Time-related factors in the study of risks in animals and humans

Data from epidemiological studies of humans exposed to potentially harmful substances are usually analyzed using methods that account for the dependence of risks on time-related factors such as age and follow-up period. Recently developed statistical procedures allow modeling of the age-specific risks as a function of dose as well as factors such as age at exposure, time since exposure, exposure duration, and dose rate. These procedures potentially allow more rigorous inferences and clearer understanding of the patterns of risk observed in epidemiological studies than has been available in the past. Statistical procedures that consider time-related factors can also be applied to laboratory animal data, providing information that is useful for the problems involved in extrapolating from animal studies to humans. By applying such procedures to data on exposure to the same substance in different species (including humans) or to different substances in the same species, better understanding of the relationship of risks across species and across substances can be achieved. In addition, such statistical procedures allow appropriate treatment of exposure that is accumulated over time and lead to improved understanding of patterns of risk over time. The approach is illustrated using data from a lifespan study of beagle dogs exposed to inhaled Pu.