Substitution model is a useful post-hoc analysis for regression
models with compositional predictors. Results from our main
brms model tell us how each compositional predictor
(expressed as ILR coordiante) is associated with an outcome. However, we
often are also interested in the changes in an outcomes when a fixed
duration of time is reallocated from one compositional component to
another, while the other components remain constant.
The Compositional Isotemporal Substitution Model can be used to
estimate this change. The multilevelcoda package implements
this method in a multilevel framework and offers functions for both
between- and within-person levels of variability. We discuss 4 different
substitution models in this vignette.
We will begin by loading necessary packages,
multilevelcoda, brms (for models fitting),
doFuture (for parallelisation), and datasets mcompd
(simulated compositional sleep and wake variables), sbp
(sequential binary partition), and psub (base possible
substitution).
library(multilevelcoda)
library(brms)
#> Loading required package: Rcpp
#> Loading 'brms' package (version 2.18.0). Useful instructions
#> can be found by typing help('brms'). A more detailed introduction
#> to the package is available through vignette('brms_overview').
#>
#> Attaching package: 'brms'
#> The following object is masked from 'package:stats':
#>
#> ar
library(doFuture)
#> Loading required package: foreach
#> Loading required package: future
data("mcompd")
data("sbp")
data("psub")
options(digits = 3) # reduce number of digits shownLet’s fit our main brms model predicting
STRESS from both between and within-person sleep-wake
behaviours (represented by isometric log ratio coordinates), with sex as
a covariate, using the brmcoda() function. We can compute
ILR coordinate predictors using compilr() function.
cilr <- compilr(data = mcompd, sbp = sbp,
parts = c("TST", "WAKE", "MVPA", "LPA", "SB"), idvar = "ID")
m <- brmcoda(compilr = cilr,
formula = STRESS ~ bilr1 + bilr2 + bilr3 + bilr4 +
wilr1 + wilr2 + wilr3 + wilr4 + Female + (1 | ID),
cores = 8, seed = 123, backend = "cmdstanr")
#> Compiling Stan program...
#> Start samplingA summary() of the model results.
summary(m$Model)
#> Family: gaussian
#> Links: mu = identity; sigma = identity
#> Formula: STRESS ~ bilr1 + bilr2 + bilr3 + bilr4 + wilr1 + wilr2 + wilr3 + wilr4 + Female + (1 | ID)
#> Data: tmp (Number of observations: 3540)
#> Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1;
#> total post-warmup draws = 4000
#>
#> Group-Level Effects:
#> ~ID (Number of levels: 266)
#> Estimate Est.Error l-95% CI u-95% CI
#> sd(Intercept) 0.99 0.06 0.87 1.11
#> Rhat Bulk_ESS Tail_ESS
#> sd(Intercept) 1.00 1574 2552
#>
#> Population-Level Effects:
#> Estimate Est.Error l-95% CI u-95% CI Rhat
#> Intercept 2.62 0.49 1.66 3.58 1.00
#> bilr1 0.11 0.31 -0.48 0.73 1.00
#> bilr2 0.52 0.33 -0.13 1.20 1.00
#> bilr3 0.13 0.22 -0.30 0.55 1.00
#> bilr4 0.02 0.28 -0.54 0.55 1.00
#> wilr1 -0.34 0.12 -0.58 -0.11 1.00
#> wilr2 0.05 0.13 -0.21 0.30 1.00
#> wilr3 -0.11 0.08 -0.26 0.05 1.00
#> wilr4 0.24 0.10 0.04 0.44 1.00
#> Female -0.39 0.17 -0.71 -0.06 1.00
#> Bulk_ESS Tail_ESS
#> Intercept 1345 2241
#> bilr1 1126 2074
#> bilr2 1366 1982
#> bilr3 1126 1436
#> bilr4 922 1831
#> wilr1 3041 2537
#> wilr2 3208 3224
#> wilr3 3269 3139
#> wilr4 3279 2873
#> Female 1515 2370
#>
#> Family Specific Parameters:
#> Estimate Est.Error l-95% CI u-95% CI Rhat
#> sigma 2.37 0.03 2.31 2.43 1.00
#> Bulk_ESS Tail_ESS
#> sigma 5233 2780
#>
#> Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
#> and Tail_ESS are effective sample size measures, and Rhat is the potential
#> scale reduction factor on split chains (at convergence, Rhat = 1).We can see that the first and forth within-person ILR coordinates were both associated with stress. Interpretation for multilevel ILR coordinates can often be less intuitive. For example, the significant coefficient for wilr1 shows that the within-person change in sleep behaviours (sleep duration and time awake in bed combined), relative to wake behaviours (moderate to vigorous physical activity, light physical activity, and sedentary behaviour) on a given day, is associated with stress. However, as there are several behaviours involved in this coordinate, we don’t know the within-person change in which of them drives the association. It could be the change in sleep, such that people sleep more than their own average on a given day, but it could also be the change in time awake. Further, we don’t know about the specific changes in time spent across behaviours. That is, if people sleep more, what behaviour do they spend less time in?
