Meta-analytic competing risks analysis — Pooling Competing Risks Evidence Across Studies
Meta-Analysis of Competing Risks Studies · Also known as: meta-analysis of competing risks, pooled competing risks analysis, systematic review competing risks
Meta-analytic competing risks analysis pools results from multiple primary studies that each used a competing risks framework, allowing summary estimates of cause-specific or subdistribution hazard ratios and cumulative incidence functions. Because standard meta-analytic methods may misrepresent competing events, specialized pooling strategies are required that respect the subdistribution hazard structure introduced by Fine and Gray and the distinction between cause-specific and all-cause hazard models.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use meta-analytic competing risks analysis when synthesizing a body of primary studies that each explicitly modeled competing events (e.g., disease-specific mortality vs. other-cause mortality, first hospitalisation vs. death before hospitalisation). The method is appropriate when at least five to ten primary studies reported comparable competing risks effect estimates under consistent model types. Do not use when primary studies used only standard Kaplan-Meier or all-cause Cox regression without accounting for competing events — in that case, pooled estimates will be biased. Also avoid pooling subdistribution and cause-specific hazard ratios together, as they measure different quantities and their combination is not interpretable.
Strengths & limitations
- Provides higher-precision pooled estimates of cause-specific or subdistribution hazard ratios by combining evidence across studies.
- Explicitly accounts for competing events, avoiding the well-documented overestimation of cumulative incidence that occurs in standard survival meta-analyses.
- Allows subgroup and meta-regression analyses to explain heterogeneity in competing risks estimates across populations or settings.
- Supports clinical decision-making by reporting pooled absolute cumulative incidence at key time points alongside relative effect estimates.
- Dependent on primary studies having correctly specified and reported competing risks models; if primary studies ignored competing events, the meta-analysis inherits that bias.
- Pooling subdistribution hazard ratios requires that all primary studies used the Fine-Gray model with consistent covariate adjustment, which is rarely the case in practice.
- Individual patient data meta-analysis, while more flexible, is resource-intensive and rarely available; aggregate-data pooling may mask important within-study heterogeneity.
- Small numbers of eligible studies with correctly reported competing risks estimates can limit power and make heterogeneity assessments unreliable.
Frequently asked
Can I pool Kaplan-Meier estimates from primary studies in a competing risks meta-analysis?
No. Kaplan-Meier estimates treat competing events as censored observations, which overestimates the cumulative incidence of the primary event. A meta-analysis that pools Kaplan-Meier-based estimates will inherit and amplify that overestimation. Only cumulative incidence functions or hazard ratios derived from proper competing risks models should be pooled.
What is the difference between pooling cause-specific and subdistribution hazard ratios?
A cause-specific hazard ratio measures the instantaneous rate of the primary event among those still event-free, treating competing events as censored. A subdistribution hazard ratio (Fine-Gray) measures the effect on the cumulative incidence of the primary event in the presence of competing risks. Both are valid but answer different questions; they should be pooled in separate analyses and interpreted accordingly.
When is individual patient data meta-analysis preferred over aggregate data meta-analysis for competing risks?
Individual patient data (IPD) meta-analysis is preferred when primary studies use different covariate adjustments, when time-varying effects matter, or when the goal is to pool cumulative incidence curves at individual-level time resolution. IPD meta-analysis is more flexible but requires access to original datasets, which study authors or data custodians must provide.
How do I handle studies that report only all-cause hazard ratios without competing risks decomposition?
Such studies should be excluded from the primary analysis. They can be included in a sensitivity analysis with explicit notation that competing events were ignored, or in a supplementary all-cause survival meta-analysis. Mixing all-cause and competing risks estimates in the primary pool is not methodologically defensible.
What software is typically used for meta-analytic competing risks analysis?
R packages such as meta, metafor, and CompetingRisksMeta provide functions for pooling hazard ratios and cumulative incidence estimates. Stata's metan and stcompet commands are widely used in clinical epidemiology. The choice of software should be reported transparently with version numbers in the methods section.
Sources
- Riley, R. D., Hayden, J. A., Steyerberg, E. W., et al. (2013). Prognosis Research Strategy (PROGRESS) 2: Prognostic Factor Research. PLOS Medicine, 10(2), e1001380. DOI: 10.1371/journal.pmed.1001380 ↗
- Wolkewitz, M., Cooper, B. S., Bonten, M. J., Barnett, A. G., & Schumacher, M. (2014). Interpreting and comparing risks in the presence of competing events. BMJ, 349, g5060. DOI: 10.1136/bmj.g5060 ↗
How to cite this page
ScholarGate. (2026, June 3). Meta-Analysis of Competing Risks Studies. ScholarGate. https://scholargate.app/en/epidemiology/meta-analytic-competing-risks-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cox proportional hazardsEpidemiology↔ compare
- Fine-Gray Competing Risks ModelStatistics↔ compare
- Kaplan-Meier AnalysisEpidemiology↔ compare
- Meta-analytic Cohort StudyEpidemiology↔ compare
- Meta-analytic survival analysisEpidemiology↔ compare