Monitoring Information for a Time-to-Event Outcome

library(impart)
library(deming) # Required for deming::theilsen()

This vignette demonstrates how to use impart to monitor the amount of information contained in the accrued data so that analyses are not under- or over-powered. To see all available vignettes in impart, use the vignettes command:

vignette(package = "impart")

Title	Item

Monitoring Ongoing Studies

In a fixed sample size design with a binary or continuous outcome, the timing of the final analysis is based on the last participant’s last visit. Group sequential analyses of such outcomes are based when pre-specified fractions of the maximum sample size have their final outcome observed. Timing analyses in such studies only depends on counting the number of final outcomes observed during the study.

In a time-to-event analysis, precision and power are determined by the number of observed events pooled across study arms [@Schoenfeld1983]. A target number of events are calculated based on Type I error and power criteria: enrollment and follow-up are adjusted to observe the target number of events, after which the analysis is conducted: this is known as an event driven trial.

With continuous, ordinal, or binary outcomes obtained at fixed points during the study, enrollment can still be adjusted to meet desired power and Type I error constraints. In information-monitored designs, the times at which analyses are performed and recruitment is stopped depend on when the information reaches pre-specified thresholds. The amount of information contained in the data depends on the number of observations, the completeness of the data, the analytic methods used, and the interrelationships among the observed data. Depending on the analysis methods used, an individual could contribute baseline covariates and treatment assignment information, post-randomization outcomes, and even post-randomization auxiliary variables.

Challenges in Information Monitoring

When analyzing data as it accrues with a continuous, binary, or ordinal outcome, one has to be careful in differentiating between data that is not-yet-collected (i.e. participants who are on-study but have not yet entered the study window for the outcome) versus data known to be missing (i.e. participants whose study window has closed but whose outcomes were not obtained).

With a time-to-event outcome, this issue is handled through censoring