A 1-year randomized prospective study was conducted to investigate the efficacy of feedback from split-sample testing as part of a capillary blood glucose quality assurance program.
A total of 124 nurses were randomized to either group A (quarterly comparisons with feedback) or group B (no feedback). The measure of nurse accuracy against the laboratory at 0, 6, and 12 months was determined by an additional five to seven split-sample tests without giving feedback to either group. Mean accuracy was determined in terms of percent absolute deviation from the laboratory result and a clinical consensus limit of ± 20% deviation from the laboratory.
By 12 months, there was a significant effect of feedback on nurse agreement with the laboratory method (P = 0.022 when agreement was scored as the mean percent absolute difference and P = 0.002 when agreement was scored in terms of the ± 20% clinical consensus limit). Nurses in the group who had received no quarterly feedback from split-sample testing produced a 3.5% greater mean percent absolute deviation from the laboratory method and 12% fewer comparisons within the acceptable ± 20% range.
Feedback received from split-sample testing has a significant effect in maintaining accuracy in capillary blood glucose monitoring.
