Survival analysis concepts including Kaplan-Meier curves, hazard ratios, and censoring, one concept per card.
35 cards · basic cards · AI-written, checked twice. Edit anything.
- What does survival analysis study, broadly speaking?
- The time until a specific event of interest occurs (such as death, relapse, or failure), not just whether the event occurred
- What is the outcome variable typically called in survival analysis?
- Time-to-event (or survival time)
- What is censoring in survival analysis?
- A situation where the exact event time for a subject is unknown because the event had not occurred (or was not observed) by the end of the observation period
- What is right censoring, the most common type of censoring in clinical studies?
- When a subject's follow-up ends (study ends, subject withdraws, or is lost to follow-up) before the event of interest occurs, so it is only known the event happened after that point
- What is left censoring?
- When the event of interest is known to have occurred before observation began, but the exact time is unknown
- Why can subjects who are censored still be included in a survival analysis, unlike in a simple average time-to-event calculation?
- Survival analysis methods use all the follow-up time each subject contributed, even if the event was never observed, rather than discarding censored subjects
- What is a Kaplan-Meier curve?
- A graph that estimates the probability of survival (or remaining event-free) over time, using a step function that drops at each observed event
- What does each downward step in a Kaplan-Meier curve represent?
- The occurrence of one or more events at that particular time point
- Why does a Kaplan-Meier curve typically not drop at a censoring time?
- Censored observations are marked (often with a tick mark) on the curve but do not themselves represent an event, so the curve does not step down at that point
- What does the y-axis of a standard Kaplan-Meier curve represent?
- The estimated proportion of subjects surviving (or remaining event-free) at each time point
- What statistical test is commonly used to compare survival curves between two or more groups?
- The log-rank test
- What is the null hypothesis tested by the log-rank test?
- That there is no difference in the survival distributions between the groups being compared
- What is the hazard rate in survival analysis?
- The instantaneous risk of experiencing the event at a given time, conditional on having survived event-free up to that time
- What is a hazard ratio?
- The ratio of the hazard rate in one group compared to another, summarizing the relative risk of the event over the entire follow-up period
- What does a hazard ratio of 1 indicate when comparing two groups?
- No difference in the risk (hazard) of the event between the two groups
FlashKeepers