Re-examination of the ED01 study--risk assessment using time.

The ED01 study undertaken by NCTR has provided extremely valuable data which should stimulate the development of improved cancer risk assessment procedures. A general model for carcinogenic risk assessment including both dose level and time to response is given herein - including several possible extensions - and shown to be capable of fitting the ED01 data. The need to include time in risk assessment is clearly demonstrated. Some important new definitions of acceptable risk which more realistically reflect the time of the response are presented and illustrated. The analysis of the ED01 data indicates 1. the need to use a relatively fine time partition if each animal's observation time is not treated individually, 2. the considerable room-to-room consistency in risk assessments, and 3. estimated models which are nonlinear in both dose and time for both bladder carcinomas and liver tumors - in contrast to the linearity of the estimated liver tumor multistage quantal response model which ignores time. The capability and need to do time based risk assessment exists now.