This bug causes the values of CumEvents for a TTE design to show incorrectly. We have the following workaround currently -
User can generate CumEvents inside their R function using MaxEvents and InfoFraction as follows - CumEvents = LookInfo$InfoFrac*DesignParam$MaxEvents
Use CumCompleters in place of CumEvents.
Once this bug is fixed, the following changes need to be made -
inst/Templates/Analyze.TimeToEvent.R file - Replace all instances of CumCompleters with CumEvents including in docstring and function body.
inst/Examples/2ArmTimeToEventOutcomeAnalysis/RCode - Replace all instances of CumCompleters with CumEvents in all the examples in this folder.
Here's a brief description of CumCompleters and CumEvents as they apply to East Horizon engines
For Binary and Continuous endpoints
CumCompleters - Cumulative completors at each look. These are total subjects who have given a response by the time of each look.
CumEvents - Not applicable
For Time to event endpoint
CumEvents - Cumulative events at each look. These are total subjects who have given an event by the time of each look.
CumCompleters - Concept of cumulative completers is not applicable for a Time to event endpoint in our engine. In R connection, CumCompleters field is a copy of CumEvents. However, due to a bug, CumEvents doesn't correctly capture the cumulative events values. CumCompleters is not affected by this bug so the user can use it in place of CumEvents as a workaround.
This bug causes the values of CumEvents for a TTE design to show incorrectly. We have the following workaround currently -
Once this bug is fixed, the following changes need to be made -
Here's a brief description of CumCompleters and CumEvents as they apply to East Horizon engines For Binary and Continuous endpoints
For Time to event endpoint