Definition A random effects model for time variables, where the random effect—frailty, has a multiplicative effect on the hazard. It can be used for univariate (independent) failure times, and describe the influence of unobserved covariates in a proportional hazards model. It can also be used for multivariate (dependent) failure times generated as conditionally independent times given the frailty.
Comments The multivariate approach can be used both for survival times of individuals–e.g., twins or family members, and for repeated events in the same individual.
The model, as applied to Danish twin data, suggests that ±50% of the variation in life-spans after age 30 is due to survival attributes fixed for a person by age 30. Half of the fixed changes are due to genetic factors, the other half are due to socioeconomic status, nutrition, medical history.
References Lifetime Data Anal. 1995;1(3):255-73
Science 1998; 280:855rv