# Substantiation files and the evidence behind every claim
A brochure lands on a surgeon's desk. It says a new hip implant delivers "40% fewer revisions." That single number, printed on glossy paper, is a legal liability unless a folder somewhere holds the exact study, the exact comparator, and the exact statistical test that produced it. That folder is called a substantiation file, and in biotech and medtech it is the difference between a defensible claim and a recall of your marketing materials.
A substantiation file (sometimes "claims dossier" or "evidence file") is the organized bundle of documentary proof behind every promotional statement your company makes. One claim, one traceable source. No exceptions.
Regulators do not accept "we believe" or "clinicians tell us." They want the source document that a reasonable expert would accept before the claim was published, not after a complaint.
Two rules govern the whole discipline:
United States. The Food and Drug Administration (FDA) regulates promotional labeling and advertising for prescription drugs and many devices. For devices, the key rule is that promotion cannot be "false or misleading" and cannot promote uses outside the cleared or approved indication (off-label promotion). The Federal Trade Commission (FTC) polices unfair or deceptive advertising more broadly and enforces the prior-substantiation doctrine, especially for lower-risk consumer devices and wellness products.
Europe. Since the Medical Device Regulation (MDR, Regulation (EU) 2017/745, fully applicable since 2021) took effect, Article 7 explicitly bans misleading claims. It prohibits text, names, or images that ascribe functions the device does not have, or that create a false impressionimpressionThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → about treatment. Enforcement sits with national Competent Authorities (for example, the BfArM in Germany or the ANSM in France).
For medicines, promotion in the EU is governed by Directive 2001/83/EC and layered national codes.
Industry self-regulation adds a second net. In the UK, the PMCPA administers the ABPI Code for pharma. In the US, AdvaMed publishes the device industry code of ethics. These codes are stricter than the law in places, and breaching them is public and reputational.
Every superiority or performance claim traces back to one of three sources. Learn to spot which one a claim needs.
Evidence from studies in humans. Randomized controlled trials, registries, post-market clinical follow-up (PMCF, the ongoing data collection MDR requires after launch).
Claim: "Reduces post-surgical infection rate." This is a clinical outcome. It needs a clinical study, ideally comparative, in the relevant patient population.
Laboratory or engineering testing, no patients involved. Tensile strength, flow rates, battery life, sterility, wear simulation.
Claim: "Withstands 10 million wear cycles." This is a bench claim. It needs a validated test protocol and the raw results, not a clinical trial.
Danger: never let a bench result masquerade as a clinical benefit. "Survives 10 million cycles in vitro" does NOT license "lasts 15 years in patients." That leap is a classic misleading-claim finding.
Any claim mentioning a competitor or a category ("vs. the leading brand," "the only device that," "40% fewer revisions than standard care"). This is the highest-risk category because it invites both regulatory scrutiny and a competitor complaint.
Comparative claims need head-to-head evidence or a robust, like-for-like indirect comparison. Comparing your best-case result to a competitor's worst-case is a guaranteed loss.
Take our opening line: "40% fewer revisions."
Work backwards, the way a reviewer will:
| Question | What the file must contain |
|---|---|
| 40% vs. what? | Named comparator or standard of care |
| In whom? | Patient population matching the indication |
| Over what period? | Follow-up duration (2 years? 10 years?) |
| Statistically real? | p-value or confidence interval, sample size |
| Same measurement? | Identical definition of "revision" both arms |
If the study measured revisions at 2 years but the brochure implies lifetime, the claim is broken. If "revision" was defined differently in the two groups, broken. If the 40% has no confidence interval and the sample was 30 patients, likely broken.
A defensible version might read: "In a 24-month randomized study (n=420), revision rate was 40% lower vs. [named comparator] (95% CI, p<0.01)." Less punchy, far more survivable.
Marketing teams manage this with a claims matrix: a living spreadsheet mapping every approved claim to its evidence and its approved wording.
| Claim ID | Approved wording | Type | Source doc | Expiry / review | Approved by |
|---|---|---|---|---|---|
| C-014 | "40% fewer revisions at 24 months vs. X" | Comparative + clinical | Study RCT-2024-07 | Review Q3 2026 | Med, Reg, Legal |
| C-021 | "Withstands 10M wear cycles" | Bench | Test WR-118 | Static | Reg, R&D |
Rules that make it work:
🎬 [VIDEO: "How the FDA Regulates Medical Device Advertising" - youtube.com - a plain-language walkthrough of promotional rules and the false-or-misleading standard for devices]
Before a single asset ships, run the review loop. This is the MLR review (Medical, Legal, Regulatory), standard in pharma and increasingly formalized in medtech.
The sequence:
1. Draft with Claim IDs. Copywriter tags each claim to the matrix.
2. Reference pack assembly. Attach the source document for every claim. Reviewers should not have to hunt.
3. Line-by-line review. Medical checks scientific accuracy. Regulatory checks it stays within the cleared indication. Legal checks comparative and IP risk.
4. Fair balance check. For risk-bearing products, benefits and risks must be presented with comparable prominence. A benefit in 24-point font and side effects in 6-point footnote fails.
5. Approval and version lock. Approved asset gets a code and expiry date. No silent edits after approval.
Consumer-protection rules run underneath all of this. The core question: would the intended audience be misled?
Context changes the answer. A dense statistical claim aimed at interventional cardiologists is fine. The same claim on a direct-to-consumer page for an at-home glucose monitor may mislead a lay reader and trigger FTC scrutiny. Same evidence, different audience, different verdict.
Watch three recurring traps:
Knowledge check
1. A marketing team assembles supporting studies only after a regulator challenges a published claim. Which principle of substantiation have they violated?
2. A brochure claims an implant delivers '40% fewer revisions,' but the underlying study measured a different population than the one targeted by the ad. Which substantiation rule is at risk?
3. Why is the standard 'clinicians tell us it works' insufficient for a substantiation file?
4. Select ALL correct answers about how promotional oversight is divided in the United States.
Select all the correct answers.
5. Select ALL correct answers describing the core purpose and structure of a substantiation file.
Select all the correct answers.
The consequences are concrete, not theoretical.
In the US, the FDA can issue an Untitled Letter or the more serious Warning Letter demanding you pull materials and run corrective communications. The FDA publishes these publicly in its Warning Letters database, so a promotional violation becomes searchable reputational damage. Many historical device and drug letters cite exactly the failures above: unsubstantiated efficacy claims, off-label promotion, and overstated comparisons.
In Europe, a Competent Authority can order withdrawal of non-compliant advertising under national law implementing MDR Article 7. Industry codes add public rulings: the PMCPA publishes named case outcomes.
There is also the competitor route. Rivals monitor each other's claims and file complaints, in the US often through the National Advertising Division (NAD) of BBB National Programs, a self-regulatory forum that rules on truth and accuracy in advertising. A NAD decision against you is fast, public, and expensive to unwind.
You do not need statistics to sanity-test a comparative claim. Try this on any "X% better" line:
Claim: "40% fewer revisions"
Step 1 Absolute rates? Comparator 5.0% → Device 3.0%
Step 2 Absolute difference 5.0% - 3.0% = 2.0 percentage points
Step 3 Relative reduction 2.0 / 5.0 = 40%
Step 4 Ask: which framing does the brochure use, and is it clear?Both numbers are true. But "40% fewer" (relative) sounds far larger than "2 in 100 fewer" (absolute). Regulators increasingly expect the absolute figure to appear alongside the relative one so readers are not misled by the bigger-sounding number. If your file only supports the relative framing, flag it.