Despite the wide distribution of devices for selfmonitoring of blood glucose (SMBG), there is no internationally accepted agreement on a standardized procedure for their evaluation. This is due to incomplete or even inappropriate technical evaluation and to inadequate evaluation criteria for their clinical acceptability. To provide adequate information on the performance of these devices over the whole clinically relevant range for SMBG (30-350 mg/dl), a standardized test procedure has been established for technical (accuracy, precision, and total deviation) and clinical (acceptance analysis) evaluation. To demonstrate the potency of this new approach, the following SMBG devices and test strips were evaluated: Chemstrip bG batch (n = 10), Glucostix batch (n = 2), Accu-Check II (n = 5), Diascan (n = 5), Diascan strip batch (n = 4), Glucocheck SC (n = 4), and Glucometer II (n = 4). The devices and test strips were examined by trained technicians, and in addition, 1 Chemstrip bG batch and 2 Accu-Check II meters were examined by 10 and 5 trained diabetic patients, respectively. Even the best-performing device did not achieve the American Diabetes Association's goal that SMBG measurements should be within 15% of the reference value. Instead, the maximal total deviation within the clinically relevant blood glucose range reached values that equalled deviations from the reference value between 16 and 108%. Three of 36 devices were classified as good, 29 as acceptable, and 4 as unacceptable for clinical use. In conclusion, this new approach to the technical and clinical evaluation of devices for SMBG is easy to perform and provides more realistic and comparable information for clinical use and approval than commonly used methods.