The accuracy and reproducibility of the CoaguChek S, and its clinical agreement with conventional laboratory international normalized ratio (INR) determination, were evaluated in an outpatient anticoagulation clinic setting. Forty-three patients provided 248 paired INR measurements for analysis. The paired results were highly correlated (r = 0.90). The mean coefficient of variation for the CoaguChek S for a random sample of 21 patients with three repeated tests each, was 4%. Clinical applicability was also measured by discrepant INR values, as defined in the literature by expanded and narrow agreement, and by INR values resulting in a different clinical decision by a blinded haematology registrar. Expanded agreement and narrow agreement between the two INR values occurred 90 and 88% of the time, respectively. The stricter criteria set down by the clinician resulted in 73% of paired results producing the same dosage decision. The CoaguChek S displayed good correlation with laboratory determination of INR and compared relatively well with expanded and narrow clinical agreement criteria.