How We Grade Medical Evidence
How we grade a claim
We grade each health claim as established, emerging, experimental or evidence against, based on the kind of study behind the claim, the size and consistency of the results, who was studied, and who funded the research. Evidence grades describe how strong the research is for one specific claim, not whether a treatment is right for you.
The four evidence grades
| Label | What it means |
|---|---|
| Established | Consistent results from two or more adequate randomized trials, a systematic review, or one large blinded trial confirmed by a review, in the people studied |
| Emerging | Some human trial evidence, but small, short, single, unblinded, mixed, maker-funded or based on surrogate markers |
| Experimental | Rests only on mechanism, animal, lab, case-report, case-series or blood-level data; no controlled human trial of the claim among the studies relied on |
| Evidence against | Human trials tested the claim and did not find the effect |
An established grade means consistent results in the people who were studied, from two or more adequately sized human randomized controlled trials, a systematic review of randomized trials, or one large blinded placebo-controlled trial whose result a systematic review confirms. An emerging grade means some controlled human evidence exists, but the trials are small, short, single, unblinded, mixed or mostly funded by the product's maker, or rely on surrogate markers. An experimental grade means the claim, as stated, rests only on mechanism studies, animal studies, lab studies, case reports, case series or measurements of blood levels, with no controlled human trial of the claim among the studies the evidence file relies on. An evidence-against grade means controlled human trials tested the claim as their main question and did not find the effect, and none of the controlled trials the evidence file relies on found it.
When a grade applies only to a guideline's wording, a drug label's wording, one population, one blood measurement or one measurement method, the grade always states that limit. Each grade in our evidence files comes with a written reason, such as the number of trials, their size, or a limit on who was studied. One claim about a treatment can carry different grades for different uses, forms or populations.
Examples from our evidence files
Emerging, for example: Studies of DEXA's visceral fat estimate found close agreement with CT scans, but our DEXA evidence file grades the finding emerging because the validation studies were funded by, or had authors employed by, the scanner makers.[1][2]
Experimental, for example: In a 1996 open-label study with no comparison group, 9 people with early, untreated Parkinson's disease received 600 mg of IV glutathione twice a day for 30 days, and the authors reported a 42% decline in disability, but a result with no comparison group cannot show that glutathione caused the change, so our IV glutathione evidence file grades the claim that IV glutathione slows Parkinson's progression experimental.[3] The same evidence file grades symptom improvement separately: in a 2009 randomized, placebo-controlled pilot of 21 people with Parkinson's disease given 1,400 mg of IV glutathione three times a week for 4 weeks, symptom scores did not differ significantly from placebo, and the authors wrote that a mild symptomatic effect remained possible and needed a larger study.[4] We read only the abstracts of these two studies. The PubMed record for the 1996 study lists no funding information, and the record for the 2009 trial lists non-US government research support without naming the funder. We did not find an FDA-approved glutathione drug product given by IV or injection (checked September 27, 2026).
Evidence grades versus regulatory status
Regulatory labels such as FDA-approved indication, off-label use and not approved come from a separate regulatory status record for each treatment, not from the evidence grade. Our editorial rules require pages to tell apart established evidence, emerging evidence, experimental use, off-label use, FDA-approved indications and treatments that are not approved.
Read more:
- What Does FDA-Approved Mean?
- Off-Label Prescribing Explained
- Compounded Medications Explained
- FDA Drug Approval Process Explained
Study types we look at
| Study type | How our evidence files use it |
|---|---|
| Systematic review or meta-analysis | Can support an established grade when results are consistent |
| Randomized controlled trial | Several consistent trials can support established; one small or single trial is usually emerging |
| Observational study | Described as an association, not proof of cause |
| Animal study | Never used to support a human lifespan claim |
| Lab (in-vitro) or mechanism study | Explains how something might work; not evidence of an effect in people |
Our editorial rules prefer systematic reviews, randomized controlled trials, clinical guidelines and US government health sources such as the FDA, NIH, CDC and National Library of Medicine. Our editorial rules treat a human randomized trial, an observational study, an animal study, a lab (in-vitro) study and a mechanism study as different levels of evidence that are never interchangeable. For each claim, our evidence files record the source and an evidence grade, and most also record the study type, who was studied, what was found and the study's limits.
Observational studies alone can earn at most an emerging grade, stated as an association and never as proof of cause, and they cannot support a claim that a treatment causes a benefit, which stays experimental. A possible harm seen in observational or regulator data is reported with its caveats and graded emerging, rather than left out. Our evidence files never use animal lifespan results to support a claim about human lifespan. Mechanism studies, such as the finding that cells use NAD+ in energy metabolism, explain how something might work but are not evidence that taking a product changes anything in people.
Our evidence files note when the people studied differ from a typical reader, for example when a trial enrolled only people with heart disease or only people with diabetes.
Examples from our evidence files
Observational, for example: A US observational study of 9,471 adults aged 20 and older, none with a major chronic disease at the start, found that the highest quarter of DEXA-measured body fat percentage was associated with higher death rates over an average of 8.8 years of follow-up, and our DEXA evidence file treats the finding as an association, which does not show whether getting a scan improves health. The PubMed record lists National Institutes of Health grants for the study.[5]
Animal, for example: In a mouse study run by the National Institute on Aging's Interventions Testing Program, nicotinamide riboside did not significantly increase lifespan at the doses tested, and animal results like that cannot support or rule out a human longevity claim.[6] We did not find an FDA-approved nicotinamide riboside drug product for any use, including anti-aging or longevity (checked September 27, 2026).
