= Net survival
{title2=$S_k^{\rm net}(t)=e^{-\Lambda_k(t)}$}
For <cause-specific hazard> $\lambda_k$, the <net survival> is $S_k^{\rm net}(t)=\exp[-\int_0^t\lambda_k(u)\,du]$. Censoring all competing events in a cause-specific <Kaplan–Meier estimator> targets this quantity under appropriate observation-censoring assumptions. Its complement generally exceeds the actual <cumulative incidence function> $\int_0^tS(u)\lambda_k(u)\,du$, because all-cause survival $S(u)$ also includes the other hazards. Identifying <net survival> with survival after a hypothetical elimination of competing events requires extra causal assumptions.
Back to article page
