diff_mean_deviance_residuals {TSDT}

diff_mean_deviance_residuals

Description

Computes the difference in the mean of deviance residuals function across treatment groups.

Usage

diff_mean_deviance_residuals(data, scoring_function_parameters = NULL)

Arguments

data

data.frame containing response data

scoring_function_parameters

named list of scoring function control parameters

Details

The deviance residual is the observed number of events at time t minus the expected number of events at time t. See documentation for mean_deviance_residuals (linked below) for more details. A smaller value for the deviance residual is preferred when the event under study is an undesirable event – i.e. it is preferred to observe fewer events than predicted by the survival model. A two-arm TSDT model computes the mean deviance residual in the treatment arm minus the mean deviance residual in the control arm. The treatment arm is superior to the control arm when the mean deviance residual in the treatment arm is less than the mean deviance residual in the control arm. Thus, the appropriate value for desirable_response is desirable_response = 'decreasing'. If the event under study is a desirable event the appropriate value for desirable_response is desirable_response = 'increasing'. It is assumed most survival models will model an undesirable event, so the default value for desirable_response when the scoring_function is diff_mean_deviance_residuals is desirable_response = 'decreasing'. Note this differs from all other TSDT configurations, for which the default value for desirable_response is desirable_response = 'increasing'.

Value

Difference in mean deviance residuals across treatment arms.

See Also

mean_deviance_residuals, Surv, coxph, survreg, residuals.coxph, residuals.survreg, TSDT