Accuracy of two systems for blood glucose monitoring without a meter (Chemstrip/Visidex).

Blood specimens collected from 159 campers were tested using Visidex (VD) (Ames Company, Elkhart, Indiana) and Chemstrip (CS) (Bio-Dynamics, Boehringer Mannheim, Indianapolis, Indiana) compared with glucose analyzer (GA) results. Readers were physicians, two with no prior experience using either strip and two with some experience using CS. Subsequently, 30 clinic patients and 4 staff contributed random specimens that were read by a trained novice as 68 discrete samples and compared with GA results. In both studies readers did not place blood on the strips and were thus blinded as to source and the possible previous reading with the other strip. Correlations of VD and CS with GA for camp and clinic data were calculated. Data were compared for reliability and errors greater than 20% of reference values (GA) were compared for VD and CS. CS correlated better with GA than did VD estimates. Reliability was also significantly greater for CS than for VD readings. Both of these visually read methods provided clinically useful estimates of glycemia.

Duke Scholars

Author Jane Frances Pendergast Biostatistics & Bioinformatics, Division of Biostatistics