Uncertainty assessment procedure

Risk characterization is the last step in the risk assessment procedure. It is the quantitative or semi-quantitative estimation, including uncertainties, of frequency and severity of known or potential adverse health effects in a given population based on the previous steps. Risk characterization is the step that integrates information on hazard and exposure to estimate the magnitude of a risk. Comparison of the numerical output of hazard characterization with the estimated intake will give an indication of whether the estimated intake is a health concern. ... 

The degree of confidence in the final estimation of risk depends on variability, uncertainty, and assumptions identified in all previous steps. The nature of the information available for risk characterization and the associated uncertainties can vary widely, and no single approach is suitable for all hazard and exposure scenarios. In cases in which risk characterization is concluded before human exposure occurs, for example, with food additives that require prior approval, both hazard identification and hazard characterization are largely dependent on animal experiments. And exposure is a theoretical estimate based on predicted uses or residue levels. In contrast, in cases of prior human exposure, hazard identification and hazard characterization may be based on studies in humans and exposure assessment can be based on real-life, actual intake measurements. The influence of estimates and assumptions can be evaluated by using sensitivity and uncertainty analyses. - Risk assessment procedures differ in a range of possible options from relatively unso-... 

Uncertainty may also affect the credibility of ecological risk assessment procedures. In the late 1970s, the USEPA presented a risk assessment for the use of granular carbofuran on com, including a detailed list of field studies and incidents. The Eederal Insecticide, Fungicide, and Rodenticide Act Science Advisory Panel concluded there was insufficient information to justify restricted use labeling and recommended further testing. Nearly 20 years later, the accumulation of additional field studies and incidents provided sufficient evidence such that approvals for use of carbofuran were withdrawn. 

The Community-Level Aquatic Systems Studies Interpretation Studies (CLASSIC) guidance document, which deals with the interpretation of results of microcosm and mesocosm tests in the risk assessment procedure of pesticides, recommends that regulatory model ecosystem experiments be conducted in spring to midsummer (Giddings et al. 2002). On the basis of the limited number of model ecosystem experiments described above, it seems that threshold concentrations for effects observed in early-season studies are reasonably predictive for threshold concentrations later in the season. Above these threshold concentrations, however, the intensity and duration of the responses (direct and indirect effects) may vary during different periods of the year. Consequently, the extrapolation of NOECcommunity values from one season to another seems to be possible with lower uncertainty than hazard estimates of higher concentrations in which both direct and indirect effects are involved. 

The comparative cholinesterase studies, as described, provide infonnation on adverse outcomes (cholinesterase inhibition) following cither acute or repeated exposures in populations at various life stages. In a review of cumulative risk assessment procedures for the OP pesticides (EPA, 2(K)2b), acute and/or repeated dose comparative cholinestera.se data in rats were available for a limited number of OP pesticides. These data raised uncertainties regarding the adequacy of adult risk potency factors to be protective of potential age-dependent. sensitivity to cholinestera.se inhibition and of potential adverse neurodevelopmental outcomes that arc a result of the inhibition t)f cholinesterase. As a re,sull, a thrcc-fold database uncertainty factor was applied in the cumulative risk calculation for those chemicals that did not have comparative cholinesterase data twailable for evaluation. This... 

Risk characterization is the final step in the risk assessment process. It comprises quantitative or semiquantitative estimations, including uncertainties, of the probability of adverse health effects in people associated with exposure to the toxic agents. Risk characterization is based on the information gathered through the first three steps in the risk assessment procedure. It is important that the weight of evidence leading to the conclusions be openly discussed. Risk characterization should include a description of the primary causes of uncertainties. 

It remains to look into assumption Z4. The results are conditional on the use of the model M. We may have some test results for model M, but say that the model is relatively new and its predictabdity is questioned. Then the output quantities of the model may be considered an uncertainty factor. Also in this case we may adopt the two-stage assessment procedure outlined above. 

Note that subcrificaT means that the maximum neutron multiplication, adjusted appropriately by including a calculafional bias, uncertainties and a subcrifical margin, should be less than 1.0. See Appendix vn for specific advice on the assessment procedure and advice on determining an upper subcrifical limit. 

The elements of the matrices F and G are calculated as follows. Each reference value x, / = 1, 2,. .., 5, of Table 2 is interpreted as the median of an HEP uncertainty distribution. Therefore, the HEP assessment problem becomes that of assessing which of the five x/s best represents the median, based on the expert estimates. The error factor (square root of the 95th and 5th percentiles) of the uncertainty distribution is assumed to be known and equal to 5 this significantly simplifies the assessment procedures, because only one parameter has to be assessed (the median). Indeed, the value of 5 for the error factor represents a commonly accepted confidence range for HEPs (Podofillini Dang 2013). 

This discussion uses a typical MC-simulation to illustrate the need of emphasizing foundational issues how to perceive risk and assure quality of the assessment procedure. Students in fields relying on the use of quantitative assessments, such as risk, reliability or system analysis, learn about MC-simulation and different ways to quantitatively describe uncertainty. It is important not to end there, but give the students, and future quantitative assessors, the ability to fully understand what this uncertainty represents and on what scientific principles it can be assessed. We would like to point at Bayesian Evidence Synthesis as a suitable framework to teach predictive statistical principles involving simulations from a computer model. BES is useful both in scientific research and risk assessment and can therefore close the gap between these two highly important processes of knowledge production. 

NRHA is allowed by the NZSEE Recommendations, although they note that NRHA should not be the sole adopted assessment procedure and should be supplemented by the results of a simplified approach. They cite such tmcer-tainties as flexure-shear-axial force interaction, effect of axial force on column stiffness, degrading strength characteristics, and modeled of beam-column joints. Presumably, the restriction on the use of NRHA could be relaxed somewhat if more detailed analysis models are used that reduce these uncertainties. The focus of the NZSEE Recommendations is on linear analysis supplemented by an analysis of expected yielding mechanisms, and therefore the recommendations for NRHA are reasonably limited. 

Although methods have been established to derive these levels (Barnes and Dourson 1988 EPA 1990c), uncertainties are associated with these techniques. Furthermore, ATSDR acknowledges additional uncertainties inherent in the application of the procedures to derive less than lifetime MRLs. As an example, acute inhalation MRLs may not be protective for health effects that are delayed in development or are acquired following repeated acute insults, such as h q)ersensitivity reactions, asthma, or chronic bronchitis. As these kinds of health effects data become available and methods to assess levels of significant human exposure improve, these MRLs will be revised. 

---