We have taken a stepwise approach to improving the dosing of continuous intravenous heparin in patients with acute coronary syndromes. Our primary objective was to use computer modeling to develop a nomogram for managing heparin therapy and to put in place a continuous quality monitoring system to evaluate the nomogram's effectiveness. We prospectively collected data on 41 patients with unstable angina or myocardial infarction who were treated with heparin. Their response to heparin was computer modeled and the dose to achieve an activated partial thromboplastin time (aPTT) ratio of 2.0 was established. This dose was regressed against all demographic characteristics to establish predictors of heparin dose (phase I). The regression formula was used prospectively in 110 patients to initiate the infusion rate of heparin and a bolus dose to achieve an aPTT ratio of 2.5. Subsequent dosage adjustments were achieved by computer modeling the patient's aPTT response (phase II). A nomogram was developed that simulated the decisions achieved using computer-assisted methods. This was retrospectively tested and then prospectively tested in 50 patients using nursing staff (phase IV). The nomogram was then made generally available (phase IV) and has been tested in an additional 310 patients. Phase I: Of the original 41 patients, 32% of the aPTT ratios were in the therapeutic range, 36% were supratherapeutic, and 32% were subtherapeutic after the first 24 hours. Phases II and III resulted in 85% of the aPTT ratios between 1.5 and 2.5 at 24 hours. Phase 4 had similar results in 310 patients. The use of computer-assisted or a computer-generated nomogram to adjust heparin therapy results in better control of heparin therapy than using standard methods.