Identifying cancer in the French National Health Data System (SNDS): an updated scoping review of algorithms, validation and applications.

[BACKGROUND] Identifying cancer cases in the French National Health Data System (SNDS) is essential for real-world oncology research, yet operational definitions remain heterogeneous and few algorithms have been validated. An initial ReDSiam review in 2017 mapped early practices. Since then, the use of SNDS data in oncology has expanded considerably, warranting an updated synthesis.

[METHODS] We conducted a scoping review following the PRISMA-ScR guidelines. A comprehensive MEDLINE search identified peer-reviewed studies using SNDS-based algorithms to detect cancer cases up to November 2024. A structured extraction grid captured algorithm characteristics, study applications and validation procedures.

[RESULTS] Among 233 included studies, publication volume increased sharply over time. Algorithms showed substantial heterogeneity in operational definition. Most studies focused on incident cancer detection using hospital data, while outpatient and long-term illness data were increasingly incorporated. Only 6% of studies reported validation against a gold standard, and validated algorithms remained limited to a few cancer sites. More recent publications more frequently provided code lists and supplementary materials, improving transparency.

[CONCLUSIONS] The use of the SNDS in oncology has expanded substantially, but validated and standardised case-identification algorithms remain scarce. Strengthening validation efforts, improving access to reference data, ensuring regular algorithm updates and fostering collaboration between data producers, methodologists and clinicians are critical to enhance methodological consistency and reproducibility.