Introduction

SPEAKERS:

- Jason Lott, VP, TA Head of Global Medical & Evidence Cell & Gene Therapy, Ophthalmology, and Women's Healthcare, Bayer
 - Chris Wright, Managing Director, Value and Access, Consulting, Syneos Health (Moderator)
 - Adam Hathorn, Head, National Accounts, US Market Access, Sanofi

KEY TAKEAWAYS:

·      Payers evaluate products 12–18 months before approval; evidence architecture must begin during Phase 2, not post-launch

·      Real-world evidence complements rebate contracting rather than replacing it, with weight varying significantly by customer type

·      Cell and gene therapy entering common disease markets breaks the average treatment effect model that has governed drug access for decades

·      Agentic AI is resolving internal data silos where HEOR and commercial analytics teams produce different answers from the same databases

·      CMS Coverage with Evidence Development—historically a device mechanism—is emerging as a post-approval evidence requirement for high-cost therapeutics

The pharmaceutical industry is running two evidence programs simultaneously, and they are optimized for entirely different audiences. At Pharma USA 2026, Chris Wright, Managing Director, Value and Access, Consulting, Syneos Health, named the structural fault line directly: "One three-letter agency is doing a regulatory approval and another three-letter agency is evaluating our reimbursement. There's a little bit of a disconnect between the evidence need for regulatory approval versus true reimbursement."

That disconnect is not narrowing. The FDA's evidentiary bar and CMS's coverage calculus have never been formally reconciled, and the commercial payer market operates on a third set of criteria entirely. Jason Lott, VP, TA Head of Global Medical & Evidence Cell & Gene Therapy, Ophthalmology, and Women's Healthcare, Bayer, made the implication explicit:

"There are now levers to pull at the clinical trial design phase to incorporate endpoints that matter, potentially external comparators that matter to payers, even if they're not going to be submitted for regulatory approval. Bring that perspective forward."

Manufacturers waiting for institutional alignment between regulatory and reimbursement frameworks will wait indefinitely. The bridge must be built from the manufacturer's side, and it must begin in trial design.

Evidence Architecture Is Now a Pre-Launch Competitive Moat

The evidence package that clears regulatory review and the evidence package that secures appropriate payer access have different authors, different timelines, and increasingly different contents. Treating trial design as a regulatory exercise produces evidence that arrives at the payer's desk already insufficient.

Adam Hathorn, Head, National Accounts, US Market Access, Sanofi, put the commercial timeline in concrete terms: "Our customers look at products that are coming to market really early on. It could be 12 months, could be 18 months, sometimes even before if it's a disease state that will be a newer disease state to them." That window means payer evidence needs are functionally a Phase 2 design input. The unmet need narrative, the demographic alignment to specific payer populations, and the shape of the forthcoming value story must be legible to payers before pivotal data exists—not assembled from it afterward.

This is where the concept of a living evidence ecosystem becomes operationally meaningful rather than aspirational. Lott argued that "the evidence should be dynamic and it should be evergreen and it should be responsive to the changes in market demands or access constraints that are popping up in real time... payer narratives can be dynamically developed over time and things that were once discrete in time become really living things." The static value dossier—assembled at launch, updated annually, submitted as a finished document—is the wrong unit of analysis. In a market where the IRA, PBM structural reform, and competitive biosimilar entry can shift a payer's formulary calculus within a single quarter, evidence that cannot adapt is evidence that decays.

Wright reinforced what specificity actually requires in practice: the data package lands better when it uses numbers the payer already uses, reflecting their own population rather than a generalized trial cohort. That reframes evidence from a broadcast document into a co-created instrument—one built with the payer's own denominators and defensible in their own internal review processes.

Hathorn cited simultaneous active variables that make static planning untenable: PBM reform shifting from a rebate model to a net sales model by 2028, IRA impacts shaping payer forecasting through 2027 and 2028, and Medicaid channel exposure from pending legislation. These are not future scenarios to monitor. They are current inputs to evidence strategy. Organizations that can generate payer-population-specific analyses inside the pre-launch window are structurally advantaged over those presenting generic value dossiers, and the distinction compounds over time.

