ISPOR 2025 Decoded: Axtria’s Must-See Insights and Emerging Themes

While catching up with old friends, meeting new ones, and sharing great meals along the way, the Axtria team presented seven posters (links at bottom) at this year’s ISPOR 2025 in Montreal, an event that offered a wide-ranging exploration of key trends shaping Health Economics and Outcomes Research (HEOR). Topics ranged from the breadth of data sources and platforms to advanced methodologies like causal inference and the promises and current limitations of generative AI (GenAI). The shift toward cost-effectiveness value frameworks and the growing influence of Health Technology Assessments (HTA) was prominent and many were looking for clarity across the evolving policy landscape, including the Inflation Reduction Act (IRA) and recent executive orders. While GenAI dominated many discussions, the conference's true strength was in bringing together diverse threads that will define the next era of HEOR. The conference revealed an underlying consensus: while promising, GenAI has yet to deliver fully on its significant hype.

Building Toward Breakthroughs

The Time for GenAI, Causal Inference, and Innovative Partnerships Has Arrived

GenAI Must Deliver on Its Promise. The parade of breakthroughs signal that superhuman AI is already operating among us. Yet, just as Google’s AlphaEvolve demonstrates algorithmic universality—applying across math, computing, or any problem expressible as an algorithm—GenAI now faces the imperative of consistent, real-world impact. The next phase will demand rigorous benchmarking, transparent evaluation, and cross-sector validation to ensure GenAI finally fulfills its hype.

Causal Inference Expressed through Game Theory

Despite methodological advances, causal inference often remains underappreciated outside specialist circles. A new narrative—one that frames causal methods as unexploitable strategies in an informational game—can reposition these techniques at the core of HEOR. By embedding game theory concepts, researchers can communicate causal inference as the strategic foundation for studies rife with uncertainty, mirroring how players navigate incomplete information with calculated, unexploitable decisions. For example, a GTO strategy would protect a researcher (player) from faulty design (blundering) as the analysis progresses (the cards are dealt).

Innovative Partnerships as the Catalyst

The complexity of modern HEOR challenges requires more than single entities can supply. The future lies in strategic alliances uniting patient-driven data models, advanced GenAI platforms, and causal expertise. Collaboration among RWE providers, AI innovators, payers, and patient advocates will be the engine that converts pilot projects into practice, translating insights into better therapies, policies, and patient outcomes.

What follows is a set of topic-level highlights that shaped our week—each area supported by our recent work and reflections for those seeking a deeper dive.

Data Options and Foundational Models

The conference highlighted an abundance of data sources and options, each with varying degrees of built-in analytic tools. Companies such as Truveta, Komodo Health, and Panalgo joined OM1 in demonstrating platforms that integrate data access with analytics to support advanced HEOR. Many organizations showcased capabilities that leverage enriched datasets to create advanced analytical frameworks. Notably, companies like OM1 demonstrated how they use healthcare data to train foundational models, like their PhenOM platform, to enhance predictive analytics and clinical insights. Other organizations, such as Clinakos, are focusing on patient-centered data models by obtaining patient consent to close the gaps in rare disease evidence needs. For a deeper dive into current data trends, we invite you to explore our white paper, Real-World Data Ecosystem and New Frontiers in Evidence Generation.

Causal Inference: Methodological Depth and Practical Application

Causal inference emerged as a methodological pillar at ISPOR 2025. A standout short course, "Causal Inference and Causal Diagrams in Big, Real-World Observational Data and Pragmatic Trials," examined methods such as directed acyclic graphs (DAGs), approaches to time-dependent confounding, and target trial emulation. These techniques were explored through case studies, helping participants apply causal thinking to real-world data challenges.

Another session, "What Causal Inference Teaches Us About the Limitations of Indirect Treatment Comparisons for Health Technology Assessment," unpacked the assumptions behind indirect comparisons and highlighted how causal inference tools can improve rigor and validity. These sessions reflected ISPOR's commitment to strengthening the methodological backbone of HEOR through actionable insights.

The Axtria team recently hosted a webinar titled Accelerating Market Access with AI: It's Time to Look Beyond Propensity Score Matching, which discussed the impact of causal inference in the context of our recent case study for a regulatory submission.  Download File

GenAI: Expectations vs. Reality

Key sessions such as "Introduction to Applied Generative AI for HEOR" and "From Prompting to Policy" highlighted ongoing explorations into AI's practical application, notably in systematic literature reviews, economic evaluations, and health technology assessments. Despite these insights, attendees expressed widespread skepticism. The tools discussed in the presentation often presented robust theoretical benefits but struggled with real-world applicability, particularly in delivering consistent, reproducible outcomes.

