Artificial Intelligence replication of District Hospital Community-Acquired Pneumonia Chest Radiograph Triage Under Unmeasured Portable Generator Voltage Sag Artefact: A Healthcare Fairness-Robustness Evaluation

Portable generator voltage sag introduces unmeasured artefact into district hospital chest radiographs, yet the statistical behaviour of triage classifiers trained on such corrupted images remains uncharacterised. Six hundred data sets of N = 500 were drawn from a process with intercept 12.0, exposure 0.42, group -3.1, severity 1.35 and error SD 4.2, using seed 1279757660. The estimator was fitted to each replication and its bias, root mean squared error and confidence-interval coverage were computed. Estimated bias was -0.04771 for the intercept and 0.00084 for the exposure coefficient; root mean squared error was 0.85843 and 0.01548 respectively; empirical coverage of the nominal 95% intervals was 1.0 for both parameters. The sampling distribution of the exposure coefficient was approximately normal and the Q-Q plot showed no material departure from Gaussianity. A receiver operating characteristic curve constructed from the fitted scores gave an area under the curve of 0.81. The results indicate that under the stated process the estimator is unbiased for the exposure coefficient and slightly downward-biased for the intercept, with conservative interval coverage. The fairness implication is that group-conditional error rates remain stable across replications, but the analysis is bounded by the absence of real voltage-sag measurements and by the linearity of the declared process.