Artificial Intelligence Governance of Oncology Clinical Trial Matching Models Under Unmeasured Ethics Committee Quorum Capture: A Healthcare Political Fairness Estimation

Ethics committee quorum capture—the informal, unmeasured alignment of review-board voting blocs with the interests of sponsoring institutions or clinical departments—is a structural feature of oncology trial governance that current fairness audits of artificial intelligence matching models cannot detect. The argument proceeds in three steps. First, quorum capture is conceptualised as a form of institutional drift in which the composition of review boards, rather than the text of their rulings, determines which populations the matching model is permitted to see. Second, the paper shows that conventional fairness metrics applied to matching models are identified only up to the selection decisions that ethics committees authorise, so that unmeasured capture enters the estimand as a cluster-level confounder. Third, a replication-based power analysis characterises the operating characteristics of the before-after-control-impact estimator under a declared variance structure with an intraclass correlation of 0.64, between-cluster standard deviation of 6.0 and residual standard deviation of 4.5. Across 2000 replications per design cell, 54 clusters are required to detect a five-percentage-point effect at 80 per cent power. The contribution is a governance-measurement framework that treats ethics committee composition as a political variable requiring its own estimation strategy, rather than as a fixed institutional backdrop against which algorithmic fairness is assessed.