Insights from ISPOR Europe 2025 – The Evolution of Market Access and HEOR to Meet Future Needs

For 2025, the ISPOR Europe conference returned to Glasgow.  Held from November 9-12, the event brought together a unique pool of global experts in health economics and outcomes research (HEOR), as well as market access and commercialization stakeholders. In all, there were more than 5,000 attendees, including policymakers, researchers, payers, and patient representatives.

This year’s theme, “Powering Value and Access Through Patient-Centered Collaboration” encompassed discussions around the global implications of changes in US healthcare policy, how healthcare stakeholders can best prepare for future evidence requirements, and how new technologies can increase efficiency and reduce the time required to reach patients. The discussions were engaging and detailed, with considerable debate regarding whether additional change is the right response to more complexity.

Three weeks after the conference, we have been able to step back and assess any ripple effects from the key themes on how to navigate the evolving regulatory, policy and health technology assessment (HTA) environment, exploring more pragmatic clinical trial design, collection of real-world evidence (RWE), the use of real-world data (RWD), and patient-centric research and equity. Inevitably, the potential power of artificial intelligence (AI) was a common thread across many discussions.

Global Implications of US Policy Changes

The potential implications of ongoing US healthcare landscape changes, and most notably Most Favoured Nation (MFN) pricing, touched most sessions. MFN is a drug pricing approach intended to lower prescription drug costs in the US, introduced by Executive Order in May of this year. This policy effectively adopts international reference pricing in the US, aligned to the lowest prices in other developed nations.

In November, after ISPOR, CMS unveiled the five-year, voluntary GENEROUS model for Medicaid, which launches in January 2026.  This provided some additional clarity on how MFN pricing would be calculated.

When MFN was first launched, the US Department of Health and Human Services (DHHS) announced that the MFN price was to be the “lowest price in any OECD country with a GDP per capita of at least 60 percent of the U.S. GDP per capita.”  Under GENEROUS, the CMS proposed a smaller basket, including the UK, France, Germany, Italy, Denmark, Switzerland, Canada, and Japan. Further, instead of the lowest price, it will anchor to “the second lowest country-specific manufacturer-reported net price, adjusted by gross domestic product per capita using a purchasing power parity method.”  Questions remain on whether manufacturers can be legally compelled to disclose confidential net prices, however, with the CMS purchasing power, there are certainly clear implications.

We are already seeing international compromise in action. A compelling analysis presented by RA Capital showed that under the MFN policy, a US-only launch is likely to be the optimal NPV strategy. It reinforces what the industry has always known and we are witnessing this consequence directly in the UK, where companies, including BMS, have chosen to effectively forego launches entirely to safeguard asset price for the US opportunity. And the UK government is responding to the MFN as part of a broader US-UK deal with a historical first – it will now raise its established ICER threshold from £20k – £30k to £25k – £35k in a bid to remain viable.

Implications of the JCA

As with last year, implementation of the Joint Clinical Assessment (JCA) remained a hot topic, especially with expected publications of the first reports early next year. It’s notable that the EMA’s ambition for 2025 was for 30 JCA processes to be started.  Yet, at the time of writing, only 10 have been confirmed, showing that we are still a cautious, risk-adverse industry. Although, once those first reports have been published and analysed to uncover the key to success (or failure) there will likely be a big influx of submissions next year.

The slow uptake does not compromise the goal of the JCA – to harmonize the clinical assessment framework across the 27 Member States, while maintaining market autonomy on clinical benefit assessments and economic evaluations. However, success – or lack thereof – in achieving a compelling, unified, JCA dossier will have a material impact on the market-level assessments.  Member states are looking to incorporate the JCA report in their own processes, with the Spanish Ministry of Health recently signalling that it will adopt the report in its entirety, and without modification.

As a result, what we know even before the first reports emerge is that it is increasingly critical that health technology developers (HTDs) are assessing the payer evidence requirements early, simulating the expected PICOs (Population, Intervention, Comparator, Outcomes), and developing modular dossiers in advance of filing. The JCA timeline, running parallel to the existing EMA marketing authorization process, does not allow for de novo development once the final PICO scope is shared with the HTD.

PICOs, Indirect Comparisons, and Uncertainty

PICOs are scoped with input from all Member States, thereby addressing differences in treatment paradigms across the 27 markets, with a consolidation step designed to streamline the scoping.  However, publication of the first JCA reports early next year will signal how proactive the (co-)assessors will be in reducing the scope down to the critical set.

