A statistical and clinical evaluation of fingerstick and routine laboratory prothrombin time measurements

To compare prothrombin time measurements by fingerstick and routine laboratory methods.

Prospective cohort study.

University-affiliated anticoagulation clinic.

Thirty-three patients receiving warfarin with stable anticoagulation for 3 months preceding the two studies.

Groups 1 (17 patients) and 2 (16 patients) provided 150 and 125 paired samples, respectively, for fingerstick and routine laboratory analysis. The fact that no patient required a dosage change allowed for a clinical assessment.

Correlation and agreement between methods were good in group 1 but poor in group 2. Fingerstick results were less variable (smaller standard deviation and smaller coefficient of repeatability) in both groups. By analysis of discrepant pairs (25 in group 1, 63 in group 2), the routine laboratory results indicated dosage changes erroneously more often than did the fingerstick method.

In these two trials, the fingerstick system was superior to the routine laboratory method in that it was more reliable (less variability and more repeatable) and less likely to indicate dosage changes erroneously.