|
Abstract:
|
Time to progression (TTP), defined as time from randomization to disease progression (PD), is an important endpoint in evaluating efficacy of cancer drug products. PD is measured at pre-planned patient visits and not on a daily basis. Missed PD assessments (MA) occur for reasons such as unevaluable measurements or missed visits. When PD is documented immediately after having MA, there are options in determining TTP: censor observation at the date of last adequate assessment visit with no PD, count observation as an event on the date of documented PD, consider as an interval-censored observation, and others. We will examine these options for the effect on type I error rate, power, and bias in estimation using simulated data with various MA patterns between treatment arms. Advantages and disadvantages of each of the options will be discussed.
|