In addressing the final PICO scoping, HTDs will be required to leverage indirect comparison to complement the core evidence package – likely a randomised controlled trial (RCT). These early JCA reports will further signal how non-RCT data has been received. Indirect treatment comparisons (ITC), network meta-analyses (NMA) and match-adjusted indirect comparisons (MAIC) are all options to supplement RCT data.  However, there are questions as to whether those will be accepted as robust in the JCA given the low acceptance from the JCA predecessor, EUnetHTA. This does not bode well, but we are looking forward to seeing how this plays out and navigating evidence requirements going forwards.

Perhaps more crucially, there was an increased focus on the JCA implications for orphan drugs, which will be required to go through the process from January 2028.  Given the recognised challenges in generating a complete evidence base, the industry is looking to orphan oncology drugs currently within the JCA process to understand if they are being treated any differently because of the sparse evidence base, potentially greater complexity in PICO scope, and greater reliance on indirect comparisons.

Artificial Intelligence in HEOR and Market Access

Artificial intelligence (AI) has been a recurring theme across numerous conferences this year, with increasing belief in the promise that it holds.  However, there is only limited clarity regarding the best applications and methods for using AI.

No doubt, AI is no longer a future concept.  It’s here now, redefining how we work and challenging us to think differently about how we accelerate patient access.  Many of the vendor booths this year were touting AI solutions to deliver better insights, faster.

Conference sessions featured similar considerations to how and when AI should be applied.  Broad application is still the hope, with use cases including AI-powered landscape insights, HTA/pricing to support access strategy, and dossier preparation to expedite regulatory and HTA submissions while ensuring compliance with complex guidelines. As payer evidence requirements become increasingly demanding, and evidence repositories increasingly broad, AI tools will be efficiency drivers.

Terms that little more than a year ago were novel to many – hallucinations, training biases etc. – are now recognized without need for explanation. The current consensus is that any AI-produced material should be caveated as such so that the risks are clear and recognized.  While AI can increase efficiency, human accountability remains the foundation and human intervention is most certainly required to ensure that AI-driven outputs are properly analyzed and vetted for reliability.

One question resonated with us; are pharmaceutical companies being more conservative on the adoption of AI than the healthcare systems themselves? Typically, we see novel technologies, trial design approaches, endpoints, and more being driven through the biopharma companies, with the healthcare systems being more conservative on adoption until there is robust validation and an overall consensus that a given new approach or technology works well.  However, there are potentially more publicly available examples of AI being used in the public space, such as the Danish Ministry of Health using AI to support the development of treatment guidelines. Of course, we are all aware of pharma investment in AI, but at this point it mostly remains behind closed doors.

Looking forward, we also want to consider how AI adoption will potentially impact the “outputs” or the dossiers themselves.  Will extensive dossiers synthesizing available data remain the right submission format?  Undoubtedly, these will remain in the short term but there is opportunity for more technology enabled platforms, to facilitate rolling reviews, “live” evidence updates and single platforms serving as submissions for multiple agencies.

Patients and Data

As we look at the new promise of AI, there is a risk that it has become a fixation. Instead, the focus must remain on the patient, who cannot be an afterthought or a final checkbox – they must be the architect. So, the conference concluded with “patient-centricity” as a familiar, comfortable, and important refrain.

In 2025, the fact that a critical industry question is still, “How do we educate payers on our data?” highlights a persistent challenge we must address: the need to first define relevance for the people who live with the disease.

The solution isn’t necessarily better “education” for payers, often coming too late; it’s earlier patient activation. We don’t need patients to validate our plans; we need them to design them. Their voice can help to better define the evidence gaps that truly matter, enabling changes to clinical development guidelines before we even begin.

We need to evolve from the “some data is better than no data” mindset. Data that does not ultimately drive decision-making can have negative consequences, by adding noise to an already noisy system.  It can waste time and money and can even harm patient trust if access is not ultimately secured.

This is where Human Intelligence must override Artificial. There is ongoing work at NICE to quantify the patient voice, which may show a potential chasm between our current reality and the desired state. Let’s not be simply content at measuring the chasm but consider how we design and redesign to close it.

Conclusion and Outlook

ISPOR Europe remains a landmark conference on the global calendar. There continues to be a search for answers regarding how we, across the healthcare system and industry, can continue to improve evidence, process, and efficiency to bring medicines to patients as fast as possible.

However, more so than in previous years, there was an increased underlying current of uncertainty. What do the changes to healthcare policy and evidence expectations mean for clinical development? How can we ensure that we have the right evidence package to address divergent needs across stakeholders? Does launching outside the US potentially pose more of a risk than it did 12 months ago?

We closed last year by saying “The next 12 months will be an exciting time…” The same is true today, and perhaps we need to be more cognizant that there are potentially more complex risks at play these days. However, we remain excited about helping our clients tackle these challenges and allowing access to patients.