Artificial Intelligence replication of District Hospital Cervical Cancer Screen-and-Treat Queue Prioritisation Under Unmeasured Visual Inspection Acetic Acid Interpretation Discordance: A Healthcare Fairness-Robustness Evaluation

Visual inspection with acetic acid (VIA) is the pragmatic cervical cancer screening modality in many district hospitals, yet its interpretation is subjective and inter-observer discordance is rarely measured in routine queues. Because no field data on discordance are available, the study is framed as a replication-based robustness evaluation: a declared data-generating process is used to represent a district-hospital screening cohort, and the behaviour of an ordinary least squares estimator of the prioritisation score is characterised across 600 replications of N = 500 observations under a fixed seed (2054533309). The estimator recovered the intercept with a bias of 0.01188 and an RMSE of 0.85132, and the exposure coefficient with a bias of -0.0003 and an RMSE of 0.01558; empirical coverage of the nominal 95% confidence intervals was 1.0 for both parameters. These results establish that, under the stated process, the linear prioritisation score is estimated without material distortion, and they isolate the conditions under which discordance-driven label noise, rather than estimator instability, becomes the dominant threat to fairness.