Are national cancer indicators decision-grade for remote regions? A stress-test using Corsica (France)

Publicly available cancer indicators are increasingly used to guide regional health-care planning and investment. Their robustness for decision-making in remote or insular regions remains insufficiently examined.

We analysed publicly available incidence, mortality, standardised incidence ratios, standardised mortality ratios, mortality-to-incidence ratios (MIR), premature mortality and temporal trends using Corsica as a stress test and comparing with other French regions using national datasets. Indicators were compared using confidence intervals, temporal aggregation and rank-stability analyses. No causal inference was attempted.

Across multiple indicators, Corsica's relative position varied substantially depending on indicator choice, time window and statistical representation. Several indicators showed wide confidence-interval overlap with most French regions and marked rank instability, despite large apparent position changes. MIR and trend-based indicators were particularly sensitive to temporal aggregation and data structure.

Our findings suggest that commonly used national cancer indicators are frequently insufficiently robust to support decision-grade regional planning in remote settings. Apparent "good" or "poor" performance in such contexts can reflect statistical instability rather than meaningful differences.

For remote regions, cancer indicators should be interpreted with particular caution. Routinely published cancer indicators are insufficiently robust to support decision-grade regional health-care planning in small or insular populations.