Cell and Gene Therapy for Common Diseases Breaks the Model

The industry's CGT pipeline is no longer confined to rare diseases with small, well-defined patient populations. Common conditions are now in scope, and that shift breaks the evidence model that has supported drug access for decades.

Lott stated the methodological challenge without softening it: "We're not talking about rare diseases now. We're moving into an era of CGT for common diseases... we have to move beyond the average treatment effect that we're estimating as part of a routine clinical trial to these individual treatment effects.  That's the holy grail." This is a category shift, not an incremental refinement. The average treatment effect estimated from a controlled trial cannot answer the question a payer must now ask: which specific patient, in my specific population, will benefit enough from a one-time, high-cost therapy to justify this expenditure—particularly if they may switch plans before long-term benefit materializes? Actuarial caution in the absence of individual-level evidence is not irrational. It is the predictable consequence of asking population-level methods to support individual-level coverage decisions.

Likewise, generalizing clinical trial results to payer-relevant populations is another emerging opportunity. Lott described a capability that would have been implausible three years ago: "You would never come to a payer and be like, hey guess what, I got an external control on this that's really reflective of your patients. Or by the way, that's actually using your data, right? And these are your patients and this is how it's stacking up. And it's not perfect, but it's a step further than what we were used to." The "not perfect" acknowledgment matters. The question is whether it is directionally accurate enough to support coverage decisions, and whether the relationship infrastructure exists to have that conversation honestly.

Hathorn confirmed that payer-data partnerships create value extending beyond the manufacturer-payer relationship:

"When you have a partnership with a payer and you are using their data, it is super powerful... it also helps them to justify when they're talking to their customers."

Payers face downstream accountability to employers, unions, and plan members for every formulary decision. Evidence built on their own population data gives them a defensible rationale for their own customers, creating a mutual interest alignment that rebate negotiations alone cannot produce.

That alignment depends on a foundation Hathorn described in operational terms: "Not all your data is going to be perfect. And so you have to acknowledge that and be upfront about that... the more trust you have and the more understanding that you're both trying to achieve the same end goal, then that can really smooth out a lot of the friction." Transparency about data limitations is not a concession. It is the mechanism by which imperfect but directionally valuable evidence earns enough credibility to influence decisions. In a domain where individual treatment effect models carry irreducible uncertainty, the relationship equity to be believed despite that uncertainty is itself a form of competitive infrastructure.

The Internal Prerequisite Most Organizations Haven't Met

The external evidence challenge has an internal prerequisite that most large pharma organizations have not cleared: their own commercial analytics and HEOR teams are producing divergent answers from the same data. Before the industry can build living evidence ecosystems for payer customers, it must first achieve internal coherence.

Lott named the organizational failure precisely: "At a lot of pharmaceutical companies, your commercial data science and analytics team is separate from your HEOR or your real-world evidence team... I've seen agentic AI act as a single source of truth and a harmonization mechanism that really democratizes how we derive insights from data so that really everyone's on the same page inside of an organization." The internal silo is not primarily a technology problem—the databases often overlap. It is a query and governance problem. Agentic AI that normalizes how questions are asked and answered across functions creates the organizational foundation on which all external evidence communication depends. Without it, manufacturers risk presenting payer customers with analyses that contradict their own internal modeling.

Wright surfaced a second internal constraint operating simultaneously: portfolio triage at the field level. National account teams carrying multiple brands must allocate limited customer-facing time against competing internal priorities, and the evidence asset that matters most to a regional payer may be irrelevant to a PBM managing national formulary decisions. Hathorn was direct about the consequence: "You have to be able to say back to the brand, like, you know, I'm sorry, this is not that big of a priority for my customer. We'll do what we can do to establish the access that you need, but we have other priorities that we need to work on."

