DOSE-EFFECT APPROACHES TO RISK ASSESSMENT

Risk assessment is the attempt to characterize the chance of obtaining an adverse effect after exposure to an agent. Traditionally, high levels of an agent have been used to estimate the likelihood a lower dose might have an effect either by using low-dose extrapolation models or by attempting to establish a dose with no observable effects (NOEL). Low-dose extrapolation models yield estimates for small effects, but these estimates may vary by orders of magnitude depending upon the function chosen to represent the data. NOEL's are imprecise because a true no-effect level is indeterminant and the inability to determine an observable effect depends primarily on background variability. Newer methods use data from portions of the dose-effect function where error is smaller to estimate risks. Risk estimates using two of these approaches are compared for two different sample sizes. Each method produced the same estimate with the larger sample at low risk, but with increasing levels of risk and smaller samples the estimates obtained using these methods diverged.