How to Read Nutrition Research Headlines
A plain guide to reading food and nutrition headlines: study type, what was measured, what was compared, relative versus absolute numbers, and how news can exaggerate a study.
Content last updated . This page has not been medically reviewed.
Safety first. Supplements can cause side effects and interact with medicines. Read mushroom safety and interactions before you use any mushroom product.
The question
A headline says a food "cuts risk" or "linked to" something. What should you check before you believe it? This page gives questions to ask and uses real examples from the sources we read. It is information, not advice about what to eat, and no single study settles a question.What kind of study was it?
Nutrition research mostly uses two designs.- Randomized trials assign people to eat different things and compare them. The kidney-disease nutrition primer we read says randomized trials provide the strongest evidence of causality, but the expense and difficulty of sustaining adherence to dietary interventions are substantial barriers. [1]
- Cohort (observational) studies follow people who eat as they choose. Dietary intakes have traditionally been self-reported and are subject to bias. [1] A review of nutritional epidemiology lists substantial measurement error, confounding, the variable effects of food items, variable reference groups, interactions and multiple testing as problems, and notes that the results of many such studies have not been replicated in subsequent studies. [2]
"Linked to" usually means an association from a cohort study. It does not show that the food caused the outcome.
Do trials and cohort studies agree?
Often, on average, yes. A 2021 BMJ meta-epidemiological study of 97 matched pairs found the difference between trial and cohort estimates was small on average. [3] A 2024 replication of 82 pairs found similar average estimates. [4] But the BMJ authors also say wide prediction intervals mean important differences in individual comparisons cannot be excluded, and that differences were mainly driven by dissimilarities in population, intervention or exposure, comparator and outcome. [3] So the question is whether the study compared what the headline implies.A food example: observational studies suggested the MIND diet might protect against cognitive decline, but a 3-year randomized trial in older adults found changes in cognition and brain MRI did not differ significantly from a control diet. [5]
Questions to ask of any headline
| Ask | Why it matters | Example from the sources |
|---|---|---|
| Humans, animals or cells? | Animal and cell findings are preclinical. Keep them apart from human trials. | A BMJ study found that 36% of UK university press releases on health research contained exaggerated inference to humans from animal research. [6] |
| Which food, exactly? | Whole food, juice, powder and extract are different. | A review of foods called superfoods looked at controlled trials of the foods and found limited evidence overall. [7] See Superfoods. |
| Compared with what? | "Instead of" a worse food is not "in addition to" the usual diet. | The PREDIMED comparison was a Mediterranean diet with olive oil or nuts against a low-fat diet. [8] |
| What was measured? | A blood marker is not a heart attack. | The first DASH trial measured blood pressure over 8 weeks in fed participants. [9] |
| Relative or absolute? | A big-sounding relative change can be a small absolute one. | PREDIMED strokes: hazard ratio 0.60, which the Cochrane review describes as a decrease from 24 per 1000 to 14 per 1000. [8] |
| Has it been corrected? | Trials can be retracted or re-analysed. | PREDIMED was retracted and re-analysed after concerns about randomisation at two of 11 sites. [8] |
| Where did the headline come from? | News often repeats the press release. | Exaggeration in news is strongly associated with exaggeration in press releases. [6] |
Where the headline comes from
The BMJ press-release study examined 462 press releases from 20 UK universities in 2011, covering biomedical and health science broadly, not only food. 33% contained exaggerated causal claims and 40% contained exaggerated advice. [6] It found that when the press release was exaggerated, news stories were far more likely to be. [6] That sample is from 2011 and from the UK, so treat it as an example of the pattern, not a measure of today's coverage.Frequently asked questions
Does "linked to" mean the food caused it?
No. A link usually comes from an observational study, where people chose their own diets, and confounding is one of the known problems. [2]
Are randomized trials of foods always better?
They are the strongest design for causation, but adherence is hard to sustain over long periods. [1] A short trial may measure a marker, not a disease outcome. [9]
Why do food studies seem to contradict each other?
