RESUMO
OBJECTIVES: Method imprecision, error rates, and explanatory causes that were identified in the Institute for Quality Management in Healthcare point-of-care (POC) glucose proficiency testing (PT) program were assessed in comparison with results obtained from laboratory glucose PT surveys. METHODS: POC and laboratory glucose PT data were assessed from September 2009 to June 2011. Peer group means and coefficients of variation (CVs) were estimated using the robust algorithm recommended in the International Organization for Standardization/International Electrotechnical Commission 13528(E). Discordant finding investigations were also reviewed to determine the causes of significant and recurring errors. RESULTS: POC glucose CVs were higher than laboratory method CVs (median CV, 4.5% and 1.6%, respectively). While all laboratory glucose results were within the performance limits, 305 (0.59%) of 51,379 POC glucose results exceeded limits. Investigations were required for 277 (0.53%) POC results. Pre- and postanalytical errors accounted for 76% of the discordant findings. Using wrong PT items, sample mix-up on the bench, and reporting results for the wrong sample were the most frequent reasons, while 21% of discordant findings identified manufacturer issues, and 3% were of unknown origin. CONCLUSIONS: Both method CVs and error rates were higher in POC than in laboratory glucose methods, even though larger performance limits were used for the assessment of POC glucose.