Health Technology Assessment (HTA) and regulatory engagement remain a central component of SYNTHIA, helping to ensure that innovations in synthetic data are developed in a way that is relevant, acceptable, and useful for healthcare decision-makers.

To support this objective, SYNTHIA established an HTA and Regulatory Advisory Group comprising experts from HTA agencies, regulatory organisations, and healthcare decision-making bodies across Europe and beyond. The group's role is to provide independent advice and stakeholder perspectives throughout the project, helping to explore how synthetic data could be used in real-world HTA and regulatory settings.

On 24 November 2025, the HTA and Regulatory Advisory Group reconvened for its second meeting. Fifteen representatives from HTA and regulatory organisations participated alongside SYNTHIA partners representing the project's work packages and disease-specific use cases.

Building on discussions from the first meeting held in April 2025, the second forum focused on specific healthcare applications of synthetic data, approaches for assessing synthetic data quality, progress across the SYNTHIA use cases, synthetic control arm development, and regulatory considerations for medical devices developed using synthetic data.

Exploring healthcare applications of synthetic data

The meeting began with the presentation of a literature review examining healthcare applications of synthetic data that had been validated in published studies. The review categorised applications into three broad areas:

  • Replication – generating synthetic data to enable privacy-preserving sharing of data.
  • Prediction – using synthetic datasets to train diagnostic, prognostic, or classification algorithms.
  • Augmentation – using synthetic data to increase sample sizes or address biases in existing datasets.

Participants discussed which applications are most relevant to HTA and regulatory decision-making and where synthetic data could provide the greatest value.

Several promising areas were identified. Forum members highlighted the potential of synthetic data to support analyses of outcomes such as overall survival and progression-free survival, which are highly influential in regulatory and HTA assessments. Synthetic data could also facilitate access to data for early study planning and protocol development, improve transparency and reproducibility of analyses, and support collaboration in time-sensitive evidence generation environments.

The discussion also explored opportunities for synthetic data augmentation. Participants noted that synthetic data could help address evidence gaps in rare diseases, under-represented populations, and small patient subgroups, as well as support the development of more robust training datasets for AI models. Interest was also expressed in hybrid approaches where synthetic data is combined with real-world or clinical trial data, for example in hybrid or external control arm designs.

At the same time, participants highlighted several challenges that need to be addressed before synthetic data can be routinely used in HTA or regulatory settings. These included the need for greater transparency regarding data generation methods, clearer expectations around appraisal and validation, and stronger evidence demonstrating that synthetic data can reliably reproduce real-world decision outcomes.

Assessing trust through synthetic data validation

The forum was also introduced to the SYNTHIA SAFE framework, which has been developed to support a more consistent assessment of synthetic data quality. The framework considers privacy, utility, fidelity, and disease-specific requirements when evaluating synthetic datasets.

Discussions highlighted the wide range of available metrics for assessing synthetic data and the need for greater harmonisation in how these metrics are selected and interpreted. Participants emphasised the importance of providing guidance on which metrics are most meaningful for different use cases and datasets. A key message from the discussion was that the acceptability of synthetic data will depend heavily on its context and purpose. Different applications may require different levels of fidelity and validation, particularly when synthetic data contributes to high-impact decisions. Participants also stressed the importance of understanding privacy-preserving techniques as part of a broader trade-off between privacy protection and fidelity.

Overall, forum members considered frameworks such as SAFE to be an encouraging step towards creating a more structured and transparent approach for evaluating synthetic data.

Progress across SYNTHIA use cases

Representatives from the SYNTHIA disease-specific use cases provided updates on ongoing work and sought feedback from the forum on maintaining HTA and regulatory relevance.

Participants highlighted the importance of clearly stating the context and purpose of synthetic data use, including why traditional methods or existing data sources may not be sufficient. The need for transparent validation plans and protocols was emphasised, alongside consideration of how synthetic data generation methods could be validated across different datasets and research questions. The forum also encouraged the inclusion of clinical and patient expertise throughout the development process to ensure that synthetic data outputs remain relevant to real-world healthcare questions. In addition, participants recommended building on existing principles and standards developed for real-world data, particularly in relation to transparency, data suitability, and analytical methods.

Advancing synthetic control arm research

The meeting included a discussion on SYNTHIA's work exploring synthetic control arms and their potential use in HTA and regulatory environments.

Participants discussed the importance of clearly distinguishing between traditional external control arms and external control arms generated using synthetic data. The level of validation required, they noted, is likely to depend on the role synthetic data plays within an evidence package, whether as supplementary information, data augmentation, or a replacement for a control arm. Questions were also raised regarding access to sufficient patient-level data to support the development of reliable synthetic control arms, particularly in rare diseases and increasingly specialised treatment indications.

Regulatory considerations for synthetic data-enabled medical devices

The final discussion focused on regulatory processes for medical devices that are developed, tested, or validated using synthetic data. Using a hypothetical synthetic data-enabled medical device as an example, participants considered what evidence would be required to support its safe use. Key themes included the need to clearly explain why synthetic data is required, how it is being used, the representativeness of the underlying data, the methodology used to generate synthetic data, and the validation processes applied throughout development. Participants also highlighted the importance of comparing outcomes against approaches that do not use synthetic data and establishing appropriate post-market surveillance to monitor performance.

A recurring message was that confidence in synthetic data will depend on demonstrating fitness for purpose. Validation in a single scenario may not be sufficient, and evidence will be needed to show that synthetic data approaches remain reliable across different applications and contexts.

Key conclusions

The second HTA and Regulatory Advisory Group meeting reinforced that the future acceptance of synthetic data in HTA and regulatory settings will depend on clearly defining the context and purpose of its use. Participants agreed that discussions around validation, quality standards, and acceptability cannot be separated from the role synthetic data plays within an evidence package. Understanding the intended application is essential for determining what level of validation is required and how data quality should be assessed.
assessed.
The meeting highlighted the importance of:

  • Developing fit-for-purpose validation frameworks.
  • Building on lessons learned from real-world data.
  • Involving clinical and patient experts in use case development.
  • Continuing to explore where synthetic data can address important evidence gaps in healthcare decision-making.

The insights gathered during this second forum meeting will help guide the next phase of SYNTHIA's research and future discussions with HTA and regulatory stakeholders as the project continues to advance the responsible use of synthetic data in healthcare.

If you are interested to learn more please send an email to communications@ihi-synthia.eu and we will connect you with Michael Bell, key contact for SYNTHIA Partner NICE.