From Evidence
to Impact

In the life sciences industry, the Target Product Profile is one of the most important strategic documents a development team will ever produce. It defines the aspirational attributes of a drug or biologic — the indication, the target population, the efficacy and safety profile, the route of administration, the competitive positioning. Done well, it becomes the compass that aligns clinical, regulatory, commercial, and medical affairs teams around a shared vision.

And yet, in too many organizations, the TPP is treated as a one-time exercise. A document drafted during early development, circulated for cross-functional sign-off, and then quietly shelved while the clinical program runs its course. By the time the asset approaches registration, the TPP is often years out of date — disconnected from the evidence that has accumulated, the competitive landscape that has shifted, and the payer expectations that have evolved.

The Problem with Static TPPs

The challenge is structural. Most TPPs are built around assumptions that made sense at the time: a projected efficacy range based on Phase II data, a safety profile informed by limited exposure, a competitive landscape that looked a certain way when the program was designed. But drug development is a decade-long process. Between the time a TPP is drafted and the time a product reaches market, nearly everything about the environment will have changed.

Competitors enter and exit the field. Regulatory agencies shift their guidance. Real-world evidence from adjacent programs reshapes expectations. Payer frameworks evolve. Biomarker science advances. The assumptions that once felt solid become increasingly fragile — and a TPP that doesn't keep pace becomes a liability rather than an asset.

What a Living TPP Looks Like

A living TPP is not simply a document that gets updated more often. It's a strategic tool that is actively managed, with a clear process for incorporating new evidence, revisiting assumptions, and triggering cross-functional alignment at defined intervals. In practice, this means several things.

First, the TPP should be formally revisited at every major data readout. When Phase II topline results come in, when interim analyses from Phase III are available, when competitive data reshapes the landscape — these are all moments where the TPP should be pressure-tested against reality. Not just by the clinical team, but by regulatory, commercial, medical affairs, and market access together.

Second, the TPP should include explicit assumptions and trigger points. Rather than stating a single efficacy target, a well-constructed TPP articulates a range of scenarios: the minimum viable profile that still supports regulatory approval, the target profile that enables competitive differentiation, and the aspirational profile that unlocks best-in-class positioning. Each scenario should have defined downstream implications for labeling strategy, launch sequencing, pricing, and evidence generation.

Third, the TPP should be version-controlled and transparent. Every revision should be documented with the rationale for changes, the data that prompted them, and the cross-functional decisions that were made. This creates an institutional memory that is invaluable during regulatory interactions, advisory boards, and commercial planning.

The Strategic Payoff

Organizations that treat their TPP as a living document tend to make better decisions at every stage of development. They identify evidence gaps earlier. They adapt their clinical programs to evolving competitive realities rather than discovering misalignment at launch. They enter regulatory discussions with a coherent story that connects early aspirations to accumulated evidence. And they give their commercial teams a head start on positioning, messaging, and payer strategy.

The best TPPs don't just describe what a product should be — they anticipate what the market will demand by the time the product arrives.

In a landscape where development timelines stretch across years and competitive dynamics shift rapidly, a static TPP is a strategic risk. The discipline of keeping it alive — actively maintained, regularly challenged, continuously aligned with reality — is one of the highest-leverage investments a development team can make.

At Vantage Health Strategy, TPP development and lifecycle management is a core part of what we do. We help teams build TPPs that aren't just documents — they're decision-making frameworks that evolve with the program and keep cross-functional teams aligned from early development through launch and beyond.

A few years ago, we were brought in to advise a medtech startup that had built something genuinely impressive: an AI-powered algorithm that could read ultrasound data in real time and detect breast biopsy markers during surgery, giving surgeons live intraoperative navigation to locate and excise non-palpable lesions.

From a clinical standpoint, the technology was elegant. It had the potential to improve surgical precision, reduce positive margin rates, and deliver a better patient experience. The team had strong clinical data. They had regulatory clearance in sight. They were ready to go to market.

There was just one problem. They were about to walk into a wall they didn't see coming.

The Wire Problem

To understand the challenge, you need to understand the current standard of care. For decades, the way surgeons have located non-palpable breast lesions has been through a process called wire-guided localization. Before surgery, the patient visits the radiology department, where an interventional radiologist inserts a thin wire through the breast under imaging guidance, anchoring the tip at or near the tumor. The patient then goes to the operating room with the wire in place, and the surgeon follows it to the lesion.

It works. It's been the standard for a long time. But it has well-known limitations: the wire can migrate, it constrains the surgical approach, the pre-operative placement creates scheduling bottlenecks, and the process is uncomfortable for patients. There is a meaningful rate of positive margins — cases where not all cancerous tissue is removed — leading to re-excision surgeries.

