Building the evidence-based value story

# Building the evidence-based value story

Edwards Lifesciences did not win transcatheter aortic valve replacement with a better delivery catheter. It won with PARTNER, a trial programme that began by testing the SAPIEN valve in patients no surgeon would operate on. In that inoperable cohort, one-year all-cause mortality was roughly 31% with the valve against roughly 51% on standard care. One comparison did the commercial work: cardiologists got a number they could quote to a referring physician, hospitals got a reason to build a structural heart programme, and the company got a decade of follow-on trials in progressively lower-risk patients.

This lesson is about that assembly. How clinical endpoints, health-economic models and outcome commitments stack into an argument a scientific buyer will accept, and where the stack breaks.

Why clearance is not a value story

Regulatory approval and commercial value answer different questions.

• Approval asks: is it safe, and does it do what it claims (or match a predicate device)?
• Value asks: should this hospital switch from what it already owns, and will the money follow?

Most medtech reaches the US market through 510(kkThe average number of new users each existing user generates through referrals. Above 1.0, growth compounds on itself and becomes exponential.View full definition →) clearance, by showing substantial equivalence to a device already sold. Clearance makes a device legally sellable. It says nothing about whether the device beats the incumbent. When a rep cannot answer "compared to what?", the committee defaults to what it has, and the institutional gatekeepers whose decision mechanics a sibling lesson takes apart never even open the file.

So the first mistake: building marketing around "FDA cleared" as a differentiator. Everyone in your category has it.

The missing ingredient: comparative evidence

Comparative evidence comes in tiers, weakest to strongest:

1. Bench testing and animal data (useful for regulators, weak for marketing)

2. Single-arm clinical studies (your device only, no comparison)

3. Retrospective or registry data (real-world, not randomised)

4. Randomised controlled trials, or RCTs (patients randomly assigned to your product or the comparator)

Your marketing is only as strong as the tier beneath it. Dexcom's argument for continuous glucose monitoring rests on randomised data: in the DIAMOND trial published in JAMA in 2017, adults with type 1 diabetes on multiple daily injections cut HbA1c by about a percentage point over 24 weeks, roughly double the improvement in the fingerstick control arm. Later randomised work extended the same claim into basal-insulin type 2 diabetes. Note what that second trial bought: the same evidence tier applied to a population an order of magnitude larger.

One distinction trips marketers repeatedly. A non-inferiority trial is designed to show your product is not meaningfully worse, within a pre-agreed margin. It cannot carry a superiority claim, however flattering the point estimate looks. If your pivotal study was powered for non-inferiority, differentiation has to come from elsewhere: workflow, cost, safety profile, duration of use.

The message architecture

Think of the value story as a pyramid. Each layer sits on the one below. If a lower layer is missing, the top collapses under questioning.

Layer 1: the clinical endpoint

An endpoint is the specific, pre-defined outcome a trial measures: 30-day mortality, infection rate, time to healing, progression-free survival.

Every claim must trace back to a measured endpoint. Not a hope. A number that appeared in a peer-reviewed paper.

Bad: "Our stent improves patient outcomes."

Good: "In [published trial], the device reduced target lesion revascularisation at 12 months versus the control arm."

The second sentence is defensible. A cardiologist can look up the paper.

Composite endpoints need care. If your one-year win combines death, stroke and rehospitalisation, the first question from any heart team is which component moved. If it was rehospitalisation alone, say so before they work it out themselves.

Layer 2: the clinical claim

Translate the endpoint into something clinicians care about, without overreaching.

Watch relative versus absolute effect. A complication rate falling from 2% to 1% is a 50% relative reduction, a 1 percentage point absolute reduction, and a number needed to treat of 100: a hundred patients treated to prevent one event. Sophisticated buyers compute the NNT themselves. Cite it first and you look like someone who has read their own paper.

Layer 3: the economic claim

Now convert clinical benefit into money, because procurement and payers think in budgets. Fewer infections mean fewer extra hospital days, which means lower cost per patient. Keep the model tied to a measured endpoint; a cost model built on assumptions is a spreadsheet, and buyers know the difference.

Two traps live in this layer.

Cost per patient and total budget impact are different arguments and can point in opposite directions. A one-off implant is a procedure-level decision. A consumable at a few hundred dollars a month across a hundred thousand covered lives is a nine-figure annual line item, cost-effective per patient and still capable of stalling in a budget meeting.

Then ask whose money you are saving. A device that shortens length of stay helps a system paid per capita, and hurts a US hospital paid a fixed DRG amount per admission unless the saving offsets the device cost. Same clinical data, opposite economic story, so the model gets rebuilt per payment context.

