William Rhine

Publication Details

  • Anticoagulation therapy advisor: a decision-support system for heparin therapy during ECMO. Proceedings / the ... Annual Symposium on Computer Application [sic] in Medical Care. Symposium on Computer Applications in Medical Care Peverini, R. L., Sale, M., Rhine, W. D., Fagan, L. M., Lenert, L. A. 1992: 567-571


    We present a case study describing our development of a mathematical model to control a clinical parameter in a patient--in this case, the degree of anticoagulation during extracorporeal membrane oxygenation (ECMO) support. During ECMO therapy, an anticoagulant agent (heparin) is administered to prevent thrombosis. Under- or over-coagulation can have grave consequences. To improve control of anticoagulation, we developed a pharmacokinetic-pharmacodynamic (PK-PD) model that predicts activated clotting times (ACT) using the NONMEM program. We then integrated this model into a decision-support system, and validated it with an independent data set. The population model had a mean absolute error of prediction for ACT values of 33.5 seconds, with a mean bias in estimation of -14.3 seconds. Individualization of model-parameter estimates using nonlinear regression improved the absolute error prediction to 25.5 seconds, and lowered the mean bias to -3.1 seconds. The PK-PD model is coupled with software for heuristic interpretation of model results to provide a complete environment for the management of anticoagulation.

    View details for PubMedID 1482937

Stanford Medicine Resources:

Footer Links: