simGFRdata: Data Set Containing Simulated Longitudinal eGFR
In SurvDisc: Discrete Time Survival and Longitudinal Data Analysis

A data set containing simulated values of log-eGFR measured longitudinally over time as a function of baseline eGFR. The data were simulated from a mixed effects model with the following form (using the lme model structure syntax; see format section below for definition of variables):

cfb ~ time + x1:time + trt + trt:time + trt:x1:time + 0

and these coefficients:

time trt time:x1 time:trt time:x1:trt -0.6447911 -0.0478315 0.1333391 0.2186963 -0.0458998

In addition, each subject has a random slope and intercept. The baseline eGFR were simulated from a log-Normal distribution.

1	data(simGFRdata)

Fixed effect coefficients used to simulate the data: fix.beta fixed.leGFR fixed.leGFRtrt fixed.time fixed.trt fixed.trttime

mu.base.leGFR: mean of baseline log-eGFR var.base.leGFR: variance of baseline log-eGFR

res.sd: residual error standard deviation. note this is for a single log-eGFR, so the standard deviation for the change from baseline is sqrt(2)*res.sd and the residual error for cfb within a patient have correlation 0.5.

Variance components of random effects distribution: sig.intercept: standard deviaiton of random intercept sig.time: standrd deviation of random slope sig.cor: correlation

A data frame named simGFR that consists of fourteen columns and 28800 rows. The variables are: PID: patient ID trt: the treatment group indicator x1: measured value of baseline log-eGFR time: time from baseline measured in years alphai: subject's random intercept betai: subject's random slope alpha: subject's intercept including fixed and random effects beta: subject's slope including fixed and random effects cfb0: the measurement error for the baseline log-eGFR x: the unobserved "true" baseline log-eGFR cfb: the change from baseline in measured log-eGFR

SurvDisc documentation built on May 2, 2019, 9:12 a.m.