Galen R S
Pediatr Clin North Am. 1980 Nov;27(4):861-9. doi: 10.1016/s0031-3955(16)33930-x.
Literature on determining reference values and reference intervals on "normal" or "healthy" individuals is abundant. It is impossible, however, to evaluate a data set of reference values and select a suitable reference interval that will be meaningful for the practice of medicine. The reference interval, no matter how derived statistically, tells us nothing about disease. This is the main reason the concepts of "normal values" have failed us and why "reference values" will prove similarly disappointing. By studying these same constituents in a variety of disease states as well, it will be possible to select "referent values" that will make the test procedure meaningful for diagnostic purposes. In order to obtain meaningful referent values for predicting disease, it is necessary to study not only the "healthy" reference population, but patients with the disease in question, and patients who are free of the disease in question but who have other diseases. Studies of this type are not frequently found for laboratory tests that are in common use today.
关于确定“正常”或“健康”个体的参考值和参考区间的文献丰富。然而，要评估一组参考值并选择一个对医学实践有意义的合适参考区间是不可能的。参考区间，无论其统计推导方式如何，都无法告诉我们任何关于疾病的信息。这就是“正常值”概念让我们失望的主要原因，也是“参考值”同样会令人失望的原因。通过同样研究处于各种疾病状态下的这些相同成分，就有可能选择出对诊断目的而言能使检测程序有意义的“参照值”。为了获得用于预测疾病的有意义的参照值，不仅有必要研究“健康”的参考人群，还要研究患有相关疾病的患者以及没有相关疾病但患有其他疾病的患者。对于当今常用的实验室检测，这类研究并不常见。