Lott also flagged a policy variable that belongs inside scenario planning but is frequently absent from it. CMS Coverage with Evidence Development decisions—historically applied to devices—are emerging as an anticipated mechanism for high-cost therapeutics, potentially generating post-approval evidence requirements resembling clinical trial protocols in scope and burden. Market access teams modeling only regulatory approval and commercial formulary access are missing a third evidence obligation that is not hypothetical.

From Deliverable to Infrastructure

The panelists described, without explicitly naming it, a category shift in what "evidence" means for market access. Evidence is moving from a deliverable—the static value dossier assembled at approval—to an infrastructure: a continuously operating system that links trial data to real-world data, generates payer-population-specific analyses on demand, and updates as policy and competitive conditions shift.

Lott pointed to a research model that illustrates where this is heading. At Stanford's Center for Digital Health, a team published work in Nature using networks of agentic AI to run drug development processes end-to-end, from modeling through candidate identification. The implication for market access is not that AI replaces account teams. It is that hybrid teams—human account directors paired with AI agents handling data synthesis, scenario modeling, and population-level analysis—are already technically feasible and will become operationally standard. "I want to see hybrid teams of humans and agents as the years progress," Lott said. "Some people are already doing it."

Hathorn offered a commercially candid recalibration of where real-world evidence actually sits in payer decisions: "I don't know that it would outweigh it. I think it complements it... you could be talking to a PBM or a GPO of a PBM and, you know, it's pretty much financial. But you could be talking to a regional payer and it's really impactful." The assumption that superior evidence translates uniformly into superior access is wrong. Evidence infrastructure investments must be calibrated to the account mix—weighted toward payer types where population-level data and trust relationships can move formulary decisions, and paired with financial contracting where they cannot.

The diagnostic question for market access leadership is not whether the organization has a value dossier. It is whether the organization can link pivotal trial data to real-world claims at the patient level after trial exit, generate a payer-population-specific external control analysis within the pre-launch planning cycle using the payer's own data denominators, and produce the same answer from HEOR and commercial analytics when querying the same underlying database. For organizations approaching cell and gene therapy in common disease markets, a no on any of these is not a capability gap to close post-launch. It is the access risk that is already accumulating.

This content reflects the personal views of the speakers and is not intended to represent the official position of their respective companies.

To get you highlights of Pharma USA 2026 faster, we are using generative AI technology to summarise the transcripts of the sessions. If you have any feedback about the summary, please contact lucy.fisher@thomsonreuters.com.

"Trust is built over time," Schlessinger said. "One small slip can make it vanish."

That asymmetry — slow to build, fast to lose — has a structural implication that the session's full argument demands: brand teams cannot manage trust. They can only create the conditions in which trust becomes possible. Those conditions are internal alignment, field force capability, real-world evidence infrastructure, and operational programs that remove friction between prescription and patient access.

The organizations that understand this are not building trust campaigns. They are building systems. In oncology, where a community physician managing 30 tumor types needs a succinct, scientifically credible answer in under five minutes, and where a rural patient over 65 may never reach a center of excellence, the difference between a campaign and a system is the difference between a patient who gets the standard of care and one who doesn't. That outcome is determined by decisions made in cross-functional strategy meetings, field training programs, and logistics contracts — not in the brand messaging document.

To get you highlights of Pharma USA 2026 faster, we are using generative AI technology to summarise the transcripts of the sessions. If you have any feedback about the summary, please contact lucy.fisher@thomsonreuters.com.

"Making sure that you tailor your governance to the appropriate risk level — just like every other business decision or tech investment you would make — is critical to ensuring that you don't get caught up in a morass of doing the same process for something that's super risky as for something that everyone would understand is not as risky."

A rep copilot surfacing pre-call planning information inside an existing Veeva workflow is not the same governance problem as an AI system synthesizing clinical trial data into brand strategy recommendations. Treating them identically wastes compliance resources and builds organizational resistance to the governance function itself.

