A quality control (QC) programme to continually monitor the reliability of hospital ward capillary blood glucose levels has been developed based on a memory blood glucose meter interfaced to a portable microcomputer. Over a 38-week period some 12,500 ward QC tests were performed by approximately 200 accredited operators. Although the mean weekly coefficients of variation were all less than 15%, there was a marked variation between wards, with the frequency of testing of patients the major factor influencing reliability. However an external laboratory-based QC programme suggested a high degree of proficiency which was not achieved by any ward on internal QC assessment.