Artificial Intelligence Governance of Community Health Worker Diagnostic Triage Models Under Unmeasured Senior Nurse Override Authority: A Healthcare Political Robustness Evaluation

Clinical decision support for community health worker (CHW) diagnostic triage is typically governed as a technical problem of model accuracy, yet the effective decision rule is a political settlement between the model and the senior nurse who may override it. It treats unmeasured senior nurse override authority as a structural feature of the governance arrangement rather than a data defect, and asks how the before-after-control-impact design used to evaluate triage models behaves when override authority is present but unrecorded. The analysis characterises the operating characteristics of that design under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64) across 2000 replications per design cell, with cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The design requires 54 clusters to detect a 5 percentage point effect at 80 per cent power. Override authority is a governance variable that the evaluation design cannot see, and that this invisibility, not model error, is the principal threat to the political robustness of CHW triage governance.