Uncertainty Assessment

Eriksson, L., Jaworska, Worth, A.P., Cronin, M.T.D. and McDowell, R.M. (2003). Methods for reliability and uncertainty assessment and applicability evaluations of classification- and regression-based QSARs. Environmental Health Perspective 111 1361-1375. 

Marini, R. D., Rozet, E., Vander Heyden, Y., Boulanger, B., Bouklouze, A., Servais, A.-C., Fillet, M., Crommen, J., and Hubert, Ph. (2007). Robustness testing of a chiral NACE method for R-timolol determination in S-timolol maleate and uncertainty assessment from quantitative data. J. Pharm. Biomed. Anal. 44, 640—651. 

T.R.M. De Beer, W.R.G. Baeyens, A. Venneire, D. Broes, J.R Remon, and C. Vervaet, Raman spectroscopic method for the determination of medroxyprogesterone acetate in a pharmaceutical suspension Validation of quantifying abilities, uncertainty assessment and comparison with the high performance liquid chromatography reference method. Anal. Chim. Acta, 589, 192-199 (2007). 

In addition to considering the FQPA-relevant areas of uncertainty, assessments of pesticide risk to children also consider applying part or all of the FQPA factors in certain situations to account for... 

More than ever before, an industrial gases company needs to learn to develop a competitive investment strategy under conditions of uncertainty. It must identify these uncertainties, assess future market scenarios and their implications, and explicitly consider the roles it might play in influencing these scenarios. The company then must create a strategic portfolio that involves a combination of actions that, collectively, are robust enough to respond to the uncertainty in the marketplace and actively manage this portfolio as uncertainties resolve themselves and the marketplace evolves. 

The achievement of smaller uncertainties than needed is usually uneconomical. In a practical way, realistic uncertainty assessments in relation to true requirements lead to economically sound planning for measurements to be fit for their intended purpose. 

L. Eriksson et al., Methods for reliability, uncertainty assessment, and applicability evaluations of regression based and classification QSARs. Environ. Health Perspect. Ill, 1361-1375 (2003)... 

Higher-tier assessments do not require the quantification of every uncertainty. Therefore, a tiered approach is proposed where each individual source of uncertainty in an assessment may be treated at one of three tiers, beginning with qualitative approaches (Tier 1) and progressing to deterministic (Tier 2) or probabilistic approaches (Tier 3) when appropriate. Within a single uncertainty assessment, different sources of uncertainty may be treated at different tiers. Higher-tier methods are targeted on those uncertainties that have most influence on the assessment outcome. It is never practical to treat all uncertainties probabilistically therefore, even in a very refined assessment, some uncertainties will still be treated at the lower tiers. 

The position of exposure and uncertainty assessment in the risk communication process... 

Uncertainties assessed at Tier 1 (qualitative) may be communicated by listing or tabulating them, together with an indication of their direction and magnitude. Possible formats for this are illustrated in chapter 5. In addition, it will generally be desirable to give a more detailed textual discussion of the more important uncertainties in the list and of their combined effect on the assessment outcome. 

Uncertainties assessed at Tier 2 (deterministic) generate alternative point estimates for exposure and may be communicated in various ways, depending on the particular methods used for sensitivity analysis. As a minimum, this should identify which sources of uncertainty have been treated at Tier 2, state and justify the alternative quantitative estimates used for each one (e.g. minimum, maximum and most likely values), present exposure estimates for those combinations of alternative estimates that are considered plausible and state and justify any combinations of estimates that are considered implausible. In addition, it will be useful (especially if upper estimates exceed levels of concern) to show which of the quantified uncertainties have most influence on the outcome. 

Uncertainties assessed at Tier 3 (probabilistic) produce probability distributions as outputs. Probability distributions can be communicated in many ways, including ... 

This is an example exposure assessment that illustrates quantitative representations of uncertainty and variability at the higher tiers of an exposure assessment. This case-study is based on human exposures to a persistent, bioaccumulative and lipid-soluble compound through fish consumption. This compound is fictional and referred to here as PBLx, but it has properties that correspond to those of known persistent compounds. Specific goals of this case-study are to illustrate (1) the types of uncertainty and variability that arise in exposure assessments, (2) quantitative uncertainty assessment, (3) how distributions are established to represent variability and uncertainty, (4) differences among alternative variance propagation methods, (5) how to distinguish uncertainty from variability and (6) how to communicate the results of an uncertainty analysis. 

Several key issues in risk assessment of chemical mixtures were identified, that is, exposure assessment of mixtures (e.g., mixture fate and sequential exposure), the concept of sufficient similarity, mixture interactions, QSARs, uncertainty assessment, and the perception of mixture risks. Resolving these key issues will significantly improve risk assessment of chemical mixtures (see next section). 

