R Functions for Reference Interval Estimation Using Resampling-Based Multiple Outputation

This repository provides R functions to estimate reference intervals from clustered or longitudinal data using repeated within-subject resampling, a method known as Multiple Outputation. The approach is designed to follow reference interval guidelines (such as ASVCP) while accounting for within-subject correlation. Features Resampling Logic: Randomly samples one observation per subject to ensure statistical independence. Core Estimation: Wraps the refLimit() function from the referenceIntervals package. Iteration: Repeats the process M times (Multiple Outputation) to maximize data use. Output: Provides mean reference limits and quantile-based intervals to represent the stability of the estimates. Methodological Background This implementation is based on the Multiple Outputation (or Within-Cluster Resampling) framework introduced by Hoffman et al. (2001) and Follmann et al. (2003). It is specifically suited for settings where multivariate methods for clustered data are impractical or where the cluster size (number of observations per subject) may be related to the outcome. Technical Requirements R Packages: dplyr, referenceIntervals. Data Structure: Requires a subject ID column and a numeric measurement variable.