily nut portion and comparing randomized clinical outcomes.
Bao Y, Han J, Hu FB, Giovannucci EL, et al.. The New England journal of medicine, 2013. PMID 24256379 · DOI
Frequency tracked with lower total mortality
Across more than three million person-years, 16,200 women and 11,229 men died. Compared with people reporting no nuts, the adjusted hazard ratio for death was 0.93 for intake less than once weekly and 0.80 for seven or more times weekly. Intermediate categories also showed inverse associations. For the highest frequency category, the 95% confidence interval was 0.73 to 0.86. These estimates describe relative mortality hazards during follow-up. They do not mean that an individual gains a specified number of years by adding nuts to an otherwise unchanged diet.
Bao Y, Han J, Hu FB, Giovannucci EL, et al.. The New England journal of medicine, 2013. PMID 24256379 · DOI
Cause-specific findings require their own labels
The abstract reports significant inverse associations with deaths attributed to cancer, heart disease, and respiratory disease. These endpoints are deaths from specified causes, not necessarily the first occurrence of each condition. Nor do the results isolate one nut variety as the explanation. The dietary exposure was frequency of nut consumption, and the analysis pooled evidence across two health-professional cohorts. It is reasonable to report consistency across total and selected cause-specific mortality, but not to translate that consistency into proof that walnuts, almonds, or one nut-containing recipe independently prevented those diseases.
Bao Y, Han J, Hu FB, Giovannucci EL, et al.. The New England journal of medicine, 2013. PMID 24256379 · DOI
Adjustment does not make the menu randomized
The investigators adjusted for other known or suspected risk factors, and the inverse association persisted. Adjustment is important, but it does not change the fact that nut intake was self-selected. The paper therefore supports an association, not a guaranteed survival benefit from a prescribed serving. Its population was nurses and other health professionals without the specified major diseases at baseline. The linked walnut-containing meal is a relevant food example, not the trial intervention, because there was no assigned menu. Preserve that distinction when moving from a compelling cohort estimate to everyday food discussion.
Bao Y, Han J, Hu FB, Giovannucci EL, et al.. The New England journal of medicine, 2013. PMID 24256379 · DOI