The startup's AI solution eliminated the wire entirely. No pre-operative radiology visit, no physical guide, no scheduling dependency. The surgeon could navigate directly to the lesion in real time using the algorithm's ultrasound overlay. Clinically, it was a clear improvement.

Economically, it was a direct threat.

The Revenue Problem No One Mentioned

Here's what the startup hadn't accounted for: wire-guided localization is a billable procedure. It generates revenue for the hospital's radiology department. The interventional radiologist's time, the imaging suite, the supplies — all of it is reimbursed. In many hospitals, this is a reliable, high-margin revenue stream.

By eliminating the wire step, the AI algorithm wasn't just improving the clinical workflow. It was removing an entire revenue line from the hospital's profit and loss statement. And the people who would need to approve the adoption of this new technology — hospital administrators, department heads, procurement committees — were the same people whose budgets would take the hit.

The startup had built a better solution for patients and surgeons. But they had inadvertently built a worse deal for the institution that had to say yes.

When we came in, the founders were puzzled by the resistance they were encountering. The clinical data was strong. Surgeons who had tried the system loved it. But hospital after hospital was slow to commit. The team assumed it was a clinical evidence problem. It wasn't. It was an economic stakeholder problem.

The SaMD Reimbursement Challenge

There was a deeper issue, too. The startup's product was a Software as a Medical Device — SaMD — and the founders had not fully grappled with what that meant commercially. SaMD occupies an awkward space in the reimbursement landscape. Unlike a physical implant or a drug with a clear billing code, AI-driven software doesn't fit neatly into existing reimbursement frameworks. There was no established CPT code for "intraoperative AI-guided surgical navigation." Hospitals couldn't bill for something that didn't have a procedure code, which meant even the hospitals that wanted to use the technology had no straightforward way to get paid for it.

The startup was brilliant at building algorithms. But they had no experience navigating the labyrinth of procedure code applications, payer engagement, and reimbursement strategy. They didn't fully understand how their own product would generate revenue — not just for themselves, but for the hospitals that would adopt it. Without that understanding, the go-to-market strategy was fundamentally incomplete.

Building the Business Model from the Outside In

Our first task was to help the company understand how they could actually make money. That sounds blunt, but it's more common than you'd think — especially with technically brilliant startups that have focused all their energy on the science and the regulatory path, and haven't yet built the commercial architecture to match.

We started by mapping the full value chain: who does what in the current workflow, who gets paid for each step, and how the introduction of the AI system would redistribute both work and revenue. Then we worked with them to construct a reimbursement strategy that didn't just accommodate the existing system — it created new value within it.

The centerpiece of this strategy was applying for a new procedure-specific reimbursement code for the intraoperative use of their software. Rather than trying to squeeze the technology into an existing code that didn't quite fit, we helped them make the case to the relevant coding bodies that AI-guided intraoperative navigation was a distinct procedure that warranted its own billable designation. This was critical: with a dedicated code, hospitals would have a clear, reimbursable pathway for using the technology. The wire localization revenue would eventually be replaced, not just eliminated.

The Marker Advantage

There was another strategic asset the startup hadn't fully leveraged. Their AI algorithm didn't just work with any biopsy marker — it worked best with a specific marker geometry. The company had developed a proprietary biopsy marker with a round, symmetrical shape that was perfectly detectable by the algorithm's pattern recognition. It was, in fact, the only marker on the market designed for AI-based detection.

This created a powerful commercial flywheel. The software was the platform, and the marker was the consumable. Every hospital that adopted the AI navigation system would also need to purchase the compatible markers — creating a recurring revenue stream on top of the software licensing. It was a classic razor-and-blade model, but the startup hadn't seen it that way until we mapped it out together.

The real business model wasn't just selling software. It was building an ecosystem — a proprietary marker, a dedicated reimbursement code, and an AI platform — where each piece reinforced the others.

With the reimbursement code in place, the economic equation for hospitals changed completely. They weren't losing a revenue line — they were gaining a new one. The AI-guided procedure would be billable. The marker would be a hospital-purchased consumable. And the clinical benefits (fewer re-excisions, shorter OR time, better patient experience) remained the same. Now the value story worked for every stakeholder: the surgeon, the administrator, the CFO, and the patient.

The Broader Lesson

This story illustrates something we see across life sciences again and again: clinical innovation alone is not enough. A product can be scientifically superior and still fail commercially if the evidence strategy, the reimbursement architecture, and the stakeholder value narrative aren't built with equal rigor.

