ISPOR 2025: Collaborating to put patients first

Like most conferences, ISPOR 2025 promised a lot. But it also delivered. Its theme was “Collaborating to Improve Healthcare Decision Making for All: Expanding HEOR Horizons.” We listened and we learned a lot – but the conference also confirmed much of our thinking about the patient voice, Generative artificial intelligence (Gen AI), Real-world evidence (RWE), US healthcare policy, and the global impact of the EU’s Joint Clinical Assessment (JCA).

The end of patient tokenism?

A conversation within and outside the conference sessions spoke to the end of the era of “token” patient involvement in healthcare research and product development. That era is being replaced by patient centricity – putting the patient at the heart of everything.

At ISPOR, speakers acknowledged that drug manufacturers routinely collect patient experience data (PED) to inform how they understand a disease and to help ensure their product development work provides the relief from symptoms that patients long for. But, to truly rise to the challenges of patient-focused drug development, right from the start of the development process we need to unite the key players – industry professionals, patient advocacy groups, and regulators. By collaborating earlier, we can identify the issues that genuinely affect patients and ensure the sponsor’s program collects the most relevant data. This level of collaboration can lead to better regulatory decisions while also ensuring that PED is reflected in labels and the resources patient actually use. Regulators, such as the Food and Drug Administration (FDA), already encourage sponsors to engage “early and often” in terms of how they collect and use PED in development programs. But industry is often unsure or hesitant about exactly when to engage with a regulator and what information is needed to support PED discussions to gain the type of valuable feedback, they need to make more meaningful decisions for patient care.

At a session titled “Enhancing regulatory engagement to streamline and advance patient-focused drug development” speakers Pujita Vaidya (Sanofi), Michelle Campbell (FDA), Ellen Janssen (Janssen Research), Rececca Noel (Amgen), discussed a way forward. Campbell reiterated: “We really mean it when we say, come early, come often.” She noted, for example, that it is problematic if a sponsor submits a marketing application, and the FDA does not understand how what the sponsor is doing fits with the program. Similarly, “If you come at end of Phase 2 and tell us about a new PED that you want to collect, our advice may look very different versus if you’d come to us sooner.” The FDA can provide advice not only based on current guidance but on its past experiences. “We may be able to steer to help you.” Early engagement can allow the FDA to pinpoint the type of additional information that it will need. Campbell said it is especially important to have pre-Investigational New Drug (pre-IND) conversations in rare disease, given that there are fewer options for collecting PED.

In conference sessions, speakers noted time and again the need to integrate the patient voice throughout the entire product life cycle – including post-market approval where patients are often overlooked. Clinical outcome assessments and patient-reported outcome measures have provided robust frameworks to track biomarker changes and their meaningful impact on patients’ lives. Early qualitative research in the development phase can be key to identifying the concepts that matter most to patients and to guide endpoint selection.

For clinical trials, speakers called for more deliberate patient engagement in trial design. Post-approval monitoring should leverage patient registries, RWE and value-based contracts to ensure surrogate or novel endpoints lead to clinical benefits.

Is AI confidently giving the wrong answers or are we confidently asking the wrong questions?

There were an array of sessions focused on Gen AI at ISPOR 2025 and for good reason. At the end of 2024, we heard a statement of intent from the National Institute for Health and Care Excellence (NICE) to develop formal evaluation criteria for AI-generated evidence, quickly followed by FDA guidance on the same topic. Just before the ISPOR conference, the FDA announced deployment of its in-house secure Gen AI system to all centers by June 30, following the success of the FDA’s first AI-assisted scientific review pilot. We also heard from the ISPOR audience that 35% of them had used Gen AI in the context of HEOR in the past year – a 10% increase in use since last year’s conference.

Various use cases were presented across forums. The most common use case in HEOR is still evidence review with several commercial Gen AI platforms now available with increasingly impressive accuracy. Evidence synthesis including meta-analyses and real-world data analyses, early modeling and dossier writing use cases were also exhibited, but with markedly reduced accuracy. More work is needed on HEOR specific models that can provide more accurate outputs for these deliverable types before we can expect them to be routinely adopted and approved for use by decision makers. A key takeaway from all sessions was the need to evolve human expertise, understanding and capabilities in parallel to Gen AI model evolutions. The future is not Gen AI alone but Gen AI + Human and our own skills development is as important as the Gen AI model developments themselves. Recognizing the limitations of Gen AI is as important as recognizing its potential and asking the right questions in the right way is critical to its effectiveness. A lot of the short courses and sessions thus focused on prompting techniques and what we can reasonably expect Gen AI to support in the context of HEOR. “AI hallucinations” are still a common phenomenon and the real question we should be asking ourselves is how we can most appropriately use Gen AI in the context of HEOR to speed up but not compromise high quality science.

