An object of class survival stores survival probabilities. It is typically returned by Psm$sim_survival() or PsmCurves$survival(); however, it can also be constructed "manually" from existing data using the survival() function as described below. The latter option is useful if survival modeling has been performed by an R package other than those that integrate with hesim ( currently flexsurv). In this case a simulation model can still be developed by using sim_stateprobs.survival() to compute simulated state probabilities and then simulating quality-adjusted life-years and costs in a typical fashion.

survival(
  data,
  sample = "sample",
  strategy_id = "strategy_id",
  patient_id = "patient_id",
  grp_id = "grp_id",
  curve = "curve",
  t = "t",
  survival = "survival"
)

Arguments

data

A tabular object that can be coerced to a data.table with as.data.table().

sample

The name of the column corresponding to sample.

strategy_id

The name of the column corresponding to strategy_id.

patient_id

The name of the column corresponding to patient_id.

grp_id

The name of the column corresponding to grp_id.

curve

The name of the column corresponding to curve.

t

The name of the column corresponding to t.

survival

The name of the column corresponding to survival.

Value

An object of class survival that inherits from data.table and contains the following columns:

sample

A random sample from the PSA.

strategy_id

The treatment strategy ID.

patient_id

The patient ID.

grp_id

The subgroup ID.

curve

One of the N-1 survival curves in an N-state partitioned survival model. Each curve corresponds to unique endpoint.

t

The time at which a survival probability is computed.

survival

The probability of surviving to time t.

The object also contains a size attribute that contains the elements n_samples, n_strategies, n_patients, n_states, and n_times denoting the number of samples, treatment strategies, patients, health states, and times.

See also

survival objects are returned by methods in the Psm and PsmCurves classes. An example in which a survival object is constructed "manually" (presumably from a preexisting survival model fit using software other than flexsurv) is provided in the documentation to sim_stateprobs.survival().