Siemens Healthineers pushes the argument one step further with multi-year value partnerships that bundle equipment, service and analytics against agreed operational metrics such as uptime and patient throughput. (It also sells consulting on exactly this sort of value case, so read its published models the way you would read any vendor's.) The second-order consequence deserves a minute: once a commitment is contractual, marketing is underwriting it. Every number in the deck becomes an obligation for operations to hit, and a claim your own service data can contradict.

Layer 4: the positioningpositioningThe mental space you want your brand to occupy in your target customer's mind relative to alternatives.View full definition → line

Only now write the line. It sits on three proven layers, so when someone challenges it you can walk them down to the published endpoint.

🎬 [VIDEO: "How to Read a Clinical Trial Paper" - youtube.com - a clear walkthrough of endpoints, control arms, and statistical significance for non-scientists]

Matching evidence to the audience

Clinicians

They trust peer-reviewed journals and guidelines. Lead with the endpoint, the study design, the sample size and the journal. Name the comparator. Physicians are trained to ask "versus what, and how many patients?"

Payers and health systems

They want cost effectiveness and budget impact. Published appraisals are the cheapest way to calibrate the bar serious payers set: NICE guidance.

Procurement committees

They want risk reduced: track record, real-world data, references from comparable institutions. A registry of 5,000 real patients can outweigh a small pristine RCT here, because it speaks to reliability at scale.

The most common way a good economic model dies at this table: it compares against a standard of care the hospital abandoned three years ago, or assumes a complication baseline higher than the one on its own dashboard. Finance checks its own numbers first. Build the model so the baseline is an input the customer supplies.

Regulatory guardrails on your claims

Your data can support more than your label allows, and where that line sits is the subject of the claims lesson in this module. For the value story, the assembly disciplines matter:

• One claim, one source. Every sentence in the deck maps to a specific result in a specific study, logged in the evidence record the substantiation lesson describes.
• No superiority language built on a non-inferiority result, and no borrowing a competitor's trial to describe your own product.
• Version the collateral. Field decks survive for years. When a follow-on trial supersedes a claim, the old PDF is still on a rep's laptop, and someone has to own its retirement.
• Absolute and relative numbers travel together, always.

The failure mode has a price tag. When a field force tells the story the data does not carry, it surfaces later as a warning letter or a settlement, long after the campaign has been forgotten and the responsible marketer has moved on.

Building the story when your evidence is thin

Not every product launches with an RCT. A pivotal device trial with a few hundred patients across dozens of sites runs into the tens of millions of dollars and several years, which is why the sequencing decision belongs to the executive team and not to the brand plan.

Be honest and be narrow.

Narrow the claim to what the data supports. With a single-arm study, do not imply superiority. Claim what you measured: "achieved X result in Y patients." Precise and modest beats broad and challengeable.

Sequence your evidence generation. Launch with what you have. Fund a comparative or registry study in parallel. Update the value story as stronger data lands. Many products launch on modest evidence and strengthen over two to three years, which only works if marketing and clinical affairs plan it together from day one.

Use real-world evidence deliberately. RWE comes from routine clinical use: registries, electronic health records, claims data. It is faster and cheaper than an RCT and answers the "does it work in messy real life?" question a pristine trial cannot.

Sometimes the obligation arrives from the coverage side. When Medicare opened national coverage for TAVR in 2012, participating centres had to enrol patients in a national registry, so every commercial case also produced data. Edwards inherited a real-world dataset far larger than anything it could have funded alone, plus a public record it had no way to spin.

A quick self-check before any claim ships

1. What endpoint, in what study, supports this exact sentence?

2. Is the comparison fair, with the right comparator and both effect sizes?

3. Is the comparator what this customer actually uses today?

4. Is the claim inside the approved indication?

If you cannot answer all four, the claim is not ready.

Putting it together

PARTNER ran for over a decade and produced trials in inoperable, high-risk, intermediate-risk and low-risk patients. Most companies could not fund a tenth of that. What transfers is the order: endpoint first, clinical claim second, economics third, words last. Reverse it and you get a tagline no clinician will repeat.

Key Takeaways

• Clearance is table stakes. The competitive story rests on comparative evidence, ideally head to head against the standard of care the customer uses today.
• Build claims as a pyramid: measured endpoint, clinical claim, economic claim, positioningpositioningThe mental space you want your brand to occupy in your target customer's mind relative to alternatives.View full definition → line. Every layer traces to a published result.
• Cite absolute effect, relative effect and NNT together, and never dress a non-inferiority result as superiority.
• Cost per patient and budget impact are separate arguments, and a saving that helps a capitated system can hurt a hospital paid per admission.
• Outcome commitments turn marketing into an obligation: once a number is contractual, operations has to hit it and your own service data can refute you.
• When evidence is thin, narrow the claim and run an evidence roadmap, with registry and real-world data doing the work an RCT cannot yet fund.