Toward a Recommended Framework — After Following the Evidence on What Has Actually Worked
This appendix could end with the easy answer that runs under most of the trilogy's preceding chapters: require the same kind of approval for supplements that we require for drugs. That intuition is largely right — a system where a manufacturer can sell a product to millions of people on the strength of a label they wrote themselves cannot be trusted to police itself. But looking at what other countries actually do — not what advocates on either side say should happen, but what has been running, successfully, for years — changes the shape of that answer. Full drug-style premarket approval for every product in an 80,000-product marketplace has no working precedent anywhere in the world. A more targeted model does. This appendix proposes that model.
Robert W. Coleman, MS Pharm.
Four findings drive everything that follows.
First: the regulatory gap isn't a design flaw that crept in over time — it's the original design, and the market it was built for no longer exists. The Dietary Supplement Health and Education Act of 1994 (DSHEA) gave the FDA no authority to review a supplement's safety or effectiveness before it reaches a shelf. That was written for a marketplace of about 4,000 products.
Second: post-market enforcement, by design, can only respond after harm is already documented — and the documentation itself is thin relative to the market's size. A widely cited analysis of the 2004–2013 window found more than 15,000 reports that included 339 deaths and nearly 4,000 hospitalizations — and researchers who have studied the system describe it as a likely undercount, given how few adverse events ever get reported or connected back to a supplement at all.
Third: a workable middle path already exists, and it isn't hypothetical. Two of the United States' closest peer countries already run functioning premarket systems for exactly this category of product, without collapsing their markets or requiring years-long drug trials for a bottle of fish oil. Canada requires every natural health product to obtain a licence — a Natural Product Number — before sale, with the depth of evidence required scaled to the product and the claim through a monograph system. Australia runs a three-tier model: low-risk products can self-certify against a pre-approved list of ingredients and claims; products making a more specific health claim must have that claim's evidence assessed before the product can carry it; and only products with the highest-risk profile face full registration. Neither country requires what DSHEA-era critics feared and what current maximalist reform proposals imply: that every vitamin be run through the same approval pathway as a new chemotherapy drug.
Fourth: Congress itself is already moving toward the first, most modest piece of this — a mandatory listing requirement, not premarket approval. The Dietary Supplement Listing Act, introduced in 2026, would require manufacturers to file their products with the FDA or have them treated as misbranded. It would not require anyone to prove a product works before selling it. That it is the piece currently gaining traction is itself informative: it suggests the politically viable sequence starts with knowing what is on the market, not with re-litigating DSHEA's core bargain all at once.
None of this changes the conclusion that runs through this investigation: American consumers are, right now, trusting a system to have already checked something it was never authorized to check. What it does is point toward a specific, sequenced fix rather than an all-or-nothing rewrite of the 1994 law that has failed twice before in Congress.
An evidence framework is easiest to evaluate against real products rather than in the abstract. Applying the three-tier structure above to the 25 best-selling U.S. supplement categories sorts them as follows — and the exercise itself illustrates why risk and evidence tier don't track market popularity.
| Tier | What it requires | Where today's best-sellers land |
|---|---|---|
| Tier 1 — Listed | Self-certified; basic nutrient-support claims only | Multivitamin/Multimineral, Vitamin C, B-Complex/B12, Magnesium (general use), Zinc (general use), Biotin, Milk Thistle |
| Tier 2 — Assessed | Evidence submission required for any claim beyond basic nutrient support | Vitamin D, Omega-3/Fish Oil, Protein Powder, Probiotics, Calcium, Magnesium (migraine claim), Collagen, Creatine, Melatonin, Turmeric/Curcumin, CoQ10, Zinc (cold-duration claim), Ashwagandha, Elderberry, Glucosamine/Chondroitin, Echinacea, Garlic Extract, Saw Palmetto |
| Tier 3 — Registered | Full premarket safety/efficacy review — documented adverse event history or disease-adjacent claim | Iron, Green Tea Extract |
Two results are worth pointing out directly. Iron — one of the oldest, most familiar nutrients on the market — lands in the highest-scrutiny tier not because it's novel, but because it's a documented leading cause of fatal pediatric poisoning. Novelty and risk are different axes, and this framework sorts by risk. Green Tea Extract lands there for the same reason: the DILIN liver injury registry data behind Book Two isn't theoretical, it's an observed pattern tied specifically to concentrated extract.
One entry doesn't fit the tiers cleanly at all. Biotin's real problem — interference with troponin and thyroid lab assays — isn't a dosing or efficacy risk the tier system is built to catch. It's a diagnostic-disclosure problem, better solved by this model's separate interaction-warning requirement than by moving biotin to a higher tier it doesn't otherwise need. Omega-3 and Garlic Extract similarly stay at Tier 2 but should each carry a mandatory anticoagulant interaction warning regardless of tier — a reminder that the tier and the warning requirement are two separate levers, not one.
The pattern underneath all of it: several of the highest-risk products on this list are also lower-tier by evidence strength for their primary marketed use — reinforcing that market prominence, evidence quality, and actual risk are three separate variables, not one.
For consumers, the clearest visible change is a label marker — something like Australia's AUST L / AUST L(A) distinction — that tells you at a glance whether a claim has been reviewed or simply self-certified.
For responsible manufacturers, the added burden is real but bounded: a listing filing, and — for products making a specific claim — an evidence submission most legitimate companies already possess internally. Companies whose entire model depends on an untested claim see the most friction, which is the intended effect, not a side effect.
For the industry as a whole, this model does not ban a category of product, does not require multi-year clinical trials for a multivitamin, and does not import the FDA's drug-approval timeline wholesale into a food-regulated category.
The most direct answer — require every supplement to clear the same bar as a prescription drug — is also the version of reform that has failed twice before and has no successful precedent anywhere. No country in our international comparison runs its entire supplement category through a full drug-approval pathway; even the stricter models in Canada and the EU reserve genuinely rigorous premarket review for the highest-risk tier, not the whole market.
DSHEA exists because a stricter FDA posture toward supplements in the early 1990s produced a backlash large enough to pass a law removing premarket authority almost entirely. A reform proposal that asks Congress to reverse that outcome in one step, rather than through the sequenced, risk-tiered model above, is the version most likely to produce the same result it produced in 1994 — not stronger oversight, but a codified retreat from it.
No framework built from an honest reading of the evidence should claim to solve everything. This model does not resolve enforcement against imported products sold through online marketplaces outside FDA's practical reach. It does not stop influencer-driven misinformation, which is a platform and cultural problem, not a labeling problem. It does not guarantee its own funding — a tiered evidence framework is only as good as the appropriations behind the FDA office that has to run it. And it does not, by itself, close the gap in physician and pharmacist training on supplement-drug interactions — that requires a professional-education fix this framework can support but not deliver on its own.
What it offers instead is a sequence: know what's on the market, sort it by actual risk, require evidence in proportion to the claim, and fund the surveillance that catches what evidence review misses. It is a narrower promise than "treat supplements like drugs," and it is the more achievable one.
This appendix draws on the regulatory history in The Hollow Aisle, the risk evidence in that same volume, and the international comparison in Prove It. It is intended, like Beyond Reform's own updated appendix in the companion healthcare trilogy, to close the investigation with a recommended model rather than a diagnosis alone.