Simulating Total Error While Excluding Results Underestimates Total Error

摘要

To the Editor:Karon et al. (1) modeled the effects of glucose meter error on insulin dosing errors. They refer to an earlier model that I had pointed out was incomplete (2). Although Boyd and Bruns (3) agreed with my assessment, the current simulation is, once again, incomplete and the same as the previous simulation.We have reviewed the statistics for glucose meter performance (4) and noted that Lawton and coworkers (5) have a simple and complete model for analytical assay error: mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="FD1" overflow="scroll"mml:mrowmml:mo%/mml:momml:mtextTE/mml:mtextmml:mo=/mml:momml:mo%/mml:momml:mtextBias/mml:mtextmml:mo+/mml:momml:mstyle displaystyle="true"mml:mrowmml:mif/mml:mimml:mrowmml:mo?/mml:momml:mrowmml:mo(/mml:momml:mrowmml:msubmml:mrowmml:mtextCV/mml:mtext/mml:mrowmml:mtextT/mml:mtext/mml:msub/mml:mrowmml:mo)/mml:mo/mml:mrow/mml:mrow/mml:mrow/mml:mstylemml:mo+/mml:momml:mstyle displaystyle="true"mml:mrowmml:mif/mml:mimml:mrowmml:mo?/mml:momml:mrowmml:mo(/mml:momml:mrowmml:msubmml:mrowmml:mtextCV/mml:mtext/mml:mrowmml:mrowmml:mtextRI/mml:mtext/mml:mrow/mml:msub/mml:mrowmml:mo)/mml:mo/mml:mrow/mml:mrow/mml:mrow/mml:mstylemml:mo,/mml:mo/mml:mrow/mml:math where %TE is the percent total error, %Bias is the percent mean bias, CVT is the total CV due to imprecision, CVRI is the total CV due to random …