Screening computer‐assisted dosage programs for anticoagulation with warfarin and other vitamin K antagonists: minimum safety requirements for individual programs

Abstract

Based on the results of the previous European Action on Anticoagulation (EAA) multicenter study, a simplified minimum procedure is described for screening the safety and effectiveness of marketed programs for dosage of oral anticoagulant drugs (vitamin K antagonists). The aim was to demonstrate non-inferiority to the manual dosage at the experienced centers in the EAA study. With the use of a cluster sampling procedure, a minimum number of centers and a minimum total of patients required to establish non-inferiority were determined. At least four centers, each recruiting 50 patients over a period of 6 months, were shown to be required (excluding the results from the first 3 weeks of treatment). To achieve non-inferiority, the lower 95% confidence interval of the 'time in target International Normalized Ratio range' (TIR) of a marketed program must be above the TIR limit set by the manual dosage group in the EAA study, that is, 57.5%. The simplified procedure proposed, although not an absolute guide to safety, is designed to screen against gross unreliability of a test program, without the need to repeat a massive clinical endpoint study for each and every program