The goal of this paper is to describe the metrics used for the evaluation of accuracy of blood glucose (BG) meters for self-monitoring of blood glucose (SMBG) and continuous-glucose monitoring (CGM) system and their limitations and to discuss the current status of SMBG and CGM accuracy. SMBG measurement is used by patients for therapy control and for calculation of appropriate insulin doses for approximately 30 years. The minimum accuracy criteria for SMBG meters are currently defined by ISO 15197:2003 (at least 95 % of results within \(\pm \)20 % or \(\pm \)15 mg/dL of the comparison method measurement results for BG concentrations above or below 75 mg/dL, respectively). In 2013, these accuracy limits were revised in the standard ISO 15197:2013: at least 95 % of results within \(\pm \)15 % or \(\pm \)15 mg/dL for BG above or below 100 mg/dL, respectively. SMBG systems are also used by patients for calibration of CGM systems. Therefore, precision and trueness of the SMBG system are influencing the accuracy of the CGM results. The timing of the BG measurement used for calibration has to be taken into account because, during rapid glucose changes, a time lag exists between BG and the tissue glucose that is measured by CGM systems. The accuracy of CGM devices is often reported by the mean absolute relative deviation (MARD) between CGM results and BG comparison results. This parameter is influenced by different factors like study procedures, glucose fluctuations during the study, and distribution of comparison BG measurements. It is important to define standard study procedures and evaluations to be able to compare MARD results from different studies. For the correct prediction of glucose concentrations, the specific prediction method as well as the accuracy of the CGM system, which may be affected by the accuracy of the SMBG system used for calibration, and the timing of the calibration are important aspects.
