ETTAA - CCW

The main workhorse function is fixITB, which takes the following arguments with default arguments based on the ETTAA application.

• data: A data.frame object,
• GP: Scalar denoting the grace period for the trial emulation (default: 365.25),
• truncation.time: Scalar denoting the truncation/administrative censoring time (default: 7 years),
• bump: Scalar, if intervention/event time occurs at time zero, small amount to add for analysis (default: 0.5),
• stablised.weights: Logical, should stabilised weights be used (default: TRUE),
• rmsts: Numeric vector containing the RMSTs of interest (default: 1,3,5,7 years),
• rmst.scale: Scalar, for presentation purposes what should the RMST be scaled by? (Default: 1 i.e. ignored),
• surgery.formula: Character, right-hand-side of the formula for the surgery IPTW/IPCW models (default: Combination of the key confounders in ETTAA).
• control.formula: Character, right-hand-side of the formula for the control IPTW/IPCW models (default: NULL, which copies surgery.formula).
• max.weight: Scalar, ad-hoc maximal weight for analysis (default: NULL which is no maximum weight),
• verbose: Logical, should statements be printed showing progress of fixITB? (Default: FALSE),
• outputs: Character vector, must be one (or both) of "composite", "CR", tells fixITB what ATE to compute, defaults to both.

A set of dummy data is supplied to demonstrate this in the file demo.R. Here, S3 methods for fixITB are presented (printing/plotting).

The file explanatoryPlot.R creates Figure 1.