This is especially true for software-based medical technologies, where the reimbursement landscape is still catching up to the innovation. But the principle applies everywhere. A drug that reduces hospitalizations is great for payers but threatening to hospitals that depend on inpatient revenue. A diagnostic that eliminates a biopsy step is great for patients but displaces a procedure that a pathology lab bills for. A SaMD that improves surgical outcomes still needs a billing code before a hospital can say yes.

The companies that navigate this successfully are the ones that think about the economic architecture as early as they think about the clinical evidence. They map every stakeholder who needs to say yes. They build reimbursement strategy into their development timeline, not as an afterthought. And they design their value narrative to speak to each audience in terms that actually matter to them.

At Vantage Health Strategy, this is core to what we do. We help companies — from early-stage startups to global pharma — see their innovations through the eyes of every decision-maker in the chain. From clinician to CFO, from regulatory to reimbursement, we build the evidence and value narratives that bridge the gap between clinical excellence and commercial reality.

Over the past two years, generative AI has moved from a curiosity to a daily tool in life sciences organizations around the world. Medical writers use it to draft first versions of manuscripts. Medical affairs teams use it to summarize advisory board transcripts. Commercial teams use it to generate internal training materials. The technology is genuinely useful, and it's here to stay.

But as AI becomes embedded in the science communication workflow, it's worth pausing to ask a fundamental question: what actually changes about how we communicate science, and what stays the same?

What AI Changes

The most obvious impact is speed. Tasks that once took days — literature reviews, first drafts, data summaries, slide deck outlines — can now be completed in hours or minutes. For organizations producing high volumes of medical communications across multiple indications and geographies, this is a genuine step change in capacity.

AI also lowers the barrier to producing structured content. A medical affairs team that previously relied entirely on agency partners for publication planning can now generate initial drafts in-house, reserving agency expertise for strategic guidance and quality control. This shifts the role of external partners from production to curation — a change that is already reshaping the medical communications landscape.

Perhaps most importantly, AI enables new forms of evidence synthesis. Large language models can process and connect findings across hundreds of publications in ways that would take a human reviewer weeks. This capacity for rapid, broad synthesis is particularly valuable in competitive intelligence, landscape assessments, and identifying evidence gaps.

What AI Doesn't Change

For all its power, AI does not change the fundamentals of credible science communication. And in some ways, it makes those fundamentals more important than ever.

First, scientific accuracy remains non-negotiable. AI models generate plausible-sounding text, but they have no understanding of whether that text is factually correct. In an industry where a misrepresented p-value, an inaccurate mechanism description, or a subtly misleading efficacy claim can have regulatory, legal, and patient safety consequences, human verification is not optional. Every AI-generated science communication must be reviewed by someone who understands the underlying data and the regulatory context.

AI can write faster than any human. But it cannot take responsibility for what it writes. In science communication, accountability is not a feature — it's the foundation.

Second, strategic framing still requires human judgment. A publication strategy is not just about generating manuscripts — it's about deciding which stories to tell, in what sequence, to which audiences, at which congresses, in service of what strategic objectives. AI can help execute that plan, but it cannot design it. The judgment calls about what to emphasize, what to hold back, how to position data in a competitive context — these are fundamentally strategic decisions that require deep domain expertise and an understanding of the organizational objectives.

Third, relationships remain central. The most impactful science communication programs are built on genuine partnerships with key opinion leaders, investigators, and scientific communities. AI cannot attend an advisory board, sense the mood in a room during a data presentation, or build the trust that comes from years of collaborative work with a medical expert. The human dimension of medical affairs is not something that can be automated.

The Real Opportunity

The organizations that will benefit most from AI in science communication are not the ones that use it to replace human expertise. They're the ones that use it to amplify it. By automating the lower-value, time-intensive tasks — first drafts, literature searches, data formatting, slide generation — teams can redirect their energy toward the work that actually moves the needle: strategic planning, evidence interpretation, stakeholder engagement, and quality oversight.

This means the skills that matter most in science communication are shifting. Being able to write a clean first draft matters less when AI can do it. Being able to evaluate whether a draft is accurate, strategically aligned, and scientifically sound matters more than ever. The value is moving from production to judgment.

Looking Ahead

We're still in the early stages of understanding how AI will reshape science communication in life sciences. The technology will get better, the use cases will expand, and the regulatory frameworks around AI-generated content will continue to evolve. What won't change is the fundamental requirement that science communication be accurate, ethical, strategically sound, and ultimately in service of better outcomes for patients.

At Vantage Health Strategy, we help organizations navigate this evolving landscape — integrating AI capabilities where they add genuine value, while maintaining the scientific rigor and strategic clarity that effective communication demands. The tools are changing. The standards should not.