Fermanian J, Chaperon J, Grond C
Nouv Rev Fr Hematol (1978). 1978 Jun 10;20(2):249-56.
The data which Hemalog D obtains from blood samples of non hematologic patients have been analyzed statistically and compared with those obtained from an eye count. As well as usual statistical techniques, i.e. correlation coefficient and paired t-test, a more sophisticated multidimensional method was used : the Factorial Analysis of Correspondences. We have analyzed two samples. For the first sample, that of "all-comers", the reproductibility of the results given by the machine was shown to be excellent, and always superior to that of the eye count. The machinen-technician agreement is reasonably good except for the monocytes and basophiles. The preceeding conclusions were found to be true for the second sample, i.e. "with discrepancies", and in addition, it was shown that an eye count on 400 elements gives a result closer to that of the machine than does a reading on 100 elements.
对Hemalog D从非血液学患者血样中获取的数据进行了统计分析，并与通过目视计数获得的数据进行了比较。除了常用的统计技术，即相关系数和配对t检验外，还使用了一种更复杂的多维方法：对应因子分析。我们分析了两个样本。对于第一个样本，即“所有来诊者”样本，结果表明该机器给出的结果具有出色的可重复性，且始终优于目视计数的结果。机器与技术人员之间的一致性相当好，但单核细胞和嗜碱性粒细胞除外。对于第二个样本，即“有差异”的样本，上述结论同样成立，此外，还表明对400个细胞进行目视计数所得到的结果比计数100个细胞更接近机器的结果。