Joshi formalized this into a three-tier framework at Takeda. Tier one covers low-risk, high-return tools — rep copilots, omnichannel copilots, market research assistants — where experimentation within existing ecosystems carries limited regulatory exposure. Tier two covers marketing automation and MLR processes, where legal and regulatory involvement becomes necessary. Tier three is the category every pharma company is racing toward: an AI system capable of synthesizing clinical trial data, conference presentations, market research, and transactional records into disease area strategy. "This is the most high, high risk area," Joshi said. "We don't have a real answer over there." The framework is useful precisely because it makes that uncertainty explicit rather than treating all three tiers as equivalent unknowns.

BMS has operationalized the time dimension of this calibration. "Gone are the days when you had unlimited funding to run pilots you wanted to run," Tantsyura said. "We condensed the timelines — before it used to be unlimited time; now you have to show the results within the three-month time period.

Leadership now is expecting that before we even start working on anything, we have very clearly defined value — not just slides, but demos — showing every two weeks what progress has been made." A hard pilot window with biweekly accountability replaces open-ended exploration. The three-month constraint is not primarily a cost control mechanism. It is a forcing function that surfaces adoption problems and compliance friction early enough to address them without sunk-cost pressure.

One governance actor most technology conversations still overlook is procurement. "Procurement is also having a seat at the table," Joshi said. "Procurement along with legal has been really driving conversation much earlier on in the journey of AI, so that you don't have to do expensive retrofitting of all this from a regulatory standpoint. That has been the big learning." Procurement's earlier entry into AI investment decisions is preventing downstream regulatory costs — but it also introduces a new friction point at the front of the pipeline that organizations haven't fully mapped into their governance timelines.

The Capability No One Named But Everyone Described

Dr. Salinas posed the question that crystallizes the regulatory specificity of the pharma context: pharma operates in a very different space, with strict regulatory oversight — how are organizations responsibly deploying AI while ensuring compliance and maintaining audit visibility? The session produced multiple frameworks in response. What it did not produce — because no panelist named it explicitly — is the capability that connects all of them.

The measurement dimension points toward it. "It's no longer about Gen AI insights," Liu said. "We need to measure also the activities that are taken upon those insights and then what's the outcome of that. That's the next level we're trying to get into." The people dimension confirms it. "Do we have the leadership alignment? Do we have the team operationally scalable? This is about your deployment, driving the change management and thinking through the end-to-end pull-through," Liu said. Both observations describe organizations that must hold multiple operational modes simultaneously, not transition from one to the next.

The capability is governance ambidexterity: the organizational capacity to run parallel governance tracks at different speeds and scrutiny levels for different risk tiers, without collapsing them into a single process. Tantsyura confirmed BMS is already doing this — federated governance for traditional AI, centralized cross-functional review for agentic AI, both live at the same time. Joshi described the "ambidextrous organization" in terms of operations versus experimentation. The evidence suggests the same structural demand applies to governance itself.

Liu's measurement evolution — tracking not just AI outputs but actions taken and outcomes produced — requires governance infrastructure that spans from content generation through field execution through business impact, across multiple compliance regimes simultaneously. That is a different organizational design problem than the one pharma solved when it first built centralized AI review boards.

The organizations advancing fastest are not the ones with the most mature governance model. They are the ones that can operate multiple governance models in parallel — applying proportionate scrutiny through risk-tiering, maintaining federated guardrails for known-risk categories, and restarting centralized review for emerging ones — without treating each new technology generation as a crisis that suspends everything else. The diagnostic question for any leadership team is not whether your governance is mature. It is whether your governance architecture can absorb the next generation before this one has finished scaling — and whether you have already started building the answer.

To get you highlights of Pharma USA 2026 faster, we are using generative AI technology to summarise the transcripts of the sessions. If you have any feedback about the summary, please contact lucy.fisher@thomsonreuters.com.

Text

Discover more on this topic at Pharma Customer Engagement USA 2026 (October 27-28, Philadelphia) - where commercial, marketing, medical, data and AI pioneers converge. Explore the agenda here.