Accuracy of the AvoSure PT Pro System Compared with a Hospital Laboratory Standard

Abstract

OBJECTIVE: To compare international normalized ratio (INR) values obtained using the AvoSure PT Pro point-of-care (POC) system with those obtained using a standard laboratory method. METHODS: Forty-one INR values obtained from the POC system were compared with those obtained from a standard laboratory method. The POC method was evaluated for both laboratory and clinical agreement. To evaluate laboratory agreement, various analyses were used, including mean-squared prediction error (MSE) and mean prediction error (ME), Bland—Altman analysis, correlation, and paired t-test comparing group INR means. For clinical accuracy, discrepant pairs were identified and evaluated to determine whether dosage adjustments would have been needed based on values obtained. RESULTS: The POC system demonstrated modest precision (MSE = 0.147, 95% CI 0.065 to 0.228) and relatively little bias (ME = 0.090, 95% CI–0.025 to 0.205). Bland—Altman analysis also suggested good agreement at average INRs from 2.0 to 3.0. At average INR values >3.0, the POC system consistently overestimated INR. Values obtained with the POC system were significantly correlated with those obtained from the hospital laboratory (r = 0.77; p < 0.001). Similarly, mean ± SD POC INR did not differ significantly from the laboratory-determined INR (2.45 ± 0.59 vs. 2.37 ± 0.48, respectively; p = 0.176). Regarding clinical accuracy, the values clinically agreed in 85.4% of the cases. CONCLUSIONS: The AvoSure PT Pro POC system appears to be useful for INR values within the 2.0–3.0 range, but values outside of this range should probably be confirmed with a standard laboratory method.