OBJECTIVE: The Bland-Altman method is often upheld as the optimal method to assess agreement between alternate measures of the same clinical parameter.  However, recent research by our group demonstrates the Bland-Altman method does not report agreement in a clinically meaningful way.  The objective was to determine if the Bland-Altman method distinguished between two point-of-care (POC) INR devices.  These devices were previously shown to have significantly different levels of agreement with our core laboratory. METHODS: In a previous experiment, 170 patients provided three separate INR measures at the same clinic visit – two by POC (Avosure™ and ProTime™ devices) and one venous sample analyzed at our core laboratory (considered the standard measure).  Agreement was achieved when the POC and lab INR values led to the same clinical decision.  Differences in agreement between the POC devices and laboratory were assessed by McNemar’s test.  In the current study, we applied the Bland-Altman method to determine if inferences regarding agreement between the POCs and laboratory were identical to the previous experiment where clinical decisions defined agreement. RESULTS:   The Avosure device was significantly more likely to lead to the same clinical decision as the laboratory versus the ProTime device (80% vs. 66%, respectively, p<0.001).  However, the Bland-Altman method produced virtually identical mean bias (0.4 and 0.5 INR units, respectively) and did not distinguish between the devices.  Statistical analysis of the Bland-Altman method produced the same findings for each device: significantly different standard deviations between the POC and the laboratory (p<0.001), significant bias in each device (p<0.001), and high correlations between the POCs and the laboratory (0.925 and 0.926, respectively). CONCLUSIONS:   The Bland-Altman method did not detect clinically important differences between the POC INR devices.  Clinically meaningful agreement between measures of INR is optimally assessed by a method that directly observes or explicitly estimates clinical decisions.