How to read the evidence

The strength word

Each finding opens with one word for how much research stands behind it. It grades the evidence, not the thing itself. A strong finding can be a small effect, and a limited one can be a large effect seen once.

• Strong Several meta-analyses or systematic reviews agree.
• Moderate One meta-analysis, or several controlled trials that agree.
• Limited One trial, a few small ones, results that vary a lot, or case reports.
• Untested No study in people. The finding says what it rests on: animal studies, lab studies, how it acts in the body, documented traditional use or what users report.
• No research yet We looked and found nothing to report yet.

What was found

The sentence after the strength word says what the research found, in words tied to the grade.

• Shown to comes from trials or reviews of trials. Something was given, and the result was measured against a comparison.
• Linked to comes from studies that watched people without changing anything. A link is not proof of cause. People who drink coffee or sleep seven hours differ in many other ways too.
• Not shown to help means it was tested and did not beat the comparison. That is a finding, and it stays on the page.
• Mixed results means the studies disagree. The finding says how.
• Checked, no increase found means a possible harm was looked for and not seen.

Compared with

"Better" only means something next to what it was better than. Diet findings, and many others, name the comparison right under the sentence: a low-fat diet, usual care, a sham treatment, people who ate no red meat. A diet defined by one rule, such as keto or intermittent fasting, states it once for the whole page.

For a therapy, beating no treatment and beating a sham are different results. The page keeps them apart.

Who was studied

When research was done in one group only, the finding carries a mark: studied in men, in women, in children, or in pregnancy. A result from one group may not hold for another. Bloodwork never offers a finding from one group to someone outside it.

How common a side effect is

Risks carry a band, taken from the rate the source reports in people taking it. We never estimate one.

• Very common: more than 1 in 10 people.
• Common: between 1 in 100 and 1 in 10 people.
• Uncommon: between 1 in 1,000 and 1 in 100 people.
• Rare: fewer than 1 in 1,000 people.
• How often is not known: the sources give no rate.

Marks beside a name

• Grey market: not approved for human use in the US, and sold anyway.
• Caution: offered where research found it may help, with a risk record heavy enough to say so beside its name every time. The reason is written out where it appears.

Bloodwork levels

Type in a result and each published range it falls in is listed, in its source's own words. The level at the top is set by rule from those ranges, never by our judgment.

• In range: every source that covers the result puts it in range. Nothing is suggested for it.
• Optimal: the same, where a source's own word for that range is optimal.
• Slightly elevated or slightly low: the sources split, or the result is in the first step a source grades outside its range.
• Elevated or low: every source puts it outside its range.
• Concerning: a source calls that range severe, very high, or the like, or it is a diabetes range. The page says to see a doctor, and lists nothing until you tick that you already have.

Ranges written for a risk group, a diagnosis or a treatment are shown apart and never set the level. With your age entered, a range written for an age band is used only if you are in it.

"Based on research, these may be helpful" lists supplements, foods, diet patterns, practices and daily life factors that research found moved that test the other way, in people like the one the result is for. It is not a promise that they will for you, or that moving the number changes your health.

Bio Hack

Bio Hack lines up what separate pages report about the things on your list: where their effects stack, which interactions research or a label names, and what each was found to do. It reports. It does not tell you to add or drop anything.

Where the findings come from

Meta-analyses and systematic reviews first, then the largest trials, then large cohort studies, plus drug labels and regulators for safety. Peer-reviewed or official sources only. Every journal citation is checked against its DOI record before it is published, and every number is taken from the source itself. Each page lists its sources, and each finding links to them.