After a few calmer summer weeks, the legislative work on the European Health Data Space (EHDS) is back to running at full steam towards adopting new policy with the potential to reshape the landscape for medtech innovation. The proposed secondary use policies outlined by EHDS will allow medtech companies to accelerate the development of artificial intelligence (AI) tools, opening up significant opportunities for new treatments and assistive technology for clinicians and patients.
In the area of surgical AI, the medtech industry has been hampered by the scarcity of large datasets available to researchers. The potential applications of data are vast, ranging from assessment of a surgeon’s training progress to providing critical intraoperative guidance. However, to realise these applications, medtech needs access to clean, comprehensive, and diverse datasets.
A fit-for-purpose legislative framework requires policy to stay closely tied to clinical use cases. The real-world application of using AI to guide surgeon training can transform the surgeon’s pathway from a novice level to expert. Rather than rely on potentially misleading sources such as a supervisor’s opinion or the surgeon’s own self-confidence, AI can analyse surgical videos to assess skill acquisition, safety, and judgment. By comparing data across surgeons, benchmarks tied to patient outcomes can be established, and surgeons can be provided and measured against a structured and personalised training roadmap.
A second application to consider is the power AI can bring to guide surgeon decision-making by offering assistance based on collective knowledge and prior outcomes. When a surgeon encounters a challenging intraoperative situation for the first time, such as a complication or atypical anatomy, AI provides the opportunity to easily leverage the expertise of others who have encountered the situation before.
Currently, gathering the data to realise these example applications is not straightforward. Navigating and gathering patient consent and then cleaning data sources is resource-intensive work. Aggregating data from multiple regions introduces further challenges of local legal considerations for data use and export, and standardisation of data structure and units. The EHDS regulation and future work of the Health Data Access Bodies could help address these concerns.
Another significant advantage where the EHDS could help advance in personalised care is to provide a framework for access to pseudonymised data when the use case requires it. This will allow the Health Data Access Bodies to incorporate novel data types and offer longitudinal association of data, important for the use cases above and comprehensive views for each patient or surgeon.
Some challenges remain in understanding the implementation of the policy. For instance, the EHDS should bring more clarity ensuring companies remain incentivised to drive forward innovation and new tools for patient care.
The future of surgical AI is undoubtedly brighter with the EHDS proposal. Policymakers, healthcare, and industry partners will need to strike a balance between innovation and privacy as well as transparency and protection. All in all, the opportunities to transform the development and adoption of AI tools in medtech abound.