### Install flexsurv Development Version

Source: https://github.com/chjackson/flexsurv/blob/master/README.md

Install the latest development version of flexsurv directly from GitHub using the devtools package. Ensure devtools is installed first.

```r
install.packages("devtools") # if devtools not already installed
devtools::install_github('chjackson/flexsurv-dev')
```

--------------------------------

### Install flexsurv Package

Source: https://context7.com/chjackson/flexsurv/llms.txt

Installs the stable CRAN version or the development version from GitHub.

```r
install.packages("flexsurv")
```

```r
devtools::install_github("chjackson/flexsurv-dev")
```

--------------------------------

### Calculate quantiles using qgeneric

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/qgeneric.html

This example demonstrates how to use qgeneric to find quantiles for a normal distribution. Ensure that all distribution parameters are explicitly named.

```R
qgeneric(pnorm, c(0.025, 0.975), mean=0, sd=1) # must name the arguments
```

--------------------------------

### Install flexsurv from CRAN

Source: https://github.com/chjackson/flexsurv/blob/master/README.md

Use this command to install the stable version of the flexsurv package from CRAN.

```r
install.packages("flexsurv")
```

--------------------------------

### Fit Custom Distribution (EV) using eha package

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/flexsurvreg.html

Demonstrates fitting a custom Extreme Value (EV) distribution using definitions from the 'eha' package. Requires the 'eha' package to be installed and loaded.

```R
if (require("eha")) {
custom.ev <- list(name="EV",
                  pars=c("shape","scale"),
                  location="scale",
                  transforms=c(log, log),
                  inv.transforms=c(exp, exp),
                  inits=function(t){ c(1, median(t)) })
flexsurvreg(formula = Surv(futime, fustat) ~ 1, data = ovarian,
            dist=custom.ev)
lines(fitev, col="purple", col.ci="purple")
}
```

--------------------------------

### fmixmsm Constructor

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/fmixmsm.html

Constructs a mixture multi-state model based on flexsurvmix. It takes named arguments, where each argument is a fitted model from flexsurvmix, and the argument name specifies the starting state.

```APIDOC
## fmixmsm

### Description
Constructor for a mixture multi-state model based on flexsurvmix.

### Usage

    fmixmsm(...)

### Arguments

*   `...`: Named arguments. Each argument should be a fitted model as returned by `[flexsurvmix](flexsurvmix.html)`. The name of each argument names the starting state for that model.

### Value

A list of `[flexsurvmix](flexsurvmix.html)` objects, with the following attribute(s):

*   `pathways`: A list of all potential pathways until absorption, for models without cycles. For models with cycles this will have an element `has_cycle=TRUE`, plus the pathways discovered before the function found the cycle and gave up.
```

--------------------------------

### Fit a flexsurvreg model and augment data

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/augment.flexsurvreg.html

This example demonstrates fitting an exponential survival model using `flexsurvreg` and then augmenting the original dataset with predictions and residuals using `augment`. The `.resid` column is included because the original data was provided.

```R
fit <- flexsurvreg(formula = Surv(futime, fustat) ~ age, data = ovarian, dist = "exp")
augment(fit, data = ovarian)
```

--------------------------------

### totlos.fs Function

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/totlos.fs.html

Calculates the expected amount of time spent in each state of a multi-state process, starting from a given state, up to a specified maximum time.

```APIDOC
## totlos.fs

### Description

Calculates the matrix whose (r,s) entry is the expected amount of time spent in state s for a time-inhomogeneous, continuous-time Markov multi-state process that starts in state r, up to a maximum time t. This is defined as the integral of the corresponding transition probability up to that time.

### Usage

