<scp>M</scp>e<scp>R</scp>ef: Multivariable extrapolated reference values in motor nerve conduction studies
Sanjeev D. Nandedkar, Erik Stålberg, Paul E. Barkhaus
Abstract
INTRODUCTION: In this study we describe a method called "multivariable extrapolated reference values" (MeRef) that derives reference values (RVs) using patient data and includes the dependence of these variables on multiple patient demographic variables, such as age and height. METHODS: Computer simulations were used to generate "normal" and "patient" nerve conduction data. Median, ulnar, and tibial motor nerve conduction data from 500 patients studied were tabulated. Data were analyzed using the MeRef method. RESULTS: The simulations showed great similarity between RVs obtained from MeRef of "patient" data and traditional analysis of "normal" data. In the real patient data, MeRef gave RVs as regression equations based on patient age and/or height. DISCUSSION: MeRef can provide RVs by including patient demographic data and does not require subject grouping. It provides parameters of multivariable linear regression and standard deviation, and requires a few hundred patient studies to define reference values.