Studies can differ in population, food form, comparator and outcome. The BMJ analysis found most disagreement came from such dissimilarities. [3] Many nutritional epidemiology findings have also not been replicated. [2]
What is the difference between relative and absolute risk?
Relative risk compares groups as a ratio. Absolute risk gives the actual numbers. In the Cochrane review, a stroke hazard ratio of 0.60 corresponded to 24 per 1000 falling to 14 per 1000. [8]
Can I trust a press release about a food study?
Check the paper itself. In one UK study, exaggerated press releases were strongly linked with exaggerated news. [6]
More details: evidence, limits and cautions
The details behind this page: cautions, limits of the evidence and where sources disagree. Open any note to read it.
What regulators allow on labels
On U.S. food labels, a health claim describes a relationship between a food substance and reduced risk of a disease or health-related condition, and the FDA reviews health claims in advance. Dietary guidance statements and structure/function claims are not subject to premarket review and authorization. [10] See Reading Food Labels.
Limits and uncertainty
Even a good study answers one narrow question about one group. Diet and cancer research, for example, faces exposure misclassification, confounding and difficulty in assessing long-term and changing dietary habits. [11] For how the Research Library labels study designs, see Evidence Types and Study Designs and Human, Animal and Laboratory Research.
Who should ask a clinician
Anyone about to change eating habits because of a headline, especially people on warfarin, people with kidney disease, people with food allergies, and people who are pregnant. Ask the prescriber or pharmacist before changing a medicine. More on Observational Studies vs Trials in Nutrition and Functional Food Research and Evidence.
References
- Sullivan VK, Rebholz CM. Nutritional Epidemiology and Dietary Assessment for Patients With Kidney Disease: A Primer. Am J Kidney Dis 2023. PMID 36610612. Read 2026-10-22 (abstract).
- Kamangar F, Karimi P. The state of nutritional epidemiology: why we are still unsure of what we should eat? Arch Iran Med 2013. PMID 23906255. Read 2026-10-22 (abstract).
- Schwingshackl L et al. Evaluating agreement between bodies of evidence from randomised controlled trials and cohort studies in nutrition research: meta-epidemiological study. BMJ 2021. PMID 34526355. Read 2026-10-22 (abstract).
- Stadelmaier J et al. Evaluating agreement between evidence from randomised controlled trials and cohort studies in nutrition: a meta-research replication study. Eur J Epidemiol 2024. PMID 38177572. Read 2026-10-22 (abstract).
- Barnes LL et al. Trial of the MIND Diet for Prevention of Cognitive Decline in Older Persons. N Engl J Med 2023. PMID 37466280. Read 2026-10-22 (abstract).
- Sumner P et al. The association between exaggeration in health related science news and academic press releases: retrospective observational study. BMJ 2014. PMID 25498121. Read 2026-10-22 (abstract).
- Effects of superfoods on risk factors of metabolic syndrome: a systematic review of human intervention trials. Food Funct 2018. PMID 29557436. Read 2026-10-22 (abstract).
- Rees K et al. Mediterranean-style diet for the primary and secondary prevention of cardiovascular disease. Cochrane Database Syst Rev 2019. PMID 30864165. Read 2026-10-22 (abstract).
- DASH Collaborative Research Group. A clinical trial of the effects of dietary patterns on blood pressure. N Engl J Med 1997. PMID 9099655. Read 2026-10-22 (abstract).
- U.S. Food and Drug Administration. Label Claims for Conventional Foods and Dietary Supplements. Page dated 03/28/2024. Read 2026-10-22.
- Challenges and promises of nutritional epidemiology to investigate cancer etiology. J Natl Cancer Inst Monogr 2026. PMID 42008722. Read 2026-10-22 (abstract).
Education, not medical advice. This page is for general information. It does not diagnose or treat any condition and is not a substitute for care from a qualified health professional. Natural does not automatically mean safe. Ask a clinician or pharmacist before changing your care, especially if you take medication, are pregnant or breastfeeding, or have a health condition.