Measuring risks in humans: the promise and practice of epidemiology

Epidemiology has been considered the fundamental science of public health policy. The use of epidemiologic data in environmental health policy has been limited particularly in the environmental regulatory arena. Epidemiologic risk assessment (ERA) is different from risk assessment and the interplay between the two has led to some misconceptions over the promise and practice of epidemiologic data. The current risk assessment process was designed in a time when the need for regulation was great and the epidemiologic information was sparse. There was little time for the consideration of conducting specific studies to improve the information base for environmental health policy. Animal bioassays could be conducted under standardized protocols within defined time periods. The limitations and uncertainties of animal studies also became standardized and risk assessors became comfortable with their models of extrapolation. As the cost of regulations have grown, the economic realities of regulating with little or no data to support actual public health benefit have become a political and legal liability. Major among epidemiology's advantages is that the information is of direct relevance. The majority of epidemiology data are observation and whether the number of studies is broad enough, the data can be generalized to major segments of the population. The uncertainties in animal-based risk assessments are likely to be greater than the uncertainties associated with epidemiologic studies. Another advantage is the range of extrapolation is often smaller. Another advantage is that epidemiologic data include the genetic diversity and variability in other endogenous factors inherent in human populations. The homogeneity of animal studies has often been cited as an advantage but is unrepresentative of the heterogeneity of the human race. Epidemiology does have its limitations. A major limitation is the time needed to obtain a database sufficient for policy-making purposes and the resources needed to conduct the research to develop the database. This has often prompted the conduct of poorly designed studies, forced the use of data collected for other purposes or improper use of existing data. Four situations where epidemiology should be pursued are discussed. Once an environmental health decision has been made, epidemiologic studies should be considered for documenting the reduction of exposure and therefore disease in the population. This traditional use of epidemiology has rarely been applied in the environmental health arena. A final consideration on the use of epidemiologic data is the need to provide a sense of perspective to set priorities in the larger context of public health priorities. The role of environmental pollutants in causing disease may in some cases be minor in comparison to other risk factors and needs to be considered in setting environmental regulations.