AgePeriodCohort

Age-period-cohort (APC) analysis is a statistical framework used to study how three temporal dimensions—an individual's age, the historical period during which data are observed, and the birth cohort to which the individual belongs—shape patterns in health, behavior, and social outcomes. The age effect captures systematic change associated with aging; period effects reflect influences that affect all ages at a given time, such as medical advances, outbreaks, or policy changes; cohort effects reflect differences among groups born in the same time window due to shared experiences or exposures.

Because cohort equals period minus age, these three dimensions are linearly dependent, creating an identification problem that prevents unique estimation of all three effects without additional assumptions. Researchers address this by imposing constraints, using alternative parameterizations, or applying models such as the intrinsic estimator, Holford's constrained models, or Bayesian approaches that incorporate prior information or smoothing.

APC analysis is widely applied to cancer incidence and mortality, fertility and mortality trends, smoking behavior, obesity, and other risk factors, to distinguish aging processes from historical shifts and generational differences. It helps interpret secular trends and to inform policy by highlighting whether changes are predominantly age-related, period-driven, or cohort-specific.

Limitations include sensitivity to model choice, potential overinterpretation of constrained components, and reliance on long-term data with stable measurement. Because of the identifiability problem, APC results are best viewed as plausible, not definitive, explanations of temporal patterns.