```r
totlos.fs(
  x,
  trans = NULL,
  t = 1,
  newdata = NULL,
  ci = FALSE,
  tvar = "trans",
  sing.inf = 1e+10,
  B = 1000,
  cl = 0.95,
  ...
)
```

### Arguments

*   `x`: A model fitted with `flexsurvreg`. This must be a Markov / clock-forward model, but can be time-inhomogeneous. `x` can also be a list of models.
*   `trans`: Matrix indicating allowed transitions. See `msfit.flexsurvreg`.
*   `t`: Time or vector of times to predict up to. Must be finite.
*   `newdata`: A data frame specifying the values of covariates in the fitted model, other than the transition number.
*   `ci`: Logical. If TRUE, return a confidence interval calculated by simulating from the asymptotic normal distribution of the maximum likelihood estimates.
*   `tvar`: Variable in the data representing the transition type. Not required if `x` is a list of models.
*   `sing.inf`: Workaround for singularities in the observed hazard. Assumed to be a large finite number at the time of singularity.
*   `B`: Number of simulations from the normal asymptotic distribution used to calculate variances.
*   `cl`: Width of symmetric confidence intervals, relative to 1.
*   `...`: Arguments passed to `ode` in deSolve.

### Value

The matrix of lengths of stay \(T(t)\), if `t` is of length 1, or a list of matrices if `t` is longer. If `ci=TRUE`, each element has attributes `"lower"` and `"upper"` giving matrices of the corresponding confidence limits. The result also has an attribute `P` giving the transition probability matrices.

### Details

This is computed by solving a second order extension of the Kolmogorov forward differential equation numerically. The equation is expressed as a linear system and solved for \(T(t)\) and \(P(t)\) simultaneously, where \(T(t)\) is the matrix of total lengths of stay, \(P(t)\) is the transition probability matrix for time \(t\), and \(Q(t)\) is the transition hazard or intensity as a function of \(t\). The initial conditions are \(T(0) = 0\) and \(P(0) = I\).

Note that the package `msm` has a similar method `totlos.msm`. `totlos.fs` should give the same results as `totlos.msm` when the time-to-event distribution is exponential for all transitions and the model is time-homogeneous.
```

--------------------------------

### Simulate Multi-State Survival Data with pmatrix.simfs

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/pmatrix.simfs.html

Use pmatrix.simfs to simulate the transition probability matrix for a multi-state model. This example demonstrates simulating probabilities at different time points (t=5 and t=10) using an exponential distribution. The results should converge to those from pmatrix.fs for exponential distributions as the number of simulations increases.

```r
bexp <- flexsurvreg(Surv(years, status) ~ trans, data=bosms3, dist="exp")
tmat <- rbind(c(NA,1,2),c(NA,NA,3),c(NA,NA,NA))

# more likely to be dead (state 3) as time moves on, or if start with
# BOS (state 2)

pmatrix.simfs(bexp, t=5, trans=tmat)
#>         1       2       3
#> 1 0.29456 0.26691 0.43853
#> 2 0.00000 0.26781 0.73219
#> 3 0.00000 0.00000 1.00000
pmatrix.simfs(bexp, t=10, trans=tmat)
#>         1       2       3
#> 1 0.08806 0.15161 0.76033
#> 2 0.00000 0.07006 0.92994
#> 3 0.00000 0.00000 1.00000

# these results should converge to those in help(pmatrix.fs), as M
# increases here and ODE solving precision increases there, since with
# an exponential distribution, the semi-Markov model is the same as the
# Markov model.
```

--------------------------------

### pmatrix.fs() / totlos.fs() — Transition probabilities and total length of stay

Source: https://context7.com/chjackson/flexsurv/llms.txt

Computes the transition probability matrix (probability of being in each state at time t, given starting state) and expected total time spent in each state, directly from a list of transition-specific flexsurvreg models, without requiring the mstate package.

```APIDOC
## pmatrix.fs() / totlos.fs() — Transition probabilities and total length of stay

### Description
Computes the transition probability matrix (probability of being in each state at time `t`, given starting state) and expected total time spent in each state, directly from a list of transition-specific `flexsurvreg` models, without requiring the `mstate` package.

### Method
```r
pmatrix.fs(fits, trans, t, start, ...)
totlos.fs(fits, trans, t, start, ...)
```

### Parameters
- **fits**: A list of fitted flexsurvreg objects, one for each transition.
- **trans**: A transition matrix defining the states and possible transitions.
- **t**: The time at which to compute the transition probabilities or total length of stay.
- **start**: The starting state.
- **...**: Additional arguments.

### Request Example
```r
library(flexsurv)

data(bosms3)
tmat <- rbind(c(NA, 1, 2), c(NA, NA, 3), c(NA, NA, NA))

# Fit separate models per transition
fits <- vector("list", 3)
for (i in 1:3) {
  fits[[i]] <- flexsurvreg(
    Surv(years, status) ~ 1,
    data = bosms3[bosms3$trans == i, ],
    dist = "weibull"
  )
}

# Transition probability matrix at t=5 starting from state 1
pmat <- pmatrix.fs(fits, trans = tmat, t = 5, start = 1)
pmat
#            to
# from        State 1 State 2 State 3
#   [starting] 0.47    0.13    0.40

# Expected total length of stay in each state over [0, 10]
totlos.fs(fits, trans = tmat, t = 10, start = 1)
```
```

--------------------------------

### Package Overview

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/index.html

Information about the flexsurv package.

```APIDOC
## Package Overview

### `flexsurv-package` / `flexsurv`

This package, `flexsurv`, provides tools for flexible parametric survival and multi-state models.
```

--------------------------------

### Getting Results from Survival Models

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/index.html

Functions for extracting and summarizing results from fitted survival models.

```APIDOC
## summary(flexsurvreg)

### Description
Summaries of fitted flexible survival models.

### Method
Call

### Endpoint
summary.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## standsurv()

### Description
Marginal survival and hazards of fitted flexsurvreg models.

### Method
Call

### Endpoint
standsurv

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## coef(flexsurvreg)

### Description
Extract model coefficients from fitted flexible survival models.

### Method
Call

### Endpoint
coef.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## normboot.flexsurvreg()

### Description
Simulate from the asymptotic normal distribution of parameter estimates.

### Method
Call

### Endpoint
normboot.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## hr_flexsurvreg()

### Description
Hazard ratio as a function of time from a parametric survival model.

### Method
Call

### Endpoint
hr_flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## vcov(flexsurvreg)

### Description
Variance-covariance matrix from a flexsurvreg model.

### Method
Call

### Endpoint
vcov.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## simulate(flexsurvreg)

### Description
Simulate censored time-to-event data from a fitted flexsurvreg model.

### Method
Call

### Endpoint
simulate.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## plot(flexsurvreg)

### Description
Plots of fitted flexible survival models.

### Method
Call

### Endpoint
plot.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## lines(flexsurvreg)

### Description
Add fitted flexible survival curves to a plot.

### Method
Call

### Endpoint
lines.flexsurvreg

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## plot(standsurv)

### Description
Plot standardized metrics from a fitted flexsurv model.

### Method
Call

### Endpoint
plot.standsurv

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## summary(flexsurvrtrunc)

### Description
Summarise quantities of interest from fitted flexsurvrtrunc models.

### Method
Call

### Endpoint
summary.flexsurvrtrunc

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## plot(survrtrunc)

### Description
Plot nonparametric estimates of survival from right-truncated data.

### Method
Call

### Endpoint
plot.survrtrunc

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

```APIDOC
## lines(survrtrunc)

### Description
Add nonparametric estimates of survival from right-truncated data to a plot.

### Method
Call

### Endpoint
lines.survrtrunc

### Parameters
(No specific parameters documented in the source)

### Request Example
(No request example provided)

### Response
(No response details provided)
```

--------------------------------

### Plot Survival Curves

Source: https://github.com/chjackson/flexsurv/blob/master/docs/articles/standsurv.html

Plots survival curves based on extracted predictions. This example combines predictions from the flexsurv model with Kaplan-Meier estimates for comparison.

```R
ggplot() + geom_line(aes(x=time, y=est, color = group2), data=predictions) + 
  geom_step(aes(x=time, y=surv, group=strata), data=kmsurvplot$data.survplot)
```

--------------------------------

### Calculate Transition Probabilities and Total Length of Stay

Source: https://context7.com/chjackson/flexsurv/llms.txt

Use pmatrix.fs and totlos.fs to compute the transition probability matrix and expected total time spent in each state, respectively. These functions work directly from a list of transition-specific flexsurvreg models and do not require the mstate package.

```r
library(flexsurv)

data(bosms3)
tmat <- rbind(c(NA, 1, 2), c(NA, NA, 3), c(NA, NA, NA))
```

```r
# Fit separate models per transition
fits <- vector("list", 3)
for (i in 1:3) {
  fits[[i]] <- flexsurvreg(
    Surv(years, status) ~ 1,
    data = bosms3[bosms3$trans == i, ],
    dist = "weibull"
  )
}
```

```r
# Transition probability matrix at t=5 starting from state 1
pmat <- pmatrix.fs(fits, trans = tmat, t = 5, start = 1)
pmat
```

```r
# Expected total length of stay in each state over [0, 10]
totlos.fs(fits, trans = tmat, t = 10, start = 1)
```

--------------------------------

### Calculate RMST using rmst_lnorm

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/rmst_generic.html

Example of calculating restricted mean survival time using the rmst_lnorm function for a log-normal distribution. Ensure all arguments are named explicitly.

