What Supplements Should I Actually Take? (2026)
Informational summary of population health data and published clinical trials — not medical advice. Blood tests, not symptoms alone, should confirm any real deficiency before you supplement for it.
Most adults need at most one or two supplements, not a shelf of them. Vitamin D is the one nutrient that's actually commonly low across the general population — 22.1% of US adults are deficient by blood test.
Everything else depends on you. B12 mainly matters after 65, if you're vegan, or on metformin or a PPI. Iron mainly matters for menstruating women. Most other popular categories are a case-by-case bet on evidence that ranges from strong to thin.
Roughly a third of the products on shelves also carry no independent quality testing at all.
Want the full grade on a specific category? Jump to the default stack or what you can probably skip →
Start from the population, not the marketing
Supplement marketing starts from what sells. This page starts from a different question: what does blood-test data actually show is common in US adults? That data comes from NHANES, the CDC's national health survey, and it's blunt about which gaps are real and which are rare.
| Item | Value (%) |
|---|---|
| Vitamin D | 22.1% |
| Iron | 3.5% |
| Vitamin B12 | 3.2% |
| Magnesium | 2.5% |
| Folate | 0.2% |
Read that chart plainly: for one nutrient, being low is common no matter who you are. For the rest, being low is uncommon overall and concentrated in specific groups. That distinction is the whole first filter.
| Nutrient | % of US adults low | Who's actually at risk | Worth testing? |
|---|---|---|---|
| Vitamin D | 22.1% overall; 32.6% at 18-30 | Everyone, but especially younger adults and non-Hispanic Black adults (53.6%) | Yes — the 25(OH)D test |
| Iron | 3.5% overall; 6.2% in women | Menstruating women, pregnancy, endurance athletes | Yes if at-risk — ferritin test before you supplement |
| Vitamin B12 | 3.2% overall; 6.0% at 65+ | Adults 65+, vegans/vegetarians, long-term metformin or PPI users | Yes if at-risk — MMA test catches what serum misses |
| Magnesium | 2.5% deficient by serum | People on thiazide or loop diuretics | Uncertain — serum poorly reflects body stores, so testing has real limits |
| Folate | 0.2% overall | Essentially nobody, per this data — folate fortification of US grain products works | No, outside pregnancy planning |
Source: CDC NHANES 2017-March 2020 Pre-Pandemic, survey-weighted; B12 and magnesium from published NHANES analyses. See our vitamin D, iron, B12, and magnesium deficiency pages for the full breakdowns.
The second filter: are you actually in a risk group?
Population averages hide a lot. The nutrients above matter far more for some people than others, and three things move you into a risk group: what you eat, what you take, and how old you are.
- Diet pattern. If you're vegan, B12 stops being optional — it comes almost entirely from animal foods, and one cohort found roughly half of vegan men were deficient versus about 7% of vegetarians. See B12 deficiency signs and who's at risk.
- Metformin. Long-term use lowers B12 by roughly 19% over 4 years, and about 1 in 9 long-term users ends up low. See metformin and B12 depletion.
- PPIs. Long-term use carries a similar B12 risk — a JAMA-reported odds ratio of 1.65 after 2+ years. See PPIs and B12 depletion.
- Diuretics. Thiazide and loop diuretics deplete magnesium and potassium. Ask your doctor before supplementing either — don't self-dose potassium. See diuretics and nutrient loss.
- Age. B12 absorption declines with age even though the RDA doesn't rise. NHANES shows deficiency more than tripling from 1.8% at 18-30 to 6.0% at 65+. See B12 dosing for seniors.
- Pregnancy. Folate needs rise sharply, and roughly 40% of women don't efficiently convert folic acid to its active form. See the prenatal supplement guide for why methylfolate is the safer default there.
If none of these apply to you, you've likely just ruled out most of the reasons to take anything beyond the population-level default below.
The third filter: does the evidence actually hold up for your goal?
Once you're past deficiency and risk group, the next question is whether a supplement's marketed benefit is actually the one the trials measured. This site grades that outcome-by-outcome, and the grades are uneven on purpose:
- Ashwagandha — strong evidence for stress and cortisol, preliminary for testosterone.
- Berberine — strong evidence for blood sugar, weak and disputed for weight loss despite the "nature's Ozempic" framing.
- Tongkat ali — real testosterone effect, but only in men who start low; nothing for men with normal levels.
- Curcumin — strong evidence for knee osteoarthritis, but only in bioavailable forms; plain turmeric is close to a placebo by comparison.
- Collagen — reasonably good evidence for skin, thin evidence for hair and nail growth.
Notice the pattern: almost nothing on this list is a flat yes or no. It's "yes, for this specific outcome, at this dose, in this population" — and reading past the headline claim is where most people's supplement budget goes to waste.
The fourth filter: is the product even verified?
A supplement with strong trial evidence is only as good as the bottle in front of you. The FDA doesn't test supplements before they reach shelves. Independent certification is the only real check that a label matches what's inside.
Across 730 products in our own catalog spanning 116 categories, 34% carry no third-party certification at all — no USP Verified, NSF Certified for Sport, or IFOS seal.
That's the last filter, not the first. A certified product that treats a gap you don't have is still a waste of money. But once you know what you actually need, certification separates a verified purchase from a guess. See the full certification-gap data and what "third-party tested" actually means.
The default stack for most adults
Run a healthy adult with no special diet, no relevant medications, and no confirmed deficiency through all four filters, and here's what's actually left:
The one near-universal default Nature Made Vitamin D3 2000 IU $0.06/day
2,000 IU/day is a reasonable maintenance dose if you haven't tested and aren't correcting a confirmed deficiency — USP Verified, so it's independently checked. If a 25(OH)D test comes back under 20 ng/mL, the correction dose is higher; see the dosage guide for that math.
