A hospital-wide blind control program for bedside glucose meters.

We have instituted hospital-wide proficiency testing of nursing personnel who operate glucose-monitoring instruments. Whole blood samples with known glucose levels were distributed to all nursing units for analysis on a biweekly basis. Measurements obtained by nurses on glucose meters were compared with one another and with true glucose levels to identify discrepant performance and institute corrective action. After 10 months of operation, we found that the nursing units using glucose meters daily showed mean absolute variation from true glucose of 12.7% (range, 10.0% to 17.7%), whereas intermittent users showed mean absolute variation of 15.3% (range, 10.7% to 24.0%). Of 552 total paired determinations, 199 (36.9%) differed from true glucose by greater than 15%. This program was successful in establishing uniform quality control procedures among nonlaboratory personnel who had no previous experience in laboratory measurements.