Cumulative Incidence Estimation in the Presence of Competing Risks

Abstract

When competing risks are present, the appropriate estimate of the failure probabilities is the cumulative incidence. stcompet creates new variables containing the estimate of this function, its standard error, and ln(− ln) transformed confidence bounds. Two examples are presented to illustrate the use of the new command and some key features of the cumulative incidence.

Peripheral blood stem cell versus bone marrow transplantation from HLA-identical sibling donors in patients with leukemia: a propensity score-based comparison from the Japan Society for Hematopoietic Stem Cell Transplantation registry

“…OS was assessed using the Kaplan-Meier product limit method [23,24]. Cumulative incidences of acute GVHD, chronic GVHD, NRM, and relapse were evaluated as 1 -(KaplanMeier estimate) instead of applying methods considering competing risks [25,26] to maintain statistical consistency between logrank tests and methods of cumulative incidence estimation. Allo-PBSCT and allo-BMT groups were compared using the propensity score in quartiles [1][2][3][4], a stratified logrank test, and a stratified Cox proportional hazards model.…”

Ten-year incident osteoporosis-related fractures in the population-based Canadian Multicentre Osteoporosis Study — Comparing site and age-specific risks in women and men

“…We tested age-sex interactions and age-gradients with a Cox proportional hazards model. Since the FRAX tool adjusts for competing mortality, we performed further analyses with death as a competing risk [23]. All analyses were performed with Stata (Version 12) (College Station, Texas, USA); we used the package "stcompet" for the competing risks calculations.…”

On the Importance of Accounting for Competing Risks in Pediatric Cancer Trials Designed to Delay or Avoid Radiotherapy: I. Basic Concepts and First Analyses

“…Competing risks in this situation include deaths from prostate cancer and other noncardiovascular causes. In studies of adult patients with supratentorial low-grade gliomas designed to evaluate radiation-related late complications and the benefit of postoperative RT on local control (21), competing risks analysis could also be implemented to assess these outcomes. For example, in the analysis of complications-free rates for somnolence, it would be important to account for other competing complications such as seizures and memory impairment.…”