Work in progress
This package:
- Compares a functional markov modelling approach to a reference approach for several example models.
- Explores approaches to speeding up Markov modelling in a principled fashion making use of C++ when required.
- Details the benefits of parallisation and provide a code structure in which parallisation is easy to make use of.
- Provides a toolkit for use in discrete Markov modelling.
- Provides optimised code that may be ported into other applications and workflows.
The work in this package was started at the Health Economic 2019 hackathon hosted at Imperial. Much of this work is based on that developed by the hermes6 team. The original reference approach was developed by Howard Thom.
Install the CRAN version (when released):
install.packages("SpeedyMarkov")Alternatively install the development version from GitHub:
# install.packages("remotes")
remotes::install_github("seabbs/SpeedyMarkov")
The first step is to specify a Markov model in the format specified by
SpeedyMarkov. An example framework is example_two_state_markov which
is a two state Markov model that compares an intervention to a baseline.
See ?example_two_state_markov for more details.
SpeedyMarkov::example_two_state_markov()
#> $transitions_list
#> $transitions_list$SoC
#> function(samples = NULL) {
#> # Sample transitions
#> tmp <- list(VGAM::rdiric(samples, c(88, 12)),
#> VGAM::rdiric(samples, c(8, 92)))
#>
#> # Arrange as matrices
#> tmp <- SpeedyMarkov::matrix_arrange(tmp)
#>
#> return(tmp)
#> }
#> <bytecode: 0x55e2b141f700>
#> <environment: 0x55e2b146c078>
#>
#> $transitions_list$`Soc with Website`
#> function(baseline = NULL) {
#>
#> #Sample transitions for each baseline matrix
#> samples <- length(baseline)
#> tmp <- VGAM::rdiric(samples,c(85,15))
#>
#> # Update baseline
#> updated <- purrr::map(1:samples, function(sample) {
#> update <- baseline[[sample]]
#> update[1, ] <- tmp[sample, ]
#> return(update)
#> })
#>
#> return(updated)
#> }
#> <bytecode: 0x55e2b14257e0>
#> <environment: 0x55e2b146c078>
#>
#>
#> $qalys
#> function(samples = NULL) {
#> qaly <- function(samples = 1) {
#> ## Sample
#> tmp <- list(stats::rnorm(samples, mean = 0.95,sd = 0.01) / 2,
#> rep(1 / 2, samples))
#>
#> out <- SpeedyMarkov::vector_arrange(tmp)
#>
#> return(out)
#> }
#>
#> soc <- qaly(samples = samples)
#> soc_with_website <- soc
#>
#> out <- list(soc, soc_with_website)
#> names(out) <- list("SoC", "Soc with Website")
#>
#> out <- purrr::transpose(out)
#>
#> return(out)
#> }
#> <bytecode: 0x55e2b14322a8>
#> <environment: 0x55e2b146c078>
#>
#> $intervention_costs
#> function(samples = NULL) {
#> ## Sample
#> tmp <- list(rep(0, samples),
#> rep(50, samples))
#>
#> out <- SpeedyMarkov::vector_arrange(tmp)
#>
#> return(out)
#> }
#> <bytecode: 0x55e2b1441ba8>
#> <environment: 0x55e2b146c078>
#>
#> $state_costs
#> function(samples = NULL) {
#> state_cost <- function(samples = 1) {
#> tmp <- list(rep(0, samples),
#> rep(0, samples))
#>
#> out <- SpeedyMarkov::vector_arrange(tmp)
#>
#> return(out)
#>
#> }
#>
#> soc <- state_cost(samples = samples)
#> soc_with_website <- soc
#>
#> out <- list(soc, soc_with_website)
#> names(out) <- list("SoC", "Soc with Website")
#>
#> out <- purrr::transpose(out)
#>
#> return(out)
#> }
#> <bytecode: 0x55e2b1449508>
#> <environment: 0x55e2b146c078>
#>
#> $cohorts
#> function(samples = NULL) {
#>
#> cohort <- function(samples = 1) {
#> tmp <- list(rep(1, samples),
#> rep(0, samples))
#>
#> out <- SpeedyMarkov::vector_arrange(tmp)
#>
#> return(out)
#> }
#>
#> soc <- cohort(samples = samples)
#> soc_with_website <- soc
#>
#> out <- list(soc, soc_with_website)
#> names(out) <- list("SoC", "Soc with Website")
#>
#> out <- purrr::transpose(out)
#>
#> return(out)
#> }
#> <bytecode: 0x55e2b1457828>
#> <environment: 0x55e2b146c078>
#>
#> attr(,"class")
#> [1] "SpeedyMarkov" "list"Once a model has been specified a cost effectiveness analysis can run using the following function call.