```r
rmst_lnorm(500, start=250, meanlog=7.4225, sdlog = 1.1138)
#> [1] 237.8849
```

--------------------------------

### Generalized Gamma Distribution Functions (Original Parameterisation)

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/GenGamma.orig.html

Provides functions for the density (dgengamma.orig), distribution function (pgengamma.orig), cumulative hazard (Hgengamma.orig), hazard (hgengamma.orig), quantile function (qgengamma.orig), and random generation (rgengamma.orig) for the generalized gamma distribution using the original parameterization.

```APIDOC
## Functions for Generalized Gamma Distribution (Original Parameterisation)

### Description
These functions provide various statistical calculations for the generalized gamma distribution using its original parameterization.

### Usage
```
dgengamma.orig(x, shape, scale = 1, k, log = FALSE)
pgengamma.orig(q, shape, scale = 1, k, lower.tail = TRUE, log.p = FALSE)
Hgengamma.orig(x, shape, scale = 1, k)
hgengamma.orig(x, shape, scale = 1, k)
qgengamma.orig(p, shape, scale = 1, k, lower.tail = TRUE, log.p = FALSE)
rgengamma.orig(n, shape, scale = 1, k)
```

### Arguments
* `x`, `q`: Vector of quantiles.
* `shape`: Vector of Weibull shape parameters.
* `scale`: Vector of scale parameters (default is 1).
* `k`: Vector of Gamma shape parameters.
* `log`: Logical; if TRUE, probabilities p are given as log(p).
* `log.p`: Logical; if TRUE, probabilities p are given as log(p).
* `lower.tail`: Logical; if TRUE (default), probabilities are P(X <= x), otherwise, P(X > x).
* `p`: Vector of probabilities.
* `n`: Number of observations to generate. If `length(n) > 1`, the length is taken to be the number required.

### Value
* `dgengamma.orig`: Density function.
* `pgengamma.orig`: Distribution function.
* `Hgengamma.orig`: Cumulative hazard function.
* `hgengamma.orig`: Hazard function.
* `qgengamma.orig`: Quantile function.
* `rgengamma.orig`: Random deviates generation.
```

--------------------------------

### Custom Integration Options for flexsurvmix

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/flexsurvmix.html

Pass custom options to the integrate function when providing a custom density or hazard without its cumulative version. Example shows setting a relative tolerance.

```R
integ.opts = list(rel.tol=1e-12)
```

--------------------------------

### Fit Custom Distribution (Exponential) by Supplying Hazard Function

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/flexsurvreg.html

Illustrates fitting a custom exponential distribution by providing only the hazard function. This method is useful when the full density or survival functions are not readily available. An error is shown when the distribution is not found, followed by a successful fit using the built-in 'exp' distribution.

```R
hexp2 <- function(x, rate=1){ rate } # exponential distribution
hexp2 <- Vectorize(hexp2)
custom.exp2 <- list(name="exp2", pars=c("rate"), location="rate",
                    transforms=c(log), inv.transforms=c(exp),
                    inits=function(t)1/mean(t))
flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist=custom.exp2)
```

```R
flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist="exp")
```

--------------------------------

### Calculate total time in states

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/totlos.fs.html

Fit an exponential model and then calculate the total time spent in each state at t=10 years. This example demonstrates the basic usage of totlos.fs with a single time point.

```R
bexp <- flexsurvreg(Surv(Tstart, Tstop, status) ~ trans,
                        data=bosms3, dist="exp")
tmat <- rbind(c(NA,1,2),c(NA,NA,3),c(NA,NA,NA))
totlos.fs(bexp, t=10, trans=tmat)
```

--------------------------------

### Simulate Right-Truncated Survival Data

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/flexsurvrtrunc.html

Simulates time to an initial event (X) and time between initial and final event (T) for right-truncated survival data. This is a common setup for demonstrating survival analysis with truncation.

