Q&A with Jeroen Jansen: Chief Scientist of HEOR at Precision AQ

Jeroen Jansen, PhD, is an internationally recognized thought leader and scholar in Health Economics & Outcomes Research (HEOR). In this Q&A, he summarizes his role at Precision AQ, explains what he would change about how industry-sponsored HEOR studies are approached, and shares his view on how AI can benefit HEOR.

1. Who are you and what do you focus on professionally?  

Chief Scientist of HEOR, Executive Vice President, at Precision AQ.

My work sits at the intersection of statistics, evidence synthesis, and health economics. With 25 years of research and consulting experience, I help stakeholders understand the clinical and economic value of novel health technologies to support drug development, market access, and health technology assessment (HTA).

2. What unique perspective or expertise do you bring to your team?  

I work across the HEOR space, but with a specific focus on quantitative evidence synthesis and health economic and simulation modeling.

At Precision AQ, I support the HEOR teams with business development and project execution. This support includes brainstorming and pitching project ideas to clients, and providing guidance on study design, analysis, and interpretation of results. It also extends to hands-on contributions such as programming and analytical support, particularly for novel project types. Furthermore, I am responsible for thought leadership, publishing innovative methodological solutions that address challenging HEOR problems to showcase our expertise.

3. If you could change one industry norm or habit that slows progress in HEOR, what would it be, and what would a better alternative look like? 

Rather than rigidly adhering to labels for project types, I believe we should focus on the research question itself and how to obtain the most relevant answer in the most credible manner.

For example, in evidence synthesis, when the aim is to compare treatment effects of alternative interventions, we have established methods such as network meta-analysis (NMA), matching-adjusted indirect comparisons (MAIC), and multilevel network meta-regression (ML-NMR). However, we need to recognize that these methods are, in essence, specific approaches along a theoretical continuum, each relying on different assumptions. If the standard or established statistical methods work, that’s great—but if they don’t, we should not hesitate to modify or adapt them to fit the specific research question and evidence base at hand.

Now, I recognize that our typical evidence synthesis study is designed to inform HTA decisions, and using an established method means one less challenge for our clients to worry about. However, if its assumptions are violated in the available evidence base, the perceived benefit of using a standard method becomes a disadvantage. In that case, a bespoke approach may be preferable. One way to support its acceptance is to publish the method separately, applied to a different example, so that reviewers can judge the approach on its own merits before it is used in the actual analysis.

In essence, I want to advocate for moving from “I need an NMA or ML-NMR for my drug” to “I want to compare the efficacy of my drug with competing interventions based on the relevant evidence using a rigorous and acceptable methodological approach.” Think of it like an ice cream shop: a client may only know vanilla, strawberry, and chocolate (the known methods), but there are many other flavors, and they may find one that suits them much better.

4. What forward-looking trend do you believe will most influence HEOR practice over the next 3–5 years?

The answer is clearly AI. Yes, it will replace certain activities and automate work streams. But what excites me most are the following opportunities:

Using AI to translate the evidence that HEOR generates into relevant information for different stakeholders.

HEOR has always been perceived as niche and overly technical. We produce methodologically sophisticated work, yet as a community we struggle to communicate why these approaches are necessary to arrive at relevant and credible answers. As a result, many studies do not achieve the impact they deserve.

Using generative AI as a “translator” has the potential to change that. For instance, generative AI can take the output of a complex network meta-analysis or health economic simulation model and render it into clear, tailored narratives for different audiences, whether that is a payer evaluating formulary decisions, a clinician weighing treatment options, or a patient advocacy group seeking accessible evidence summaries.

Rather than expecting every stakeholder to interpret forest plots or cost-effectiveness planes, AI can distill the key findings into language and framing that resonate with each audience, ultimately making our work more actionable and impactful.

Using AI to help HTA decision-makers more easily understand and therefore accept novel or non-standard study approaches.

Generative AI can help decision-makers understand and engage with non-standard methods or approaches. For example, Excel has long been the dominant platform for health economic models, and we have been advocating for R-based models to improve transparency and reproducibility. Yet the uptake of R-based models remains limited, partly because of perceived reluctance or limited familiarity among HTA decision-makers — rightly or wrongly.

Generative AI can bridge that gap by making code-based models more accessible and interpretable for reviewers and decision-makers. Concretely, AI can generate plain-language explanations of complex analytical code and methods, produce interactive documentation that walks reviewers through each step of an analysis, and even allow them to query the underlying approach in natural language, for example, asking “Why were certain studies excluded from the evidence synthesis?” or “What happens to the incremental cost-effectiveness ratio if I add this additional health state to the model?” and receiving both the result and an intuitive explanation.

Whether the task involves a novel evidence synthesis method or a code-based health economic model, this approach lowers the technical barrier for HTA reviewers and builds confidence in non-standard approaches, which ultimately accelerates acceptance of more transparent, reproducible methods across the HEOR space.

Enabling novel approaches that are only feasible given the speed and power of generative AI.

One example I’m currently developing is a generative AI-based evidence synthesis approach designed to handle extremely diverse studies and distill the findings into relevant information, either for decision-makers directly or as inputs for modeling studies. I look forward to sharing this work in the future.

Meet Jeroen at ISPOR Europe: November 8-11, 2026

Interested in finding out more about Precision AQ’s HEOR capabilities and how Jeroen can support your next project? Schedule time to meet with him at ISPOR Europe 2026 in Vienna, Austria.