Closer inspection of Equation A1.4 shows that substances with a high expected risk ratio (nE//iRfD) contribute most to the uncertainty (or variance) in the HI. If 1 or 2 components dominate the mixture, it seems sufficient to base the uncertainty assessment on these dominant components. However, mixtures are often dominated by more than 2 components. Furthermore, the covariance between the individual risk ratios should not be ignored, since exposure estimates (E,) of individual mixture components can be (positively) correlated, as well as their reference values IA>fDr). The uncertainty in the HI may be severely underestimated if these correlations are not accounted for, which is evident from the last part of Equation A1.4. The central limit theorem states that the final HI will approach a normal distribution when the number of substances in the mixture becomes large or if no single risk ratio dominates the sum (De Groot 1986). 

In both human and ecological risk assessment, there is considerable scientific latitude to develop novel methods (e.g., those that exist in only one of the subdisciplines could be useful in the other one) and to refine approaches (e.g., by considering complex reaction networks and more specific attention for modes of action). The refinements are needed to improve the scientific evidence that is available for underpinning risk assessments. Several key issues in risk assessment of chemical mixtures were identified, that is, exposure assessment of mixtures (e.g., mixture fate and sequential exposure), the concept of sufficient similarity (requires clear criteria), mixture interactions, QSARs, uncertainty assessment, and the perception of mixture risks. Resolving these key issues will significantly improve risk assessment of chemical mixtures. 

Van der Sluijs, J.P., Risbey, J.S., Kloprogge, R, Ravetz, J.R., Funtowicz, S.O., Corral Quintana, S., Guimaraes Pereira, A., De Marchi, B., Petersen, A.C., Janssen, P.H.M., Hoppe, R. and Huijs, S.W.F. (2003) RIVM/MNP guidance for uncertainty assessment and communication. Report No. NWS-E-2003-163. Copernicus Institute for Sustainable Development and Innovation and Netherlands Environmental Assessment Agency, Utrecht and Bilthoven. 

In estimating the cumulative risk of a chemical in LCA, dose-response extrapolations can be based on toxicological benchmarks. Such a benchmark approach is considered more appropriate for use in comparative assessment contexts, such as in an LCA study. Benchmarks are an exposure measure associated with a consistent change in response, such as the 10% or even the 50% effect level. Regulatory-based measures do not necessarily provide a consistent risk basis for comparison, as they were often never developed for use in such a comparative context or to facilitate low dose-response extrapolation. Other data differences include the use of median, rather than extreme, data in the fate and exposure modeling, as well as the consideration of safety factors only as part of the uncertainty assessment and not as an integral part of the toxicological effects data. 

An uncertainty assessment is an integral part of the characterization of exposure. In the majority of assessments, data will not be available for all aspects of the characterization of exposure, and those data that are available may be of questionable or unknown quality. Typically, the assessor will have to rely on a number of assumptions with varying degrees of uncertainty associated with each. These assumptions will be based on a combination of professional judgment, inferences based on analogy with similar chemicals and conditions and estimation techniques, all of which contribute to the overall uncertainty. It is important that the assessor characterize each of the various sources of uncertainty and carry them forward to the risk characterization so that they may be combined with a similar analysis conducted as part of the characterization of ecological effects. 

Greenland, S. (2001). Sensitivity analysis, Monte Carlo risk analysis, and Bayesian uncertainty assessment. Risk Anal 21, 579-583. 

During the whole process it is important to involve stakeholders, to ensure that dieir interests and views are correctly understood in the problem identification and problem solution phases. They must also share the conceptual understanding of the system and have confidence in the scientific basis of the assessments of the programme of measures. Furthermore, it is important to evaluate the uncertainty in the different steps and tools used, for instance in the model simulations and in the data collected. Other factors such as social aspects, socioeconomics and restrictions in terms of economy, etc. have to be taken into account in the implementation plan. Stakeholder involvement and uncertainty assessments are not shown as individual steps in die flow chart in Figure 4.1.1, because they should be considered through the whole process. 

First of all, the decision must be made whether and where models are to be apphed and what types of model (e.g., detailed, parsimonious) could be used. The most important selection criterion is the required accuracy of the results if there is demand for very accurate and detailed model results, a more sophisticated model has to be applied, and relevant data have to be collected accordingly (Hpjberg et al., 2006). Important aspects should be uncertainty assessment and quality assurance. 

After the individual discipline modelling and uncertainty assessment, a phase of overall... 

Grunwald, S., K. R. Reddy, S. Newman, and W. E. DeBusk. 2004. Spatial variability, distribution, and uncertainty assessment of soil phosphorus in a south Florida wetland. Environmetrics 15 811-825. 

---