Artificial Intelligence Governance of Hospital Bed Surge Allocation Models Under Unmeasured Provincial Political Interference in Census Updating: A Healthcare Political Fairness Estimation

Hospital bed surge allocation models increasingly rely on census updates that provincial political authorities can influence before the data reach the model. the present analysis treats that influence as an unmeasured confounder in the before-after-control-impact design that governance evaluations of such models typically employ, and characterises the operating characteristics of that design under a declared variance structure. The analysis is a replication-based power evaluation, not an observational study: no hospital, province or respondent data are used or claimed. With a between-cluster standard deviation of 6.0, a residual standard deviation of 4.5 and an intraclass correlation of 0.640, the design was evaluated across cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points, with 2000 replications per cell. Power rose with cluster count and effect size, and 56 clusters were required to detect a 5 percentage point effect at 80 per cent power. The fairness of an artificial intelligence allocation model cannot be assessed independently of the political process that updates its census inputs, and that governance evaluations should report the intraclass correlation and the minimum detectable effect alongside any fairness claim. The contribution is a design-level account of what a fairness estimate can and cannot identify when provincial interference is unmeasured.