Federated analytics for non-communicable disease surveillance in the European health data space: a scoping review and...

The growing burden of non-communicable diseases (NCDs) in Europe has intensified the need for timely, comparable, and policy-relevant health indicators derived from increasingly heterogeneous health data ecosystems. The European Health Data Space (EHDS) represents a major policy initiative to facilitate the secondary use of health data while preserving privacy, security, and national data sovereignty. In this context, federated analytics can foster international comparisons without requiring the transfer of person-level data.

This scoping review aimed to map federated analytical approaches relevant to the production of NCD indicators within the EHDS and to develop a conceptual framework linking distributed statistical methods, privacy-preserving infrastructures, and policy-oriented surveillance requirements.

A scoping review was conducted following the PRISMA Extension for Scoping Reviews. Searches were structured into three complementary conceptual domains: (a) federated analytical approaches, (b) distributed statistical inference methods, and (c) governance and health-data infrastructures relevant to the EHDS. Searches were performed in PubMed, Scopus, and IEEE Xplore, for studies published between 2010-26. Records were exported with abstracts and full bibliographic metadata. Deduplication and metadata-aware merging were conducted across databases using DOI and normalized-title matching.

We retrieved a total of 2,362 records, of which 1,285 were unique records and 104 were finally retained. The literature revealed a heterogeneous but rapidly expanding ecosystem of distributed analytical approaches. We identified three major domains: (1) distributed regression; (2) federated or distributed analytical infrastructures; and (3) privacy-preserving federated epidemiological analysis.

An increasing range of solutions for federated analytics is available for policy-grade NCD indicators. Longitudinal and survival models remain methodologically complex because of covariance structures and globally coupled risk sets. A modular approach is needed to incorporate public-health intelligence and a dynamic set of interoperable tools in the EHDS, with the active support of a decentralized network of active stakeholders.

Federated analytics is shifting the paradigm from centralized data pooling to computation-to-data architectures. The successful implementation of NCD surveillance in the EHDS require integrating statistics with legal, IT, policy and governance mechanisms. The review outlined a conceptual framework that can couple technical innovation with the best architecture for public-health knowledge production.