Biomarkers versus outcomes
| Term | What it means |
|---|---|
| Biomarker | A measurement taken from the body, such as a blood level |
| Outcome | What happens to the person, such as a fracture, a heart attack or death |
| Surrogate marker | A biomarker used in place of an outcome, such as an epigenetic clock |
In our evidence files, a biomarker is a measurement taken from the body, such as a level in the blood, while an outcome is what happens to the person, such as a fracture, a heart attack or death. Our evidence files treat a change in a biomarker as evidence about that biomarker only, not as evidence that a treatment improves health or extends life. In weight-loss trials, body weight is itself the measured outcome, and our evidence files do not treat weight loss as proof of longer survival. US osteoporosis screening guidance rests on trials that measured fractures, an outcome, rather than only bone density readings.
Examples from our evidence files
Surrogate marker, for example: Epigenetic clock results, sometimes described as biological age, are surrogate markers in our evidence files, not aging outcomes.
Biomarker, for example: Oral nicotinamide riboside consistently raised NAD+ levels in blood in clinical trials, which our NAD+ evidence file grades established for the blood measurement only, not as evidence of a health benefit.[7][8][9] We did not find an FDA-approved nicotinamide riboside drug product for any use (checked September 27, 2026).
Outcome, for example: Hip fractures in osteoporosis screening trials.[10]
Funding and conflicts of interest
Our evidence files record who funded a study and whether authors work for the product's maker, whenever the source states it. When a study's funding statement could not be read, our evidence files say so instead of assuming there was no industry funding. When all or most of the evidence for a claim comes from studies funded by the product's maker or written by its employees, the claim can be graded no higher than emerging, unless independent reviews, not funded by the maker, pooled the same result in the same group of people and found it consistent.
When the evidence is thin, unclear or unread
When only a study's abstract could be read, our evidence files mark the source as abstract only. When a source could not be opened, our evidence files list the source as not read, never as evidence that the study does not exist. When our searches find no study for a claim, our evidence files record the result as no support found, which is a search result and not proof that no study exists.
What we don't say
Our editorial rules forbid claims that a treatment reverses aging, adds a set number of years to your life, repairs your DNA, cures a disease or guarantees results. Each of our evidence files includes a list of claims we cannot support, and our pages do not make those claims.
Who does this grading
Evidence grading is research and not a medical review, and the grading is not done by a clinician.
References
- Kaul S, Rothney MP, Peters DM, et al. Dual-energy X-ray absorptiometry for quantification of visceral fat. Obesity. 2012;20:1313-1318. doi:10.1038/oby.2011.393. PMID 22282048. Accessed 2026-09-27.
- Micklesfield LK, Goedecke JH, Punyanitya M, Wilson KE, Kelly TL. Dual-energy X-ray performs as well as clinical computed tomography for the measurement of visceral fat. Obesity. 2012;20:1109-1114. doi:10.1038/oby.2011.367. PMID 22240726. Accessed 2026-09-27.
- Sechi G, Deledda MG, Bua G, Satta WM, Deiana GA, Pes GM, Rosati G. Reduced intravenous glutathione in the treatment of early Parkinson's disease. Prog Neuropsychopharmacol Biol Psychiatry. 1996;20(7):1159-1170. PMID 8938817. doi:10.1016/s0278-5846(96)00103-0. Accessed 2026-10-01.
- Hauser RA, Lyons KE, McClain T, Carter S, Perlmutter D. Randomized, double-blind, pilot evaluation of intravenous glutathione in Parkinson's disease. Mov Disord. 2009;24(7):979-983. PMID 19230029. doi:10.1002/mds.22401. Accessed 2026-10-01.
- Zong G, Zhang Z, Yang Q, Wu H, Hu FB, Sun Q. Total and regional adiposity measured by dual-energy X-ray absorptiometry and mortality in NHANES 1999-2006. Obesity. 2016;24(11):2414-2421. doi:10.1002/oby.21659. PMID 27667735. Accessed 2026-09-27.
- Harrison DE, Strong R, Miller RA, et al. 17-a-estradiol late in life extends lifespan in aging UM-HET3 male mice; nicotinamide riboside and three other drugs do not affect lifespan in either sex. Aging Cell. 2021;20(5):e13328. doi:10.1111/acel.13328. PMID 33788371. Accessed 2026-09-27.
- Trammell SAJ, Schmidt MS, Weidemann BJ, et al. Nicotinamide riboside is uniquely and orally bioavailable in mice and humans. Nat Commun. 2016;7:12948. doi:10.1038/ncomms12948. PMID 27721479. Accessed 2026-09-27.
- Airhart SE, Shireman LM, Risler LJ, et al. An open-label, non-randomized study of the pharmacokinetics of the nutritional supplement nicotinamide riboside (NR) and its effects on blood NAD+ levels in healthy volunteers. PLoS One. 2017. doi:10.1371/journal.pone.0186459. PMID 29211728. Accessed 2026-09-27.
- Martens CR, Denman BA, Mazzo MR, et al. Chronic nicotinamide riboside supplementation is well-tolerated and elevates NAD+ in healthy middle-aged and older adults. Nat Commun. 2018;9:1286. doi:10.1038/s41467-018-03421-7. PMID 29599478. Accessed 2026-09-27.
- US Preventive Services Task Force. Osteoporosis to Prevent Fractures: Screening. Final recommendation, January 14, 2025. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/osteoporosis-screening. Accessed 2026-09-27.