Modifier effects depend on who remains observable

The risk set at each age must be defined before a penetrance modifier can be separated from selective observation. A carrier may leave observation through death, prophylactic intervention, loss to follow-up, or diagnosis of a competing phenotype. If those events depend on genotype, family history, or testing route, a modifier association can partly reflect informative censoring.

For the perspective indexed as PMID 41990334, the source evidence should identify which estimands accommodate time-varying surveillance, competing events, and interventions after genotype disclosure. Reporting cause-specific and cumulative-incidence estimates under alternative censoring assumptions would show whether the proposed modifier changes disease occurrence or only the probability that disease is observed.