SpeedyMarkov::markov_ce_pipeline(SpeedyMarkov::example_two_state_markov(),
duration = 100, samples = 10, discount = 1.035,
baseline = 1, willingness_to_pay_thresold = 20000)
#> $simulations_with_ce
#> # A tibble: 20 x 12
#> sample intervention transition state_cost intervention_co… cohort qalys
#> <int> <chr> <list> <list> <dbl> <list> <lis>
#> 1 1 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 2 1 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 3 2 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 4 2 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 5 3 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 6 3 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 7 4 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 8 4 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 9 5 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 10 5 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 11 6 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 12 6 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 13 7 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 14 7 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 15 8 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 16 8 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 17 9 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 18 9 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> 19 10 SoC <dbl[,2] … <dbl [2]> 0 <dbl … <dbl…
#> 20 10 Soc with We… <dbl[,2] … <dbl [2]> 50 <dbl … <dbl…
#> # … with 5 more variables: total_costs <dbl>, total_qalys <dbl>,
#> # incremental_costs <dbl>, incremental_qalys <dbl>,
#> # incremental_net_benefit <dbl>
#>
#> $summarised_ce
#> # A tibble: 2 x 13
#> intervention mean_costs sd_costs mean_qalys sd_qlays mean_incrementa…
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 SoC 0 0 14.0 0.0861 0
#> 2 Soc with We… 50 0 14.0 0.103 -0.00248
#> # … with 7 more variables: sd_incremental_qlays <dbl>,
#> # mean_incremental_costs <dbl>, sd_incremental_costs <dbl>,
#> # mean_incremental_net_benefit <dbl>, sd_incremental_net_benefit <dbl>,
#> # probability_cost_effective <dbl>, icer <dbl>See
Functions
for more details of the functions included in the package
(markov_ce_pipeline for example wraps multiple modular - user
customisable - functions). The package vignettes also provide more
detail.
File an issue here if there is a feature that you think is missing from the package, or better yet submit a pull request.
Please note that the SpeedyMarkov project is released with a
Contributor Code of
Conduct.
By contributing to this project, you agree to abide by its terms.
If using SpeedyMarkov please consider citing the package in the
relevant work. Citation information can be generated in R using the
following (after installing the package),
citation("SpeedyMarkov")
#>
#> To cite SpeedyMarkov in publications use:
#>
#> Sam Abbott (2019). SpeedyMarkov - -
#>
#> A BibTeX entry for LaTeX users is
#>
#> @Article{,
#> title = {SpeedyMarkov},
#> author = {Sam Abbott and Howard Thom},
#> journal = {-},
#> year = {2019},
#> }This package has been developed in docker based on the
rocker/tidyverse image, to access the development environment enter
the following at the command line (with an active docker daemon
running),
docker pull seabbs/SpeedyMarkov
docker run -d -p 8787:8787 -e USER=SpeedyMarkov -e PASSWORD=SpeedyMarkov --name SpeedyMarkov seabbs/speedymarkovThe rstudio client can be accessed on port 8787 at localhost (or
your machines ip). The default username is ceplotr and the default
password is SpeedyMarkov. Alternatively, access the development
environment via
binder.