This is common issue when working with multilevel compositional data
as ILR coordinates often contains information about multiple
compositional components. It is further inconvenient in the case of
within-person ILR coordinates, as they represent the deviation from the
mean (between-person) ILR coordinates. To gain further insights into
these associations and help with interpretation, we can conduct post-hoc
analyses using the substitution models from our multilevel
package.
multilevelcoda package provides 4 different
functions to compute substitution models, using the
substitution() function.
Basic substitution model:
Average marginal substitution model:
Tips: Substitution models are often computationally
demanding tasks. You can speed up the models using parallel execution,
for example, using doFuture package.
The below example examines the changes in stress for different
pairwise substitution of sleep-wake behaviours for a period of 1 to 5
minutes, at between-person level. We specify
level = between to indicate substitutional change would be
at the between-person level, and type = conditional to
indicate basic substitution model. If your model contains covariates,
substitution() will average predictions across levels of
covariates as the default.
bsubm1 <- substitution(object = m, delta = 1:5,
level = "between", type = "conditional")Output from substitution() contains multiple data set of
results for all available compositional component. Here are the results
for changes in stress when sleep (TST) is substituted for 5 minutes.
knitr::kable(bsubm1$TST)|| || || ||
None of them are significant, given that the credible intervals did not cross 0, showing that increasing sleep (TST) at the expense of any other behaviours was not associated in changes in stress at between-person level.
These results can be plotted to see the patterns more easily using
the plotsub() function.
plotsub(data = bsubm1$TST, x = "sleep", y = "stress")#> Error in plotsub(data = bsubm1$TST, x = "sleep", y = "stress"): data must be a data table or data frame,and is an element of a wsub, bsub, wsubmargins, bsubmargins object.Let’s now take a look at how stress changes when different pairwise
of sleep-wake behaviours are substituted for 5 minutes, at within-person
level. We can obtain prediction for each level of covariates by adding
an argument summary = FALSE to
substitution().
wsubm1 <- substitution(object = m, delta = 5,
level = "within", type = "conditional",
summary = FALSE)Results for 5 minute substitution for each level of covariates. In this example, we get separate predictions for males and females.
knitr::kable(wsubm1$TST)|| || || ||
At within-person level, we got some significant results for substitution of sleep (TST) and time awake in bed (WAKE) for 5 minutes for both males and females, but not other behaviours. For males (Female = 0), increasing 5 minutes in sleep at the expense of time spent awake in bed predicted 0.02 higher stress [95% CI 0.00, 0.03], on a given day. Conversely, less sleep and more time awake in bed predicted less stress (b = -0.02 [95% CI -0.03, -0.00]).
Let’s also plot theses results.
plotsub(data = wsubm1$TST, x = "sleep", y = "stress")#> Error in plotsub(data = wsubm1$TST, x = "sleep", y = "stress"): data must be a data table or data frame,and is an element of a wsub, bsub, wsubmargins, bsubmargins object.Average substitution models models are generally more computationally expensive than basic subsitution models. The average marginal models use the group- level compositional mean as the reference composition to obtain the average of the predicted group-level changes in the outcome when every person in the sample reallocates a specific unit from one compositional part to another. This is difference from the basic substitution model which yields prediction conditioned on an “average” person in the data set (by using the population- level compositional mean as the reference composition). All models can be run faster in shorter walltime using parallel execution.
In this example, we use package doFuture for
parallelisation. substitution() will run 5 substitution
models for 5 sleep-wake behaviours, so we will parallel them across 5
workers.
registerDoFuture()
plan(multisession, workers = 5)bsubm2 <- substitution(object = m, delta = 1:5,
level = "between", type = "marginal")
knitr::kable(bsubm2$TST[abs(MinSubstituted) == 5])
#> Error in knitr::kable(bsubm2$TST[abs(MinSubstituted) == 5]): object 'MinSubstituted' not foundwsubm2 <- substitution(object = m, delta = 1:5,
level = "within", type = "marginal")
knitr::kable(wsubm2$TST[abs(MinSubstituted) == 5])
#> Error in knitr::kable(wsubm2$TST[abs(MinSubstituted) == 5]): object 'MinSubstituted' not found
registerDoSEQ()A comparison between between- and within-person substitution model of
sleep on stress, plot using plotsub() and
ggpubr::ggarrange() functions.
library(ggpubr)
p1 <- plotsub(data = bsubm2$TST, x = "between-person sleep", y = "stress")
p2 <- plotsub(data = wsubm2$TST, x = "within-person sleep", y = "stress")
ggarrange(p1, p2,
ncol = 1, nrow = 2)#> Loading required package: ggplot2
#>
#> Attaching package: 'ggpubr'
#> The following object is masked from 'package:cowplot':
#>
#> get_legend
#> Error in plotsub(data = bsubm2$TST, x = "between-person sleep", y = "stress"): data must be a data table or data frame,and is an element of a wsub, bsub, wsubmargins, bsubmargins object.
#> Error in plotsub(data = wsubm2$TST, x = "within-person sleep", y = "stress"): data must be a data table or data frame,and is an element of a wsub, bsub, wsubmargins, bsubmargins object.
#> Error in ggarrange(p1, p2, ncol = 1, nrow = 2): object 'p1' not found