The threshold effect analysis:application and implementation in R software
Sheng Song
Zhang Yanhong
Huan Luyao
Guo Manping
Ma Hangkun
Gao Hongyang
Zhao Yang
Huang Ye
Abstract：When the effect size or direction of a study factor changes significantly after surpassing a certain threshold,this is known as a threshold effect.As a statistical tool,the threshold effect analysis can predict the occurrence,development and prognosis of diseases,which is of great significance.However,there is no open-source,free,simple and effective tool for the threshold effect analysis.Determining threshold values is complicated,requiring a heavy workload and high programming skills.Therefore,this paper demonstrates how to streamline this process and reduce the requirements on programming skills with written R script files and simplified codes by example,and enable non-statistical professionals to perform the threshold effect analysis.First,smooth curve fitting was used to observe whether there is a nonlinear relationship between the study factor(X)and the outcome(Y).If there is a visual threshold,the maximum likelihood method,recursion method,piecewise linear model and likelihood ratio test were carried out to determine the threshold,the effect size and statistical significance before and after the threshold.
Keywords：Threshold effectMaximum likelihood methodRecursive methodPiecewise linear regressionR software
Publication Date：2025-02-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：4( 161-163,174 )