Fig. 2
Results of method comparison between the POCT and two automated analyzers. The point-of-care analyzer was compared to a CRP assay run on two automated large bench top analyzers (ABX Pentra 400, AU 5800). A1–3: Passing- Bablok regression analysis with 95% confidence interval (CI) of CRP measurements performed with the three analyzers. A1: 95% CI of slope and intercept of the regression line were 0.74 to 0.90 and − 4.52 to 4.10, respectively. A2: 95% CI of slope and intercept were 0.86 to 0.95 and − 2.15 to 2.25, respectively. A3: There was a 95% CI of slope and intercept of 0.85 to 0.95 and − 2.75 to 2.02, respectively. B1–3: Bland-Altman difference plot demonstrating mean %Biasobs with its 95% confidence interval and its 1.96fold standard deviation (SD) indicative of its limits of agreement. The minimally acceptable total allowable error (TEmin) of 44.4% [32] represents the quality limit. C1–3: Bland-Altman difference plot demonstrating the absolute bias between CRP measurements obtained with three analyzers. For remainder of key, refer to Fig. 1b