care: Care trajectories

In modal-inria/cfda: Categorical Functional Data Analysis

Care trajectories

Description

Care trajectories of patients diagnosed with a serious and chronic condition

Usage

data(care)

Format

A data.frame containing three columns:

• id id of individuals (2929 different ids)

• time number of months since the diagnosis

• state new state.

Details

In this study, patients were followed from the time they were diagnosed with a serious and chronic condition and their care trajectories were tracked monthly from the time of diagnosis. The status variable contains the care status of each individual for each month of follow-up. Trajectories have different lengths.

The four states are:

• D: diagnosed, but not in care

• C: in care, but not on treatment

• T: on treatment, but infection not suppressed

• S: on treatment and suppressed infection

Source

https://larmarange.github.io/analyse-R/data/care_trajectories.RData https://larmarange.github.io/analyse-R/trajectoires-de-soins.html

See Also

Other datasets: biofam2, flours

Examples

data(care)
head(care)

plotData(care)

# Individuals has not the same length. In order to compute the encoding,
# we keep individuals with at least 18 months of history and work
# with the 18 first months.
duration <- compute_duration(care)
idToKeep <- as.numeric(names(duration[duration >= 18]))
care2 <- cut_data(care[care$id %in% idToKeep, ], 18)
head(care2)

# It is recommended to increase the number of cores to reduce computation time
set.seed(42)
basis <- create.bspline.basis(c(0, 18), nbasis = 10, norder = 4)
fmca <- compute_optimal_encoding(care2, basis, nCores = 2)

plotEigenvalues(fmca, cumulative = TRUE, normalize = TRUE)
plot(fmca)
plot(fmca, addCI = TRUE)
plotComponent(fmca, addNames = FALSE)

modal-inria/cfda documentation built on Oct. 19, 2023, 10:03 a.m.