```r
set.seed(1) 
## simulate time to initial event
X <- rexp(1000, 0.2)
## simulate time between initial and final event
T <- rgamma(1000, 2, 10) 

```

--------------------------------

### fmixmsm

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/index.html

Constructor for a mixture multi-state model based on flexsurvmix.

```APIDOC
## fmixmsm()

### Description
Constructor for a mixture multi-state model based on flexsurvmix.

### Method
Function Call

### Endpoint
fmixmsm()

### Parameters
None explicitly documented.

### Request Example
```
fmixmsm(formula, data, ...)
```

### Response
None explicitly documented.
```

--------------------------------

### Generate Tidymodels-Compatible Predictions with predict.flexsurvreg

Source: https://context7.com/chjackson/flexsurv/llms.txt

Use predict.flexsurvreg to get predictions in a tidy tibble format, compatible with the tidymodels ecosystem. Supports various outcome types and returns list-columns for multiple time points.

```r
library(flexsurv)

fit <- flexsurvreg(Surv(futime, fustat) ~ rx, data = ovarian, dist = "weibull")
nd <- data.frame(rx = 1:2)

# Mean survival time
predict(fit, newdata = nd, type = "response")
```

```r
# Survival probabilities at specified times (returns list-column)
predict(fit, newdata = nd, type = "survival", times = c(365, 730))
```

```r
# Quantiles of the survival distribution
predict(fit, newdata = nd, type = "quantile", p = c(0.25, 0.5, 0.75))
```

```r
# With confidence intervals
predict(fit, newdata = nd, type = "survival",
        times = c(365, 730), conf.int = TRUE, conf.level = 0.95)
```

--------------------------------

### Summarize flexsurvreg model with glance

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/glance.flexsurvreg.html

Use `glance` on a fitted `flexsurvreg` object to get a tibble with model summaries like N, events, logLik, AIC, and BIC. Ensure the model is fitted first.

```R
fitg <- flexsurvreg(formula = Surv(futime, fustat) ~ age, data = ovarian, dist = "gengamma")
glance(fitg)
```

--------------------------------

### Simulate parameter estimates with normboot.flexsurvreg

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/normboot.flexsurvreg.html

Simulate parameter estimates from the asymptotic normal distribution for a fitted flexsurvreg model. Use `newdata` to specify covariate values as a list, or `X` to provide them as a matrix.

```R
fite <- flexsurvreg(Surv(futime, fustat) ~ age, data = ovarian, dist="exp")
normboot.flexsurvreg(fite, B=10, newdata=list(age=50))
```

```R
normboot.flexsurvreg(fite, B=10, X=matrix(50,nrow=1))
```

--------------------------------

### Define and use custom survival distributions in flexsurvreg

Source: https://context7.com/chjackson/flexsurv/llms.txt

Supply a list to the `dist` argument in `flexsurvreg` to define custom distributions. This requires specifying distribution name, parameters, transformations, and initialization functions. The distribution must have a density or hazard function available in the environment.

```R
library(flexsurv)
library(eha)  # provides dEV, pEV

# Define a custom extreme value distribution
custom_ev <- list(
  name           = "EV",
  pars           = c("shape", "scale"),
  location       = "scale",
  transforms     = c(log, log),
  inv.transforms = c(exp, exp),
  inits          = function(t) c(1, median(t))
)

fit_ev <- flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist = custom_ev)
fit_ev

# Custom distribution defined via hazard function only (exponential example)
hexp2 <- Vectorize(function(x, rate = 1) rate)
custom_exp2 <- list(
  name           = "exp2",
  pars           = "rate",
  location       = "rate",
  transforms     = list(log),
  inv.transforms = list(exp),
  inits          = function(t) 1 / mean(t)
)
fit_custom <- flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist = custom_exp2)
# Should match:
fit_exp    <- flexsurvreg(Surv(futime, fustat) ~ 1, data = ovarian, dist = "exp")
cbind(fit_custom$res, fit_exp$res)
```

--------------------------------

### coef.flexsurvreg

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/coef.flexsurvreg.html

Extract model coefficients from fitted flexible survival models. This presents all parameter estimates, transformed to the real line if necessary. For example, shape or scale parameters, which are constrained to be positive, are returned on the log scale.

