Artificial Intelligence replication of District Hospital Bed Occupancy Forecasting Under Unmeasured Seasonal Migration Reporting Gaps: A Healthcare Robustness Evaluation

District hospital bed occupancy forecasts are routinely produced from admission records that omit seasonal migrant patients, yet the consequences of this reporting gap for forecast robustness have not been formally characterised. the present analysis treats the gap as a structural measurement problem and evaluates, through a replication-based replication, how an ordinary least squares estimator of occupancy behaves when the exposure variable is observed with a seasonal, group-dependent omission. Six hundred data sets of 500 observations each were drawn from a declared data-generating process with intercept 12.0, exposure 0.42, group -3.1, severity 1.35 and error standard deviation 4.2, and the estimator was applied to each. The intercept was estimated with a bias of 0.0422 and a root mean squared error of 0.86006; the exposure coefficient was estimated with a bias of -0.00057 and a root mean squared error of 0.01479; nominal 95% confidence intervals covered the true parameter in every replication for both coefficients. The sampling distribution of the exposure coefficient was approximately normal and its normal Q-Q plot was close to the identity line. A receiver operating characteristic curve constructed from the same replications located the discrimination threshold well above chance. The results establish that under the stated omission mechanism the estimator remains consistent and interval coverage is not degraded, and they delimit the conditions under which this robustness can be expected to hold. The contribution is a formal statement of when reporting gaps of this kind can be tolerated in district hospital forecasting and when they cannot.