Title:	Blended Survival Curves
Version:	1.0.0
Description:	Create a blended curve from two survival curves, which is particularly useful for survival extrapolation in health technology assessment. The main idea is to mix a flexible model that fits the observed data well with a parametric model that encodes assumptions about long-term survival. The two curves are blended into a single survival curve that is identical to the first model over the range of observed times and gradually approaches the parametric model over the extrapolation period based on a given weight function. This approach allows for the inclusion of external information, such as data from registries or expert opinion, to guide long-term extrapolations, especially when dealing with immature trial data. See Che et al. (2022) <doi:10.1177/0272989X221134545>.
License:	GPL (≥ 3)
Additional_repositories:	https://giabaio.r-universe.dev, https://inla.r-inla-download.org/R/stable
Encoding:	UTF-8
RoxygenNote:	7.3.2.9000
Imports:	dplyr, flexsurv, ggplot2, manipulate, sn, survHE, tibble
Depends:	R (≥ 4.4.0)
Suggests:	INLA, knitr, remotes, rlang, rmarkdown, survHEhmc, survival, testthat (≥ 3.0.0)
URL:	https://github.com/StatisticsHealthEconomics/blendR/, https://StatisticsHealthEconomics.github.io/blendR/
BugReports:	https://github.com/StatisticsHealthEconomics/blendR/issues/
Config/testthat/edition:	3
NeedsCompilation:	no
Packaged:	2025-08-29 13:46:02 UTC; n8than
Author:	Nathan Green [aut], Zhaojing Che [aut, cph, cre]
Maintainer:	Zhaojing Che <blendr-pkg@proton.me>
Repository:	CRAN
Date/Publication:	2025-09-03 20:50:02 UTC

Blended survival object

Description

This is the main function in the blendR package. Two survival curves are supplied and blended according to the blending distribution characterised by the blending interval and the beta distribution parameters.

Usage

blendsurv(
  obs_Surv,
  ext_Surv,
  blend_interv,
  beta_params = list(alpha = 3, beta = 3),
  times = NULL,
  nsim = 100
)

Arguments

Value

List of S for observed, external and blended curves.

Examples

library(survHE)

## trial data
data("TA174_FCR", package = "blendR")

## externally estimated data
data_sim <- ext_surv_sim(t_info = 144,
                         S_info = 0.05,
                         T_max = 180)

obs_Surv <- fit.models(formula = Surv(death_t, death) ~ 1,
                       data = dat_FCR,
                       distr = "exponential",
                       method = "hmc")

ext_Surv <- fit.models(formula = Surv(time, event) ~ 1,
                       data = data_sim,
                       distr = "exponential",
                       method = "hmc")

blend_interv <- list(min = 48, max = 150)
beta_params <- list(alpha = 3, beta = 3)

ble_Surv <- blendsurv(obs_Surv, ext_Surv, blend_interv, beta_params)

plot(ble_Surv)

Survival data

Description

Survival data

Get event time data

Description

Get event time data

Usage

data_times(S)

## S3 method for class 'inla'
data_times(S)

## S3 method for class 'survHE'
data_times(S)

## S3 method for class 'flexsurvreg'
data_times(S)

## Default S3 method:
data_times(S)

Arguments

Value

A vector of event times

Create an external survival data based on expert opinion

Description

Generally, the sampling is done is two steps

p(T) = p(T | interval i) p(interval i)

Usage

ext_surv_sim(t_info, S_info, T_max, n = 100)

Arguments

Details

In particular T ~ U(x_{i}, x_{i+1}) i ~ multinomial(\hat{\pi})

Value

Dataframe of times and censoring status

Examples

dat <- ext_surv_sim(t_info = c(10,20,50),
                    S_info = c(0.9, 0.8, 0.2),
                    T_max = 100, n = 100)
if (require(survival)) {
    km_fit <- survfit(Surv(time, event) ~ 1, data = dat)
    plot(km_fit)
}

Generate survival estimates with a piecewise exponential Cox model (using INLA)

Description

Generate survival estimates with a piecewise exponential Cox model (using INLA)

Usage

fit_inla_pw(
  inla.formula = inla.surv(death_t, death) ~ -1,
  data,
  cutpoints,
  nsim = 100,
  ...
)

Arguments

Value

INLA object

Examples

 if (requireNamespace("INLA", quietly = TRUE)) {
  data("TA174_FCR", package = "blendR")
  head(dat_FCR)
  obs_Surv <- fit_inla_pw(data = dat_FCR, cutpoints = seq(0, 180, by = 5))
 }

Create survival probabilities

Description

These function are version of the survHE::make.surv() function from survHE. These are needed prior to blending.

Usage

make_surv(Surv, ...)

## S3 method for class 'survHE'
make_surv(Surv, t, nsim = 100, ...)

## S3 method for class 'flexsurvreg'
make_surv(Surv, t = NULL, nsim = 100, ...)

## S3 method for class 'inla'
make_surv(Surv, t = NULL, nsim = 100, ...)

## Default S3 method:
make_surv(Surv, t = NULL, nsim = 100, ...)

Arguments

Value

Matrix of survival probabilities

Examples

library(survHE)

## trial data
data("TA174_FCR", package = "blendR")

## externally estimated data
data_sim <- ext_surv_sim(t_info = 144,
                         S_info = 0.05,
                         T_max = 180)

ext_Surv <- fit.models(formula = Surv(time, event) ~ 1,
                       data = data_sim,
                       distr = "exponential",
                       method = "hmc")

S_ext <- make_surv(ext_Surv, t = 1:100, nsim = 100)

Blended survival plot with manipulate

Description

RStudio bug need to run base R first manipulate(plot(1:x), x = slider(5, 10))

Usage

manip_plot(obs_Surv, ext_Surv, blend_interv)

Arguments

Value

Blended survival plot

Blended survival curve based on short-term data and external information

Description

Blended survival curve based on short-term data and external information

Usage

## S3 method for class 'blended'
plot(x, alpha = c(0.1, 0.05), ...)

Arguments

Value

A ggplot2 object

See Also

blendsurv()

Examples

library(survHE)

## trial data
data("TA174_FCR", package = "blendR")

## externally estimated data
data_sim <- ext_surv_sim(t_info = 144,
                         S_info = 0.05,
                         T_max = 180)

obs_Surv <- fit.models(formula = Surv(death_t, death) ~ 1,
                       data = dat_FCR,
                       distr = "exponential",
                       method = "hmc")

ext_Surv <- fit.models(formula = Surv(time, event) ~ 1,
                       data = data_sim,
                       distr = "exponential",
                       method = "hmc")

blend_interv <- list(min = 48, max = 150)
beta_params <- list(alpha = 3, beta = 3)

ble_Surv <- blendsurv(obs_Surv, ext_Surv, blend_interv, beta_params)

plot(ble_Surv)

Plots the weights for the blending procedure

Description

Plots the weights for the blending procedure

Usage

weightplot(x, ...)

Arguments

Value

ggplot2 object

See Also

blendsurv()