```APIDOC
## coef.flexsurvreg

### Description
Extract model coefficients from fitted flexible survival models. This presents all parameter estimates, transformed to the real line if necessary. For example, shape or scale parameters, which are constrained to be positive, are returned on the log scale.

### Usage

```
coef(object, ...)
```

### Arguments

*   `object` (flexsurvreg or flexsurvspline object) - Output from `flexsurvreg` or `flexsurvspline`, representing a fitted survival model object.
*   `...` - Further arguments passed to or from other methods. Currently unused.

### Value

This returns the `mod$res.t[,"est"]` component of the fitted model object `mod`. See `flexsurvreg`, `flexsurvspline` for full documentation of all components.

### Details

This matches the behaviour of `coef.default` for standard R model families such as `glm`, where intercepts in regression models are presented on the same scale as the covariate effects. Note that any parameter in a distribution fitted by `flexsurvreg` or `flexsurvreg` may be an intercept in a regression model.

### See Also

`flexsurvreg`, `flexsurvspline`.
```

--------------------------------

### Model Comparison and Accessing Estimates in flexsurvreg

Source: https://context7.com/chjackson/flexsurv/llms.txt

Compares fitted models using AIC and demonstrates how to access parameter estimates on the natural scale, log-likelihood, coefficients on the transformed scale, and the variance-covariance matrix.

```r
# Model comparison via AIC
fit_w$AIC
fit_gg$AIC

# Access parameter estimates on natural scale
fit_w$res

# Log-likelihood
logLik(fit_w)
# 'log Lik.' -97.0 (df=3)

# Coefficients on log/transformed scale
coef(fit_w)
# log(shape) log(scale)         rx
#  -0.007614   6.972496   0.443717

# Variance-covariance matrix
vcov(fit_w)
```

--------------------------------

### Calculate RMST using rmst_generic with plnorm

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/rmst_generic.html

Example of calculating restricted mean survival time using the generic rmst_generic function with the plnorm distribution. This demonstrates the flexibility of the generic function. Ensure all arguments are named explicitly.

```r
rmst_generic(plnorm, 500, start=250, meanlog=7.4225, sdlog = 1.1138)
#> [1] 237.8849
# must name the arguments
```

--------------------------------

### Generalized Gamma to Gamma Equivalence

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/GenGamma.html

Illustrates the equivalence between the generalized gamma distribution with Q=sigma and the gamma distribution. This shows how GenGamma can represent the gamma distribution with specific parameter mappings.

```R
dgengamma(x, mu, sigma, Q=sigma)

=

dgamma(x, shape=1/sigma^2, rate=exp(-mu) / sigma^2)
```

--------------------------------

### Broom-compatible model summaries for flexsurvreg objects

Source: https://context7.com/chjackson/flexsurv/llms.txt

Use `tidy()`, `glance()`, and `augment()` from the broom package to get model summaries. `tidy()` provides parameter estimates, `glance()` offers model-level statistics, and `augment()` appends residuals and fitted values to the original data.

```R
library(flexsurv)
library(broom)

fit <- flexsurvreg(Surv(futime, fustat) ~ age + rx, data = ovarian, dist = "gengamma")

# tidy(): one row per parameter
tidy(fit)
# # A tibble: 5 × 6
#   term       estimate std.error statistic  p.value conf.low
# ...

# With confidence intervals and exponentiated covariate effects (time ratios)
tidy(fit, conf.int = TRUE, transform = "coefs.exp")

# Baseline parameters only, on real (log) scale
tidy(fit, pars = "baseline", transform = "baseline.real")

# glance(): one row model-level statistics
glance(fit)
# # A tibble: 1 × 5
#    N events logLik   AIC   BIC
# <int>  <int>  <dbl> <dbl> <dbl>
#    26     12  -88.3  186.   193.

# augment(): residuals and fitted values appended to data
augment(fit)
```

--------------------------------

### Using built-in flexsurv distribution functions (GenGamma, Weibull, etc.)

Source: https://context7.com/chjackson/flexsurv/llms.txt

flexsurv provides d/p/q/r/h/H functions for various distributions like generalized gamma, Gompertz, and log-logistic. These functions have a consistent API and support vectorized parameters. The Royston-Parmar spline distribution is also available.

