Competing Risks Regression Analysis

Competing risks regression models the cumulative incidence of a specific event when other event types can occur first and prevent it, most commonly using the Fine-Gray subdistribution hazard model.

경쟁위험 회귀는 다른 유형의 사건이 먼저 일어나 관심 사건을 막을 수 있을 때 특정 사건의 누적발생률을 모형화하며, 가장 흔히 Fine-Gray 하위분포 위험모형을 사용한다.

How do I analyze time to an event when other events can prevent it from happening?

다른 사건이 먼저 일어나 관심 사건을 막을 수 있을 때 사건발생 시간을 어떻게 분석하나요?

Competing risks regression estimates the probability over time of a specific event when another event, such as death from another cause, can occur first and preclude it. It reports cumulative incidence functions and subdistribution hazard ratios, avoiding the bias that comes from treating competing events as censored in Kaplan-Meier analysis.

When to use Competing Risks Regression Analysis

• Time-to-event data with more than one possible event type
• Estimating cumulative incidence of a specific cause such as relapse versus death
• Comparing groups or covariates on the probability of a specific event

When to use something else

• Only one event type with ordinary censoring; use the Cox proportional hazards model
• Interest is in the cause-specific event rate (etiology) rather than absolute risk; use cause-specific Cox models

Assumptions and requirements

• Event type and time recorded for each subject
• Non-informative censoring
• Proportional subdistribution hazards for the Fine-Gray model

What the analysis reports

• Cumulative incidence functions
• Gray's test comparing groups
• Subdistribution hazard ratios with 95% confidence intervals
• Cumulative incidence plots

How to report it

In the Fine-Gray model, [predictor] was associated with a higher cumulative incidence of [event], subdistribution HR = [value], 95% CI [[value], [value]], p = [value].

Related methods

• Cox Proportional Hazards Model
• Kaplan-Meier Estimator
• Cause-Specific Hazards Model

Run Competing Risks Regression Analysis on Quorant