US health policy: balancing innovation and affordability in the era of drug price controls

Ever mindful of the local and global impact of health policies on patients, ISPOR conference sessions discussed health policy changes in the US and Europe. For the US, one of the questions discussed was “Drug price controls—what are the unintended consequences to innovation?” A clear message here was that drug price controls could result in drug manufacturers pulling back from pursuing some of their drug development plans and this will undoubtedly affect patient care. The discussion focused on the impact of price setting on pharmaceutical innovation; the difficulty in conveying those facts to policy makers, and why proposals for International Reference Pricing may not be the best approach to take.

The US Inflation Reduction Act (IRA) and proposals for International Reference Pricing could affect investment, particularly in terms of high-risk, early-stage and small-molecule research & development. Speakers called for evidence-based policymaking and for the HEOR sector to be more active in shaping future policy. Going forward the HEOR sector should focus even more on delivering useful data, evidence, research, clear messaging, and practical solutions that can guide health policy that benefits patients. Doing so will help to ensure that reforms support affordability and innovation.

For the Centers for Medicare & Medicaid Services (CMS) the phased implementation of the IRA price negotiations started with Part D drugs but have moved on toward including Part B therapies – which include high-cost biologics and other single-source drugs. The difference between the two negotiations is that Part D relies on negotiated plan prices while Part B drugs are reimbursed based on the Average Sales Price plus 6%. Given that difference, manufacturers must be prepared to meet rigorous Information Collection Request standards for Part B drugs. A roadmap for integrated evidence planning is to: anticipate IRA selection, generate value-based evidence, and ensure timely submissions that include RWE on unmet need, therapeutic value, affordability, and real-world use. Use of AI and hyper automation in dossier development, cost-effectiveness modeling, and literature reviews may help to streamline timelines while maintaining quality through human oversight. Speakers noted that this approach can enable HEOR and access teams to develop CMS-ready value dossiers and build viable market access strategies across indications. The key takeaway here was that AI and RWE are no longer tools for future use – they are now essential for developing value, access, and pricing strategies in a post-IRA world.

Is your HCEI “CARSE”?

Interest in communicating health care economic information (HCEI) to US payer audiences has grown in recent years, as manufacturers continue to look for ways to strengthen their value narrative for new treatments. While the FDA has provided guidance on what should be communicated to this audience, there is still uncertainty and some frustration over what constitutes HCEI, who the appropriate audience is, and what the appropriate evidentiary standard for these data actually is.
 
At a workshop session moderated by Jeff Lee, who leads Lumanity’s North America consulting team in HEOR, the discussion turned to insights on standards to evaluate whether HCEI meets the bar for competent and reliable scientific evidence (CARSE) from academic, manufacturer, and payer perspectives. Speakers included Diana Brixner, (University of Utah), Kristin Gillard (Bristol Myers Squib), and Jessica Daw, (Sentara Health Plans). The speakers discussed key considerations regarding balancing rigor with relevance when delivering HCEI to US payer audiences and provided insights on building credible economic evidence and communicating this evidence transparently to ease payer skepticism. The workshop concluded that communicating HCEI based on CARSE can broaden the typical scientific dialogue between manufacturers and payers; while there are guidelines to support these discussions, currently manufacturers might apply them differently. Having a plan to assess CARSE systematically will drive success when manufacturers want to communicate HCEI more accurately. But, designing studies to generate HCEI for payers must be tailored to specific audiences.

Could JCA be about to go global?

Alongside plenty of talk about emerging US health policy, ISPOR 2025 also looked to health policy in Europe. JCA was launched in Europe just five months ago (in January 2025) as part of the European Union’s Health Technology Assessment Regulation (EU HTAR). At ISPOR audiences considered the possibly “global impact of JCA” on health technology assessment (HTA) systems outside of the EU.

Ultimately the conclusion was that it is too soon to tell but global markets will certainly be keeping a close eye on JCA outcomes and its perceived success. The more general move toward joint processes and harmonization at various points within the HEOR / Market Access pathway is supported by all as the “new direction” for the ongoing evolution of global HTA. Regional collaborations outside of the EU continue to grow, as well as collaborations beyond the scope of the JCA with payers also looking to the potential of joint pricing negotiations.

We heard some great success stories here including formal collaborations such as the Joint Nordic HTA-Bodies (JNHB) group reducing the mean assessment time for oncology and rare disease medicines to only 82 days; and informal collaborations such as the Canadian Drugs Agency (CDA-AMC) adapting evidence synthesis used for decision making by other established HTA bodies, such as NICE and the Institute for Clinical and Economic Review (ICER), resulting in faster decision making in Canada. Let’s hope the JCA lives up to its intent and we hear of great success stories from this collaboration at next years’ conference.

Conclusion

In traveling to ISPOR 2025 we had high expectations. We left feeling fulfilled by the knowledge that while health systems are increasingly complex, we are all working toward the same goal – improving patient therapeutics, their access and patient experience while becoming more collaborative in how we work and the tools we use.