Key result
Distributionally robust triage acuity assignment requires 54 clusters to detect a ~5-point effect.
Why the study?
Point-of-care troponin assays used in emergency department triage undergo variable calibration drift, causing acuity assignment rules that treat the assay as fixed to systematically misclassify patients near the decision threshold.
Does a distributionally robust optimisation triage acuity assignment rule improve worst-case patient safety under point-of-care troponin assay calibration drift compared to a nominal plug-in rule in a simulated emergency department triage setting?
Does a distributionally robust optimisation triage acuity assignment rule improve worst-case patient safety under point-of-care troponin assay calibration drift compared to a nominal plug-in rule in a simulated emergency department triage setting?
This computational modeling study demonstrates how distributionally robust optimisation can address systematic patient misclassification caused by point-of-care troponin calibration drift during emergency department triage.
Calibration drift may cause systematic misclassification near troponin thresholds in ED triage; leaves open whether lot-specific adjustments improve safety.
Point-of-care troponin assays used in emergency department triage are subject to calibration drift that varies across reagent lots, device generations and operator conditions, so that the same measured concentration can correspond to different true concentrations at different sites and times. Acuity assignment rules that treat the assay as fixed therefore misclassify patients in a way that is systematic rather than random, and the resulting safety loss is concentrated in the patients nearest the decision threshold. the present analysis formulates triage acuity assignment as a distributionally robust optimisation problem over an ambiguity set of calibration-drift distributions, and derives a tractable convex reformulation whose conservatism is governed by a single radius parameter. The framework is evaluated in a computed healthcare misclassification study in which the operating characteristics of the robust assignment rule are compared with those of a nominal plug-in rule across a declared design grid. The design uses a before-after-control-impact structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64; power was estimated across 2000 design cells with 2000 replications per cell, and 54 clusters were required to detect a 5 percentage-point effect at 80% power. The analysis characterises how the robust rule trades nominal accuracy for worst-case safety, and identifies the drift radius beyond which conservatism dominates any achievable safety gain.
No takes yet. Share an insight, caveat, or question.
Naa Lamptey Quaye (2026) studied Emergency department triage acuity assignment using point-of-care troponin assays. Distributionally robust optimisation triage acuity assignment rule vs. Nominal plug-in rule was evaluated on Healthcare misclassification safety and nominal accuracy. Distributionally robust triage acuity assignment vs a nominal plug-in rule required 54 clusters to detect a 5 percentage-point effect at 80% power (ICC 0.64) across 2000 design cells.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: