【摘要】 从实践中提出了当前应用回归分析确定医学参考值中存在的参考界值有脱离实际情况,回归方程中可能有医学意义不大的变量,以及对同一器官功能确定多个方程有增大误诊等主要问题。考虑数学线性模型的正态分布理论要求和医学实践意义,提出了“确定参考值的回归分析步骤”,几年实践显示已基本解决了存在的几个主要问题。更多还原

【Abstract】 There are several methods, used to determine multivariate reference ranges in the current literature, both at home and abroad. However, the method chosen to determine medical reference ranges should be suitable for the specific type of medical data under study. Some questions often encountered in using this method are the following: the reference value is sometimes not suitable for the particular medical situation, some of the variables in the regression equation need not be there, and are not reported in the present literature at home, but it was found that the regression values are smaller than zero in the literature of MEFV. It was in 1976 that the zero reference value was reported by Knudson RJ. The percentile method suggested by him was hosed on regression values. The question also still exists, how will the percentiles of each equation be determined when there are only a few regression equations to determine the reference? After we considered the normal distribution of the mathematical model and the medical sense, the authors suggest a program to determine medical reference values with regression analysis.更多还原