Considerations in reference values for nerve conduction studies.

Despite the universal application of reference values, understanding their production and underlying limitations is often overlooked. Without a deeper consideration for this issue, opportunities for treatment may be missed or misdiagnosis may occur. This article reviews relevant biostatistical issues in the generation of reference intervals and subsequent errors that may occur. Further discussion includes recommendations on how to correct statistical errors, as well as more recent methods for distinguishing between "healthy" and "diseased."