Nature Made Vitamin D3 2000 IU →Past that single default, the rest of the stack is conditional, not universal:
- If you're vegan, 65+, or on long-term metformin or a PPI: add Nature Made Vitamin B12 1000 mcg Sublingual Fast Dissolve $0.11/day, USP Verified.
- If you're a menstruating woman and haven't had ferritin checked: get the test first; if it confirms low iron, Solgar Gentle Iron (Iron Bisglycinate) 25 mg $0.11/day is a gentler-tolerated bisglycinate form.
- If you're on a thiazide or loop diuretic (talk to your doctor first): Vitamin Shoppe Magnesium Glycinate 400mg $0.23/day is a well-absorbed glycinate form.
That's the honest default list for most adults: one supplement everyone should consider, and up to three more depending on who you specifically are. Not ten. Not a multivitamin plus five "just in case" bottles.
What you can probably skip
This is the part worth sitting with, because it's the opposite of what most supplement content tells you. Specifically, and by name:
- A daily multivitamin, if you eat a varied diet. It can't target the vitamin D gap that's actually common, and it bundles ingredients you likely don't need alongside the one or two you might. See do you need a multivitamin? for the full evidence, including where it does help (65+, pregnancy, restrictive diets).
- Folate or folic acid, if you're not pregnant or planning to be. At 0.2% deficient, it's the lowest gap in this site's entire dataset, thanks to US grain fortification.
- Ashwagandha specifically for testosterone. The stress and sleep evidence is strong; the hormone evidence is a handful of small, often manufacturer-funded trials that a dedicated meta-analysis calls inconclusive. See the full verdict.
- Berberine specifically for weight loss. The blood-sugar effect is real; meta-analyses disagree on whether the weight effect exists at all, let alone at the size "nature's Ozempic" implies. See the full verdict.
- Tongkat ali, if you're a young man with normal testosterone. The evidence is real for men who start low and weakest exactly where most buyers assume it'll help. It also carries a documented heavy-metal contamination risk in a meaningful share of tested products. See the full verdict.
- Plain turmeric capsules for anything. Turmeric is only 1-6% curcuminoids by weight and barely absorbed without an enhanced formulation. If you're not taking a bioavailable curcumin extract, the joint-pain evidence doesn't apply to what's in the bottle. See the full verdict.
- Collagen for hair or nail growth specifically. The skin evidence is real; the hair and nail evidence is thin. If hair or nails are your only goal, checking iron, protein, and zinc status first is the better-evidenced move. See the full verdict.
- Any uncertified product priced the same as a certified one. With 34% of this site's own catalog carrying no independent testing, and a certified alternative usually available for a comparable cost per day, there's rarely a reason to choose the unverified bottle.
Frequently asked questions
What is the one supplement most people should actually take?
Vitamin D, for most adults who don't get regular midday sun. It's the only nutrient in our population data where deficiency is common across the board — 22.1% of US adults, rising to 32.6% in adults 18-30 — rather than concentrated in one risk group. Everything else on a reasonable default list depends on your age, diet, or medications, not on being a universal gap.
Do I need a multivitamin?
Probably not, if you eat a varied diet. A multivitamin can't target a specific gap the way a single nutrient can, and it bundles several ingredients you likely don't need alongside the one or two you might. If population data says vitamin D and, for some people, B12 are the actual gaps, a $0.07/day vitamin D softgel treats that gap directly. A multivitamin is a reasonable insurance policy for a restrictive diet or a picky eater, not a default for everyone.
Which popular supplements have the weakest evidence behind them?
Weight loss from berberine, testosterone from ashwagandha or tongkat ali in men with already-normal hormones, and hair or nail growth from collagen. In each case the marketing claim and the actual trial data point in different directions — there's usually a real effect somewhere in the evidence, just not the one on the label.
How do I know if a supplement brand is actually independently tested?
Look for a USP Verified, NSF Certified for Sport, or IFOS (for fish oil) seal on the label — those are the only three programs that independently confirm a bottle contains what it claims. The FDA does not test supplements before they reach shelves. About a third of the products in our own catalog carry none of these seals, so the label alone is not enough to assume verification.
Should I get blood tests before starting supplements?
For anything beyond a baseline vitamin D check, yes, if it's easy to do. A 25(OH)D test confirms whether you're actually low, an MMA test catches B12 deficiency a normal serum panel can miss, and a ferritin test should come before any iron supplement, since iron has real risks in people who aren't actually deficient. Guessing your dose from symptoms alone is how people end up on supplements that were never the problem.
Related guides
- Vitamin D Deficiency: Signs & Who's at Risk — the 22.1% figure, by age and race
- B12 Deficiency: Signs & Who's at Risk — vegans, seniors, and medication users
- Best Iron for Women — why menstruating women need more than twice the RDA
- Metformin and B12 Depletion · PPIs and B12 Depletion · Diuretics and Nutrient Loss
- 34% of Supplements Have No Third-Party Certification
- Do You Need a Multivitamin? — the 5.5-million-participant evidence review
- Our Methodology — how we grade evidence and verify products
Sources
This page synthesizes data and citations already published on the pages linked throughout — deficiency prevalence from CDC NHANES (see the vitamin D, iron, B12, and magnesium pages for full citations), evidence verdicts from the ashwagandha, berberine, tongkat ali, curcumin, and collagen verdict pages, and certification data from our supplement certification gap research. No new PMIDs are introduced on this page — every clinical claim traces to one of the pages above.