Supplements are regulated as food rather than as medicine in most jurisdictions, which means efficacy need not be demonstrated before sale. Working out what is worth taking requires going to the evidence directly.
Where supplementation clearly works
Correcting an identified deficiency. This is uncontroversial and it is the original purpose of the category.
Which requires knowing that a deficiency exists, generally through a blood test, rather than assuming it.
Specific supplementation during pregnancy is recommended in essentially all national guidance, with strong trial evidence behind it.
Vitamin D in populations with limited sun exposure is recommended in several countries, particularly during winter months at higher latitudes.
Vitamin B12 for people on entirely plant-based diets, since it is not reliably available from plant sources and deficiency causes irreversible neurological damage.
Iron where deficiency is identified, which is common in some groups and should be confirmed rather than assumed since excess iron is harmful.
Where the evidence is genuinely mixed
Omega-3 supplementation has a large and inconsistent literature. Early observational findings were promising and large trials have been more equivocal, with effects appearing in some outcomes and populations and not others.
Creatine has good evidence for exercise performance and increasing evidence being examined for other applications, and it is among the better studied supplements in the category.
Probiotics have evidence for specific strains in specific conditions and very little supporting general use, and the strain specificity is almost never reflected in consumer products.
Where the evidence is weak or negative
Multivitamins in well-nourished populations have been examined in several large trials without demonstrating benefit for the outcomes studied.
Antioxidant supplementation in high doses has, in some trials, been associated with worse outcomes, which was a genuinely surprising result and has been replicated for specific compounds.
Which illustrates that supplementation is not simply a safer version of eating well, and that isolated compounds at high doses behave differently from the same compounds in food.
The quality problem
Independent testing has repeatedly found products containing more, less or different substances than labelled.
Contamination with undeclared pharmaceutical compounds has been documented, particularly in products marketed for weight loss, sexual function and sports performance.
Third-party certification schemes test for content and contamination, and looking for one is the most practical quality check available to a consumer.
Athletes subject to testing have additional reason to care, since contaminated supplements have caused sanctions.
Interactions
The genuine safety concern.
Several common supplements affect the metabolism of prescription medicines, altering their effectiveness.
Some affect blood clotting, which matters before surgery.
Which means anyone taking medication should mention supplements to their doctor or pharmacist, and most people do not because they do not categorise them as drugs.
Upper limits
Fat-soluble vitamins accumulate rather than being excreted, so excess is a real risk in a way it is not for water-soluble ones.
Established upper intake levels exist for most nutrients and are published by health authorities.
Taking multiple products containing the same nutrient is the common route to exceeding them accidentally.
How to think about it
Test before treating, where a test exists.
Prefer food where the nutrient is available from food, since the evidence for whole foods is consistently stronger than for isolated compounds.
Check for third-party certification.
And tell your doctor what you are taking.
This is a summary of published evidence rather than medical advice, and decisions about supplementation should involve a doctor or registered dietitian, particularly alongside any medication or condition.
Reading a label
The practical skill, and it is quick.
Check the amount per serving against the amount used in whatever evidence is claimed, since a fraction of the studied dose is a different product.
Check the serving size, since a product may require several capsules to deliver the stated amount.
Check the form of the compound, since bioavailability differs substantially between forms of the same nutrient and the cheaper forms are frequently less well absorbed.
And check for proprietary blends, where a total amount is given for a mixture without stating individual quantities, which makes assessment impossible by design.
Botanicals specifically
The category with the widest variation in quality and the most interaction risk.
Plant extracts vary by species, part used, growing conditions, extraction method and standardisation, and two products with the same name can differ substantially.
Several are pharmacologically active enough to affect medication metabolism significantly, and the interactions are documented for the common ones.
Which makes disclosure to a doctor or pharmacist more important here than anywhere else in the category.
Timing and absorption
A detail that matters for a few specific nutrients and is overstated for most.
Fat-soluble vitamins are better absorbed with food containing fat, which is a real effect.
Some minerals compete for absorption with each other, so taking them together reduces uptake of both.
Beyond those cases, most timing advice in the category has little behind it, and consistency matters far more than schedule.