How is time-to-event data used in Markov or survival models?

• A. To estimate baseline risk only.
• B. To determine final outcome measure.
• C. To inform transition probabilities or hazard rates that govern state changes over cycles.
• D. To calculate the sample size required for trials.

Answer: (C)

Time-to-event data provide the rates at which events occur over time, which survival or Markov models translate into transition probabilities (or hazard rates) between health states in each cycle. This lets the model capture how likely it is for a patient to move from one state to another—for example, from progression-free to progressed, or from progressed to death—at any point in time, and to accumulate costs and utilities accordingly. It’s not just about baseline risk or a single final outcome; the strength of time-to-event data is in shaping the dynamic transitions that drive the model’s trajectory over time. While trial design uses time-to-event information for planning, in Markov or survival models the key role is to define the state-change dynamics that reflect the observed timing of events.