How We Label Evidence in Specialty Topics

How we label evidence

Many topics in this section are new, debated or sold with big promises. To keep claims honest, each page names the kind of evidence behind a statement. The labels below are used the same way everywhere in Specialty Topics. They describe the research, not whether something is right for you. Please talk with a clinician or pharmacist before using any product or practice, especially if you take medicines or have a health condition.

The labels

Where we looked and found nothing usable, the page says "Proper evidence not found" and states what we checked.

Why "approved medicine" is a special label

Before a drug is approved, the FDA reviews the full record, from early lab work through large human trials, and decides whether it is safe and effective for its intended use in the population studied. [1] Dietary supplements work differently. The FDA says it does not have the authority to approve dietary supplements before they are marketed, and it generally does not approve supplement claims or labeling before use. [2] So a supplement is never labeled "approved medicine" here. For how that affects product claims, see how supplements are regulated.

Why cell and animal results get their own label

Early research asks whether a substance might cause harm before it is tested in people, using studies in cell cultures (in vitro) and in living animals (in vivo). [3] The FDA states that preclinical research is not a substitute for studies done in people. [4]

Animal results can point the wrong way. A systematic review compared animal experiments with human trials for six treatments. Corticosteroids showed benefit in animal models of head injury but not in clinical trials, and tirilazad improved results in animal stroke models but was associated with worse outcomes in patients. [5] The authors say differences may come from bias or from animal models that do not match the human disease. [5] That is why we do not carry an animal finding over to people.

Why a lab marker is not the same as a health outcome

A biomarker is a measurement such as a blood level. A clinical outcome is something a person feels or experiences, such as pain, a fracture or survival. A change in a marker does not always mean a better outcome. In cancer research, a review of 65 reported links between a stand-in measure and survival found more than half were of low strength. [6] A study of FDA cancer approvals also found that formal checks of that link were often missing. [7] We therefore describe marker changes as marker changes only.

Study size and design

The FDA describes Phase 2 studies as typically a few hundred patients, too few to show whether a drug will be beneficial, while Phase 3 studies use 300 to 3,000 people. [4] In a review of 93 meta-analyses, effects were on average 48% larger in trials with fewer than 50 patients than in trials with 1,000 or more. [8] Good trials also reduce bias by randomly assigning people and, where possible, keeping participants and researchers blind to who gets which treatment. [9] Our labels take size and design into account. See reading early-stage research for how to judge a study yourself.

Where to go next

Start with what counts as emerging wellness science, return to the Emerging Wellness Science overview, or read evaluating safety claims. Pages explain research; they do not give personal medical advice.