Shamah Levy Teresa, Méndez-Gómez-Humarán Ignacio, Morales Ruán María Del Carmen, Martinez Tapia Brenda, Villalpando Hernández Salvador, Hernández Ávila Mauricio
National Public Health Institute (INSP), Cuernavaca, Morelos, Mexico.
Center for Research in Mathematics, Aguascalientes, Aguascalientes, Mexico.
PLoS One. 2017 Feb 7;12(2):e0170990. doi: 10.1371/journal.pone.0170990. eCollection 2017.
To evaluate the accuracy and precision of HemoCue 201 (HemoCue) and Masimo Pronto 7 (Masimo) devices for measuring hemoglobin (Hb) in epidemiological studies, having venous blood samples as a gold standard.
We measured Hb concentrations in a field sample of 148 children from one to five years of age. Masimo and HemoCue were used for capillary blood samples and an automatic analyzer for venous blood samples. Regression models with no intercept were constructed to measure precision and predictability, concordance correlations to measure accuracy and precision, and Bland-Altman limits of agreement as well as hierarchical linear models to estimate variance.
Both HemoCue and Masimo underestimated Hb concentrations compared to the gold standard. They respectively yielded the following results: regression coefficients of 0.887 and 0.876 with 98.7% and 98.6% predictability; concordance correlation coefficients of 0.183 (p<0.001) and 0.166 (p<0.001); and Bland-Altman variances of -1.51 and -1.62. With regard to Masimo specifically, the three-level Hierarchical Linear Model showed that 57.9% of total variance stemmed from random errors in repeated measures from the same subject.
HemoCue and Masimo measure lower Hb concentrations than the gold standard. Their accuracy and precision levels are comparable. It is essential to ensure proper use of devices through enhanced training of field workers.
以静脉血样本作为金标准，评估在流行病学研究中用于测量血红蛋白（Hb）的HemoCue 201（HemoCue）和Masimo Pronto 7（Masimo）设备的准确性和精密度。
我们测量了148名1至5岁儿童的现场样本中的Hb浓度。Masimo和HemoCue用于采集毛细血管血样本，自动分析仪用于采集静脉血样本。构建无截距的回归模型以测量精密度和可预测性，一致性相关性以测量准确性和精密度，Bland-Altman一致性界限以及分层线性模型以估计方差。
与金标准相比，HemoCue和Masimo均低估了Hb浓度。它们分别得出以下结果：回归系数分别为0.887和0.876，可预测性分别为98.7%和98.6%；一致性相关系数分别为0.183（p<0.001）和0.166（p<0.001）；Bland-Altman方差分别为-1.51和-1.62。具体就Masimo而言，三级分层线性模型显示，总方差的57.9%源于同一受试者重复测量中的随机误差。
HemoCue和Masimo测量的Hb浓度低于金标准。它们的准确性和精密度水平相当。通过加强对现场工作人员的培训来确保设备的正确使用至关重要。
