In 1999, The Japan Diabetes Society (IDS) Committee on Standardization of Laboratory Testing Related to Diabetes Mellitus in cooperation with the Japan Society of Clinical Chemistry (JSCC) revealed significant device-to-device variations in glucose levels measured with devices intended for use in the self-monitoring of blood glucose (SMBG). The main cause of the discrepancy seemed to arise from differences in the manufacturers' comparative methods. The Committee recommended the use of a venous blood method as the standardized comparative method. In 2001, the Committee confirmed that blood glucose measurements of venous blood performed using 10 SMBG devices did not show any large variations. However, the ordinary use of SMBG is to measure glucose levels in fingertip blood. In 2002, the Committee repeated the same experiment in cooperation with volunteer patients with diabetes mellitus. The results suggested the presence of two groups of SMBG devices. In one group, consisting of the Free Style, Medisafe Reader, Xtra and AccuChek Comfort, the measurement of glucose in fingertip blood samples was almost the same as the glucose levels measured in arterial blood measured with the hexokinase method. In the other group, consisting of the Glucocard α and Dexter Z II, the measurements were close to the glucose levels measured in venous blood measured with the hexokinase method. Clinical diabetes educators, who advise patients with diabetes mellitus about the use of SMBG devices, should be aware of these characteristics of SMBG devices.