Competing Risks Analysis Calculator

Use competing risks analysis when an event can prevent the event of interest from ever occurring. For example, a patient who dies from another cause can no longer experience the relapse being studied. In this situation, treating that death as ordinary censoring can overestimate the probability of relapse.

Data format

Select the follow-up time, the status variable, and optional group or predictor variables. Code status as:

• 0 = censored,
• 1 = event of interest,
• 2 = competing event.

When the selected status contains code 2, numiqo automatically displays the competing risks results instead of the ordinary Kaplan-Meier analysis.

Results provided by numiqo

• Cumulative incidence functions for event 1 and event 2
• Pointwise 95% confidence bands
• Numbers at risk and estimates at selected time points
• Complete cumulative incidence tables with standard errors and confidence limits
• Gray's test for comparing cumulative incidence between groups
• Fine-Gray regression with subdistribution hazard ratios and 95% confidence intervals
• Reference-category reporting, convergence information, and residual diagnostics

Interpreting the analysis

The cumulative incidence function estimates the probability of experiencing a specified event by a given time while accounting for the fact that another event may occur first. Gray's test compares these curves across groups. Fine-Gray regression relates predictors to the cumulative incidence through subdistribution hazard ratios; these are not ordinary Cox hazard ratios or risk ratios.

Confidence intervals describe uncertainty around the estimates. Fine-Gray coefficients should only be interpreted when the model has converged. The residual plots are a visual check of the proportional subdistribution hazards assumption, not a formal hypothesis test.

Read the competing risks analysis tutorial for the difference between Kaplan-Meier, cumulative incidence, Cox regression, and Fine-Gray regression.