```R
library(flexsurv)

# Generalized gamma: d/p/q/r/h/H
dgengamma(x = 500, mu = 6.5, sigma = 0.8, Q = 0.5)       # density
pgengamma(q = 500, mu = 6.5, sigma = 0.8, Q = 0.5)        # CDF
qgengamma(p = 0.5, mu = 6.5, sigma = 0.8, Q = 0.5)        # median
rgengamma(n = 10,  mu = 6.5, sigma = 0.8, Q = 0.5)        # random deviates
hgengamma(x = 500, mu = 6.5, sigma = 0.8, Q = 0.5)        # hazard
Hgengamma(x = 500, mu = 6.5, sigma = 0.8, Q = 0.5)        # cumulative hazard

# Gompertz: rate is the scale (PH) parameter
pgompertz(q = 3, shape = 0.1, rate = 0.05, lower.tail = FALSE)  # survival

# Log-logistic
hllogis(x = 2, shape = 1.5, scale = 1.2)                    # hazard

# Royston-Parmar spline distribution (reduces to Weibull when k=0)
# gamma = c(gamma0, gamma1): intercept + slope of log cumulative hazard
knots <- c(-2, 2)
dsurvspline(x = 1, gamma = c(-2, 1.5), knots = knots, scale = "hazard")
psurvspline(q = 1, gamma = c(-2, 1.5), knots = knots, scale = "hazard",
            lower.tail = FALSE)   # survival probability
hsurvspline(x = 1, gamma = c(-2, 1.5), knots = knots, scale = "hazard")

# Natural cubic spline basis matrix (Royston-Parmar parameterisation)
basis(knots = c(-2, 0, 2), x = c(-1, 0, 1))
```

--------------------------------

### Calculate total time spent in states using totlos.simfs

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/totlos.simfs.html

Use totlos.simfs with a fitted flexsurvreg model and a transition matrix to predict time spent in states. The results converge as the time horizon (t) increases.

```R
# BOS example in vignette, and in msfit.flexsurvreg
bexp <- flexsurvreg(Surv(years, status) ~ trans, data=bosms3, dist="exp")
tmat <- rbind(c(NA,1,2),c(NA,NA,3),c(NA,NA,NA))

# predict 4 years spent without BOS, 3 years with BOS, before death
# As t increases, this should converge
totlos.simfs(bexp, t=10, trans=tmat)
#>        1        2        3 
#> 3.744657 2.124928 4.130415 
totlos.simfs(bexp, t=1000, trans=tmat)
#>          1          2          3 
#>   4.127463   2.957653 992.914884 

```

--------------------------------

### Bootstrap confidence intervals for custom quantities using normboot.flexsurvreg()

Source: https://context7.com/chjackson/flexsurv/llms.txt

Generate bootstrap confidence intervals for custom functions of model parameters. `normboot.flexsurvreg()` simulates parameters from the MLEs' asymptotic normal distribution to create these intervals, extending beyond built-in summaries.

```R
library(flexsurv)

fit <- flexsurvreg(Surv(futime, fustat) ~ rx, data = ovarian, dist = "weibull")

# Draw 1000 bootstrap samples of all parameters
boot_samp <- normboot.flexsurvreg(fit, B = 1000)
head(boot_samp)
#        shape    scale         rx
# [1,] 0.8612 1022.432  0.5213
# ...
```

--------------------------------

### Using Alternative Spline Basis Functions

Source: https://context7.com/chjackson/flexsurv/llms.txt

Demonstrates fitting a spline model using an alternative orthogonal spline basis from the 'splines2' package.

```r
library(flexsurv)

# Use alternative orthogonal spline basis (splines2 package)
spl2 <- flexsurvspline(Surv(recyrs, censrec) ~ group, data = bc, k = 1,
                       scale = "hazard", spline = "splines2ns")
```

--------------------------------

### Generalized F Distribution Functions (Original Parameterisation)

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/GenF.orig.html

This section details the functions available for the original parameterisation of the generalized F distribution, including density, distribution, quantile, and random generation.

```APIDOC
## dgenf.orig

### Description
Calculates the density of the original generalized F distribution.

### Usage
```
dgenf.orig(x, mu = 0, sigma = 1, s1, s2, log = FALSE)
```

### Arguments
* `x` (vector): Vector of quantiles.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.
* `log` (logical): If TRUE, probabilities p are given as log(p).