Integrating AI generated significant interest while emphasizing the necessity for human oversight. The "AI Agents and Guardrails in HEOR" session aptly captured the prevailing sentiment: while automated processes can substantially enhance productivity, effective guardrails and clear ethical guidelines remain paramount.

Despite these insights, the ability to make them useful in real applications remains a work in progress. We've seen this pilot inertia firsthand and have therefore developed a platform to scale into the future, and it goes like this: Axtria InsightsMAx.ai

Health Technology Assessment: Evolving Frameworks and Global Collaboration

Health technology assessment (HTA) was a central theme at ISPOR 2025, with sessions highlighting its expanding role in healthcare decision-making. The plenary "Evolution of Evidence—Innovating for the Future of HTA" emphasized a more holistic HTA model that integrates real-world data, patient input, and broader societal values. Joint Clinical Assessments (JCA) aim to harmonize the process and sessions like "Global Impact of JCA" and others explored international efforts to align HTA processes and improve access.

Another key focus was methodological innovation, with discussions on applying causal inference and other analytical approaches to enhance HTA rigor. Ethical and patient-centered perspectives also featured prominently, reflecting a move toward more inclusive and balanced evaluations.
Read more about Health Technology Assessment in a Multi-Indication Era.

The Inflation Reduction Act: Navigating Drug Pricing and Innovation

The Inflation Reduction Act (IRA) was a prominent topic, with sessions addressing its implications for drug pricing, innovation, and patient access. Discussions explored potential unintended consequences, such as reduced investment in novel therapies and narrower treatment options for patients. The IRA sessions emphasized balancing cost containment with ongoing pharmaceutical innovation and maintaining robust patient access to essential therapies.

Forward Thinking and Policy Preparedness

Reflecting broader geopolitical influences, there was substantial discussion on the need for integrated evidence planning (IEP), particularly in response to the FDA’s AI Guidance and recent political developments such as President Trump's Executive Order, Delivering Most-Favored-Nation Prescription Drug Pricing to American Patients. These conversations highlighted the importance of flexibility and resilience in evidence strategies, preparing stakeholders for multiple potential regulatory futures. Accelerating Market Readiness with Integrated Evidence Planning discusses IEP as a proven method for this purpose, and we share detailed insights in the white paper, The Evolution and Future of Integrated Evidence Planning.

In summary, ISPOR 2025 was a reflective yet forward-looking event that acknowledged GenAI's current shortcomings while underscoring its substantial future promise. Attendees left with a deeper insight into the limitations and potential of AI, ready to strategically apply these insights to their ongoing research and development initiatives.

The Axtria team presented the following posters:

1. Real-world evidence (RWE) in health technology assessment (HTA): global guidelines on the use of RWE in evidence generation for HTA evaluation
2. Software as a Medical Device: Current Regulations in North America and Europe
3. The Cost Burden of Infertility Therapy for Women with a History of Hysteroscopic Adhesiolysis Compared to Those with Other or no Previous Intrauterine Surgery
4. A Retrospective Real-World Data Analysis of Pregnancy Outcomes Among Women with a History of Hysteroscopic Intrauterine Adhesiolysis (HA), Other Intrauterine Surgery (IUs), or no Prior Ius
5. The Role of Social Determinants of Health (SDoH) Data in Improving Risk Predictions
6. Integrating Artificial Intelligence (AI) and Machine Learning (ML) Techniques With Real-World Data (RWD) and Real-World Evidence (RWE) to Inform Precision Medicine: A Scoping Review
7. Predicting Future Type 1 Diabetes Onset Risk; A Machine Learning Approach

Won Chan Lee, Ph.D.
Principal, RWE & HEOR Practice Lead – Scientific and Strategic Oversight

David Hood
Associate Director, RWE & HEOR – Client Engagement Lead

Coby Martin
Senior Manger, RWE & HEOR – Client Engagement Lead

Alexandra Koumas
Manager, RWE & HEOR – Client Engagement Lead

 

Read our latest article on the evolution and future of integrated evidence planning: Lee, W. C., Blanchette, C., Pokras, S., Shaikh, J., & Miller, J. (2025). The evolution and future of integrated evidence planning. Expert Review of Pharmacoeconomics & Outcomes Research, 1–8. https://doi.org/10.1080/14737167.2025.2497876