Reflections from the GetReal Institute Conference 2026 in Utrecht on decision-ready evidence, uncertainty, and why health technology assessment needs to evolve alongside the technologies it evaluates.
In June, I had the opportunity to participate in the GetReal Institute Conference in Utrecht, where researchers, regulators, HTA bodies, industry representatives, and other stakeholders came together to discuss how real-world data can be translated into evidence that supports healthcare decisions.
I particularly enjoyed the opportunity to bring my perspective from the work we have been doing on Lifecycle and Dynamic HTA.
The conference theme, Translating Real-World Data into Decision-Ready Evidence, captured a challenge that I have encountered repeatedly throughout my career: generating evidence is one thing; generating evidence that can actually inform a decision is something quite different.
And it made me think about a question that goes beyond the methods we use to produce real-world evidence.
Why do we continue to organize so much of health technology assessment around individual decisions when the evidence, technologies, and clinical context rarely stand still?
One evidence package, different questions
One of the recurring challenges in healthcare evidence generation is the assumption that a sufficiently comprehensive evidence package should be able to satisfy everyone.
In reality, regulators, HTA bodies, payers, clinicians, and patients often need to answer different questions.
A regulator may need to determine whether the evidence supports a favorable benefit-risk balance. An HTA body may need to understand comparative clinical effectiveness against relevant alternatives. Healthcare systems may also need to consider how a technology performs within existing care pathways.
Real-world evidence can contribute to these questions, but its usefulness depends on the decision context.
A dataset may be extensive and technically sophisticated without being appropriate for a particular assessment. Conversely, carefully designed evidence using more limited data may be highly informative when the research question, population, comparator, and analytical approach are well aligned.
This is why I find the concept of fitness for purpose particularly important.
Before asking what we can do with available data, perhaps we should spend more time asking what decision needs to be made, what uncertainty matters for that decision, and what evidence could realistically help address it.
Uncertainty does not disappear after an HTA assessment
Much of the discussion around RWE focuses on its potential to address evidence gaps at the time of regulatory or HTA submission.
This is particularly relevant for technologies supported by single-arm trials, small populations, immature outcomes, or limited comparative evidence.
Methods such as external controls and target trial emulation can provide important opportunities, although their credibility depends on the underlying data, assumptions, and methodological choices.
But I think there is another question worth asking.
What happens to the uncertainties that remain after an assessment is completed?
An HTA body may recognize limitations in the available evidence and still need to reach a conclusion. A healthcare system may introduce a technology while expecting further evidence to emerge.
Over time, we may learn more about long-term effectiveness, safety, treatment patterns, patient outcomes, or performance in populations that were underrepresented in clinical trials.
Meanwhile, the standard of care may change. New comparators may become available. The technology itself may evolve.
The original assessment was not necessarily wrong. It was conducted using the evidence and decision context available at that moment.
The problem arises when we treat that assessment as though its underlying assumptions will remain valid indefinitely.
From evidence generation to continuous learning
This is where I see an important connection between RWE and the work we have been advancing on Lifecycle HTA.
Lifecycle HTA recognizes that assessment activities need to be considered across the evolution of a health technology and its evidence base, rather than as isolated exercises.
Dynamic HTA brings attention to the need for assessment and decision-making processes to respond when relevant evidence or circumstances change.
These ideas are closely related to the opportunities offered by real-world data.
If we can generate evidence throughout the use of a technology, we should also consider how that evidence might inform subsequent assessments.
But continuous evidence generation does not automatically create continuous learning.
We need to know which uncertainties justify further investigation, when new evidence should trigger reassessment, who is responsible for initiating that process, and how the findings will be incorporated into decisions.
Otherwise, we risk generating increasingly large amounts of data without creating a clear pathway for that information to influence healthcare practice or policy.
Not everything needs to be reassessed continuously
I also think we need to be pragmatic.
It would be neither feasible nor desirable to reassess every technology whenever a new study is published or another dataset becomes available.
HTA organizations operate with limited resources, and healthcare systems must balance many competing priorities.
The challenge is therefore not to create a permanently open assessment process for everything.
It is to identify where changes in evidence, technology, clinical practice, or the decision context are sufficiently important to justify revisiting an earlier assessment.
This requires clear triggers, proportionate methods, defined responsibilities, and an understanding of the consequences of maintaining—or changing—a previous conclusion.
It also requires closer dialogue between those who generate evidence and those who use it.
For me, that is one of the most valuable aspects of forums such as GetReal. They bring different stakeholders into the same conversation, allowing us to examine not only whether a method is scientifically credible, but whether the resulting evidence can address the questions decision-makers actually face.
What would make RWE genuinely decision-ready?
I left Utrecht reflecting on how much progress has been made in real-world evidence methods, data infrastructure, and the recognition of RWE across regulatory and HTA settings.
Yet methodological progress is only part of the picture.
We also need to think about how healthcare institutions are organized to receive and act on new evidence.
A well-designed real-world study may reduce uncertainty, but its value for decision-making depends on whether there is a process through which its findings can be considered.
That means moving beyond the idea that the purpose of evidence generation is simply to complete a submission or support an initial assessment.
Evidence should remain relevant as technologies move through their lifecycle and as healthcare systems learn from their use.
Perhaps the next challenge for RWE is not only demonstrating that we can generate decision-ready evidence, but ensuring that our decision-making processes are ready to use evidence as it evolves.
Thank you to the GetReal Institute, the organizers, fellow speakers, and participants for the discussions in Utrecht. They reinforced why I believe the future of HTA will depend as much on adapting our processes and governance as on advancing the methods we use to generate evidence.
Further reading