Delivered clearance and timed concentrations for methanol dialysis duration in a simulation study

Abstract Background Methanol dialysis-duration estimates may benefit from observed delivery and timed concentrations, but their incremental value over conservative fixed prescriptions is uncertain. Methods We compared 17 policies in 5,000 sessions per scenario across 18 scenarios. Latent truth was separated from observable inputs. Fixed comparators included a seven-target frontier and constant extensions calibrated on separate timing simulations. HYBRID-95 required both nominal integral and two-level criteria. Exploratory additions examined duration tails, assay delays and 20,000 additional primary sessions with unchanged calibration constants. Results In the original primary sample, HYBRID-95 produced 1/4,999 threshold exceedances (≥ 20 mg/dL; 0.02%), versus 1.14% for its calibrated fixed timer, both at median 8.25 hours. Hybrid minus fixed paired mean duration was 9.5 minutes; completed-session 95th percentiles were 13.25 versus 11.50 hours. The additional sample produced 2/20,000 hybrid exceedances (0.010%; Wilson interval 0.003–0.036%), versus 1.32% for the fixed timer, again at equal medians. Combined mapping-volume bias and later delivery loss caused 24.85% hybrid exceedance and 54 additional horizon overruns in the original stress scenario. Conclusions Combining observations improved simulated concentration outcomes at comparable median duration, with additional mean and upper-tail treatment burden. Systematic errors remained failure modes. Empirical calibration and independent review are required before clinical evaluation.