The nlive R package allows to fit a Sigmoidal Mixed Model with 4 parameters, a Piecewise Linear Mixed Model with abrupt change, or a Piecewise Linear Mixed Model with a smooth polynomial transition in the context of longitudinal Gaussian outcomes. This function was designed to be intuitive enough to the less sophisticated users, while using recent developments such as the stochastic approximation expectation-maximization (SAEM) algorithm for efficient estimation. It was designed to optimize the initial values of the main parameters and help interpretation of the output by providing different features such as annotated outputs and graphs.
Install the CRAN version with:
install.packages("nlive")
Install the most recent update version from Github with:
library(remotes)
library(ggplot2)
library(dplyr)
library(lcmm)
library(sitar)
library(Rmisc)
library(graphics)
library(saemix)
library(stats)
library(sqldf)
library(nlraa)
library(janitor)
library(fastDummies)
# remotes::install_github("MaudeWagner/nlive")
install.packages("nlive", repos = c('https://maudewagner.r-universe.dev', 'https://cloud.r-project.org'))
A detailed companion paper, including illustrative examples, is available in BMC Medical Research Methodology:
Capuano AW, Wagner M. nlive: an R package to facilitate the application of the sigmoidal and random changepoint mixed models. BMC Med Res Methodol. 2023;23(1):257. https://doi.org/10.1186/s12874-023-02075-4
Issues and questions about the use of the nlive R package can be reported on https://github.com/MaudeWagner/nlive/issues. When reporting a bug, please provide a reproducible example.
The dataset dataCog (automatically loaded with the package) contains 1200 individuals with one line per visit. On each line, we can read ID the participant identifier, time the negative retrospective time before death (in years), cognition the repeated measures of global cognition collected during the follow-up, and ageDeath the age at death of individuals (in years). For interpretation purposes, ageDeath was centered at its mean (ageDeath90). The following lines display the first lines of dataCog:
head(dataCog)
## ID time cognition ageDeath ageDeath90
## 1000 -10.00 0.45 91 1
## 1000 -9.08 0.27 91 1
## 1000 -8.04 0.19 91 1
## 1000 -6.82 0.15 91 1
## 1000 -5.99 0.05 91 1
## 1000 -4.98 0.15 91 1
Before fitting a model, the user can inspect the longitudinal outcome of interest cognition using the function nlive.inspect():
nlive.inspect(dataset=dataCog, ID="ID", variable="cognition", time="time")
Key plots are generated, including the distribution of the longitudinal cognition, a spaghetti plot of the observed individual cognition trajectories before death for 70 individuals randomly selected in dataCog, and boxplots of the longitudinal observed cognition measures, obtained every year before death, for the whole population.
pmm.abrupt.fit = nlive.pmma(dataset=dataCog, ID="ID", outcome="cognition", time="time", var.all=c("ageDeath90"), traj.marg=TRUE, traj.marg.group=c("ageDeath90"))
## ...
## ----------------------------------------------------
## ----------- Variance of random effects -----------
## ----------------------------------------------------
## Parameter Estimate SE CV(%)
## last.level omega2.last.level 1.07196 4.7e-02 4.4
## slope1 omega2.slope1 0.00062 7.4e-05 11.9
## slope2 omega2.slope2 0.03830 2.0e-03 5.2
## changepoint omega2.changepoint 0.58980 7.9e-02 13.4
## covar cov.slope1.slope2 0.00378 3.2e-04 8.4
## ----------------------------------------------------
## ------ Correlation matrix of random effects ------
## ----------------------------------------------------
## omega2.last.level omega2.slope1 omega2.slope2
## omega2.last.level 1 0.00 0.00
## omega2.slope1 0 1.00 0.78
## omega2.slope2 0 0.78 1.00
## omega2.changepoint 0 0.00 0.00
## omega2.changepoint
## omega2.last.level 0
## omega2.slope1 0
## omega2.slope2 0
## omega2.changepoint 1
## ----------------------------------------------------
## --------------- Statistical criteria -------------
## ----------------------------------------------------
## Likelihood computed by linearisation
## -2LL= 12349.9
## AIC = 12377.9
## BIC = 12449.16
##
## Likelihood computed by importance sampling
## -2LL= 12296.83
## AIC = 12324.83
## BIC = 12396.1
## ----------------------------------------------------
## Parameter Estimate SE p-value
## 1 last.level -1.103 0.031 P<.0001
## 2 beta_ageDeath90(last.level) -0.062 0.004 P<.0001
## 3 slope1 -0.017 0.002 P<.0001
## 4 beta_ageDeath90(slope1) -0.0003 0.0004 0.082
## 5 slope2 -0.249 0.007 P<.0001
## 6 beta_ageDeath90(slope2) -0.001 0.001 0.159
## 7 changepoint -4.25 0.048 P<.0001
## 8 beta_ageDeath90(changepoint) -0.059 0.006 P<.0001
## 9 residual standard error 0.281 0.002 P<.0001
## ----------------------------------------------------
## The program took 168.58 seconds
For the most common profile of covariates (traj.marg=TRUE) - that is here participants aged 90 years at death (var.all=ageDeath90).
According to age at death (traj.marg.group=c("ageDeath90")) - values at the 10th vs. 90th percentiles are considered by default.
head(psi(pmm.abrupt.fit, type="mean"))
## [,1] [,2] [,3] [,4]
## [1,] -0.4407346 -0.010094826 -0.1829749 -3.795890
## [2,] -1.0962085 -0.021387023 -0.2702190 -4.402796
## [3,] -1.4489273 -0.011734242 -0.1828391 -4.222868
## [4,] -2.3125194 -0.037893158 -0.4280412 -4.191715
## [5,] -1.4742977 -0.026969207 -0.3554523 -3.960455
## [6,] -0.3757513 -0.008739808 -0.1729773 -3.857671
head(eta(pmm.abrupt.fit, type="mean"))
## [,1] [,2] [,3] [,4]
## [1,] 0.56429219 0.007543470 0.011370603 0.88655030
## [2,] 0.03368492 0.001122928 0.036148300 -0.92399512
## [3,] -0.55159983 0.028859466 0.007349072 -0.22031669
## [4,] -1.30823969 -0.022026587 -0.172103574 -0.06386854
## [5,] -0.53423330 -0.007728255 -0.115472760 0.11876845
## [6,] 0.53464877 0.001995156 0.058633033 -0.25910762
saemix.plot.randeff(pmm.abrupt.fit)
