Dose-response assessment by a fuzzy linear-regression method.

Regression analysis has been used to characterize the relationship between an exposure dose and the incidence of an adverse health effect such as cancer. However, the regression rarely describes the true relationship due to uncertainties in dose-response data and relationships. Therefore, a method is developed to perform dose-response assessments by a fuzzy linear regression which explicitly exhibit these uncertainties. This method is applied to define the relationship between a particular nitrate dose to humans and its corresponding cancer risk.