Variability in the Calculation of Time in Therapeutic Range for the Quality Control Measurement of Warfarin

Time in therapeutic range (TTR), a well-recognized performance metric of oral anticoagulation, measures the time when patients’ international normalized ratios (INRs) are within the desired range. The TTR value can vary significantly depending on the type of method used and can be a skewed indicator of the overall quality of anticoagulation. As such, the present study was designed to compare three methods for TTR calculation (cross-sectional, traditional, and Rosendaal) to quantify their differences, biases, and trends. As part of this investigation, a 21-week retrospective analysis of patients on warfarin was conducted to compare TTR values obtained by these three methods. Paired t-tests, correlation studies between size and bias, and Bland–Altman plots were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). It was revealed that the TTR values for the cross-sectional, Rosendaal, and traditional methods were 65.97, 58.12, and 51.55, respectively. The addition of tolerances to INR ranges of ± 0.2 and ± 0.5 increased TTR values to 81.79 and 91.53, respectively, for the cross-sectional method, and 66.86 and 82.69, respectively, for the traditional method. The use of the traditional method resulted in significantly higher TTR values than did use of the Rosendaal method, with high variability between the methods in both positive and negative directions. There was a demonstrated lack of independence between the methods, and zero bias could not be assumed. In conclusion, the different methods considered in the present study do not accurately measure whether a patient is in or out of the therapeutic range, and the addition of tolerances can further distort the perception of anticoagulation achieved. We recommend a standardized TTR calculation method as well as a uniform tolerance for use in clinical trials and quality control efforts.