### Value
The density of the distribution.

## pgenf.orig

### Description
Calculates the distribution function (cumulative density) of the original generalized F distribution.

### Usage
```
pgenf.orig(q, mu = 0, sigma = 1, s1, s2, lower.tail = TRUE, log.p = FALSE)
```

### Arguments
* `q` (vector): Vector of quantiles.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.
* `lower.tail` (logical): If TRUE (default), probabilities are P(X <= x), otherwise, P(X > x).
* `log.p` (logical): If TRUE, probabilities p are given as log(p).

### Value
The distribution function of the distribution.

## Hgenf.orig

### Description
Calculates the cumulative hazard function of the original generalized F distribution.

### Usage
```
Hgenf.orig(x, mu = 0, sigma = 1, s1, s2)
```

### Arguments
* `x` (vector): Vector of quantiles.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.

### Value
The cumulative hazard function of the distribution.

## hgenf.orig

### Description
Calculates the hazard function of the original generalized F distribution.

### Usage
```
hgenf.orig(x, mu = 0, sigma = 1, s1, s2)
```

### Arguments
* `x` (vector): Vector of quantiles.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.

### Value
The hazard function of the distribution.

## qgenf.orig

### Description
Calculates the quantile function (inverse distribution function) of the original generalized F distribution.

### Usage
```
qgenf.orig(p, mu = 0, sigma = 1, s1, s2, lower.tail = TRUE, log.p = FALSE)
```

### Arguments
* `p` (vector): Vector of probabilities.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.
* `lower.tail` (logical): If TRUE (default), probabilities are P(X <= x), otherwise, P(X > x).
* `log.p` (logical): If TRUE, probabilities p are given as log(p).

### Value
The quantile function of the distribution.

## rgenf.orig

### Description
Generates random deviates from the original generalized F distribution.

### Usage
```
rgenf.orig(n, mu = 0, sigma = 1, s1, s2)
```

### Arguments
* `n` (integer): Number of observations. If `length(n) > 1`, the length is taken to be the number required.
* `mu` (vector): Vector of location parameters.
* `sigma` (vector): Vector of scale parameters.
* `s1` (vector): Vector of first F shape parameters.
* `s2` (vector): Vector of second F shape parameters.

### Value
Randomly generated observations from the distribution.
```

--------------------------------

### Weibull Distribution Reduction

Source: https://github.com/chjackson/flexsurv/blob/master/docs/reference/Survspline.html

Demonstrates how dsurvspline can reduce to the dweibull function when specific parameters are used, effectively showing a special case of the spline model.

```R
regscale <- 0.786; cf <- 1.82
a <- 1/regscale; b <- exp(cf)
dweibull(1, shape=a, scale=b)
#> [1] 0.1137858
dsurvspline(1, gamma=c(log(1 / b^a), a)) # should be the same
#> [1] 0.1137858
```

--------------------------------

### simulate.flexsurvreg()

Source: https://context7.com/chjackson/flexsurv/llms.txt

Generates simulated time-to-event datasets from a fitted `flexsurvreg` or `flexsurvspline` model, optionally applying right-censoring. Useful for posterior predictive checks, power calculations, and trial simulation.

```APIDOC
## `simulate.flexsurvreg()` — Simulate survival data from a fitted model

Generates simulated time-to-event datasets from a fitted `flexsurvreg` or `flexsurvspline` model, optionally applying right-censoring. Useful for posterior predictive checks, power calculations, and trial simulation.

```r
library(flexsurv)

fit  <- flexsurvreg(Surv(futime, fustat) ~ rx, data = ovarian, dist = "weibull")
fit2 <- flexsurvspline(Surv(futime, fustat) ~ rx, data = ovarian, k = 2)

nd <- data.frame(rx = 1:2)

# Single simulation per covariate row (wide format)
simulate(fit, seed = 42, newdata = nd)
#   sim_1_time sim_1_status
# 1  1403.58        1
# 2  2095.17        1

# Multiple simulations in tidy (long) format
simulate(fit2, nsim = 5, seed = 42, newdata = nd, tidy = TRUE)

# Simulate with administrative censoring at t=1000
simulate(fit, nsim = 10, seed = 42, newdata = nd, censtime = 1000, tidy = TRUE)

# Simulate from original data with delayed entry
simulate(fit, seed = 42, start = 100)
```
```