Trials using composite outcomes neglect the presence of competing risks: a methodological survey of cardiovascular studies

J Clin Epidemiol. 2023 Aug;160:1-13. doi: 10.1016/j.jclinepi.2023.05.015. Epub 2023 May 27.


Objective: To determine how frequently competing risks were accounted for in recently published cardiovascular disease (CVD) trials with composite end points.

Study design and setting: We conducted a methodological survey of CVD trials that used composite end points and published from January 1 to September 27, 2021. The following databases were searched: PubMed, Medline, Embase, CINAHL, and Web of Science. Eligible studies were categorized according to whether they mentioned a competing risk analysis plan. If yes, whether a competing risk analysis was proposed as the primary or sensitivity analysis.

Results: Among the 136 included studies, only 14 (10.3%) conducted a competing risk analysis and reported the corresponding results. Seven (50%) of them conducted a competing risk analysis as their primary analysis, whereas the other seven (50%) as a sensitivity analysis to assess the robustness of their findings. The most commonly used competing risk analysis methods were the subdistribution hazard model (nine studies), followed by the cause-specific hazard model (four studies) and restricted mean time lost method (one study). None of the studies accounted for competing risks in their sample size calculations.

Conclusion: Our findings underscore the pressing need for and importance of applying appropriate competing risk analysis in this field to disseminate clinically meaningful and unbiased results.

Keywords: Cardiovascular disease; Cause-specific hazard model; Competing risk; Composite outcome; Randomized controlled trial; Subdistribution hazard model.

Publication types

  • Review

MeSH terms

  • Cardiovascular Diseases* / epidemiology
  • Humans
  • Proportional Hazards Models
  • Risk Assessment