We prospectively evaluated a capillary whole blood prothrombin time (PT) monitor (Biotrack, Ciba Corning) in an outpatient pediatric anticoagulation clinic (40 clinic patients) and in age-matched healthy subjects (30 control subjects). Subsequently, 23 children requiring warfarin therapy were placed on a home program (home patients) using the PT monitor; their parents were trained and the results followed by clinic staff. The PT results were reported as internationalized normalized ratios (INRs). The laboratory and PT-monitor INR values were similar for the clinic patients and the control subjects (y = 0.76x + 0.38; r = 0.93; p < 0.001). The accuracy of the PT monitor (the difference between INR values and the laboratory INR) was best at an INR of 2.5 to 3.5; 90% of paired INR values were within 0.8 INR units. The average duration of monitoring for home patients was 13 months (range, 2 to 60 months). They had an average of 3 dose measurements (range, 1 to 11 measurements) and 1.8 dose changes (range, 0.6 to 4.5 changes) per month. Of the 599 measurements, 63% were within the therapeutic range, similar to those for clinic patients; the dose requirements were also similar. There was 1 significant bleeding event, a subdural hematoma in a patient with an INR of 4.1, and 1 catheter-related thrombotic event with an INR of 1.2; both children recovered. Of the 23 families, one discontinued home monitoring because of parental discomfort, 2 children died of their primary disease, 6 completed warfarin therapy, and 14 remain on the home program. We conclude that the whole blood PT/INR monitor is safe and offers practical advantages to children requiring anticoagulation.