Hierarchy of sources
Not all sources carry the same weight. Where possible, we rely on the following hierarchy, from strongest to weakest:
| Level | Source type | Typical use |
|---|---|---|
| 1 | Clinical practice guidelines from recognized professional bodies. | Framing what is generally recommended. |
| 2 | Systematic reviews and meta-analyses of randomized trials. | Describing the strength of evidence for an intervention. |
| 3 | Individual randomized controlled trials. | Specific findings, with attention to sample size and population. |
| 4 | Observational studies (cohort, case-control). | Describing associations, not causation. |
| 5 | Public health agency materials (e.g., CDC, NIH, FDA, WHO). | General information, definitions, and public guidance. |
| 6 | Laboratory or animal studies. | Context only, with a clear note that findings may not apply to humans. |
| 7 | Expert commentary, news reports, and other secondary sources. | Background, with primary sources preferred where available. |
Where a statement rests on a lower-tier source, we try to note that in the text. Where a topic has no high-tier evidence, we say so rather than presenting weak evidence as though it were strong.
How we cite sources
Sources are cited inline, next to the statement they support, using numbered references. The reference list appears at the end of each article. Where possible, we link directly to the source, so that readers can check it themselves.
We prefer primary sources — the original guideline, study, or agency publication — over summaries of those sources. Where a primary source is not publicly accessible, we may cite a reliable secondary source and note that limitation.
Dates and limitations
For each source, we consider:
- When it was published, and whether it has been superseded.
- Whether it has been updated, and if so, what changed.
- The population studied, and whether it applies to the reader's situation.
- The size and design of the study, and any stated limitations.
- Whether the findings have been replicated or contradicted.
Older sources are not automatically excluded — some foundational research remains relevant. But where guidance has changed, we try to cite the current version.
Conflicting evidence
Medical evidence is often mixed. When sources disagree, we describe the disagreement rather than presenting one side as settled. In practice, this may mean:
- Noting that some studies show an effect and others do not.
- Describing the range of findings rather than a single number.
- Explaining why the disagreement may exist — different populations, different outcome measures, different study designs.
- Avoiding language that implies certainty where none exists.
Where evidence is genuinely unclear, we say so. This is more useful to readers than false confidence in either direction.
Use of organizational names
We cite organizations such as the CDC, NIH, NHLBI, FDA, and the American Lung Association because their materials are useful primary sources. Citing them does not imply that they endorse this site, review our content, or have any relationship with us. We do not use their logos to suggest endorsement or partnership.
Numbers and examples
Where we include numbers, statistics, or examples, they come from the cited source. We do not invent figures for illustration. Where a hypothetical example is used to explain a concept, it is described as hypothetical rather than presented as data.
Charts and graphs, where used, are based on published data, and the source is noted near the graphic. We do not adjust scales or presentation in ways that would mislead about the underlying numbers.
Contact
If you believe a source has been misrepresented, or if you would like to suggest a stronger source for a specific claim, please use our Contact page. To report a factual error, see our Corrections Policy.