Competing Risk of Death - An Important Consideration in Studies of Older Adults

  • Kaplan Meier (KM) survival curve and Cox are used to study disease outcome for unequal follow up time
    • They are only useful when describing all cause mortality
  • Competing risk is another outcome that can alter the chance of primary outcome. Examples:
    • elderly patient die from another comorbidity before benefit from new medicine
    • cancer patient die from something else unrelated to cancer
  • KM will overestimate incidence when there's competing risks
    • patients lost to follow up are "censored" and considered at risk for incidence
    • patients who die are grouped similarly to patient lost to follow up
    • e.g. patient death will add to loss to follow up and drive up incidence
  • remove patients who died in analysis => smaller sample size, survivor bias
  • Cumulative Incidence Competing Risk will adjust incidence with competing risks