Challenges of implementing Point-of-Care Testing (POCT) glucose meters in a pediatric acute care setting

Objectives: To investigate factors contributing to analytical bias in POCT glucose values generated by the NICU versus the core laboratory. Methods: The LifeScan Flexx hospital system glucose meters (SureStep) were used in precision and comparison studies between the NICU and laboratory (ABL715 and Vitros 950). Results: Analysis of 40 neonatal blood samples revealed a positive bias between the NICU glucose meters versus either the laboratory glucose meter or instrument (mean difference of 0.28 and 0.21 mmol/L, respectively). Linear regression analysis (R-2 = 0.0584) of the difference in glucose results versus time elapsed between measurements indicated that the bias observed between the NICU and laboratory glucose meters was not due to in vitro glycolysis for samples transported on ice. Further analysis indicated that the bias appeared to be mostly operator driven, with different NICU operators exhibiting different mean biases. Increasing the amount of blood applied to the SureStep Pro test strip (e.g., 60 vs. 20 muL), led to higher values for glucose concentration for the same blood. Nearly 50% of all glucose values reported for the NICU were obtained by the SureStep Flexx glucose meters in a 3-month period following the introduction of POCT, yet the number of laboratory-reported glucose results for the same period increased by 21% as compared to the previous year. Conclusions: Operator error appears to be a source of bias present between the NICU and laboratory, and despite glucose meter utilization in the NICU, the number of glucose measurements by the central laboratory increased after POCT introduction. (C) 2004 The Canadian Society of Clinical Chemists. All rights reserved.