Risk estimates, uncertainty bounds

Practitioners of ecological risk assessments will frequently experience large uncertainty bounds on the estimates of risk. Unfortunately, characterizing and/or reducing uncertainty can be very costly. However, these costs must be balanced with the need to conduct sufficient analysis to make an informed decision. 

Communication between risk managers, risk assessors, and analysts is essential from the start of the assessment process, not just in communicating results. For example, the choice of uncertainty analysis methods will be dependent on 1) the questions posed by decision makers, 2) the closeness of the risk estimate and its bounds to thresholds of acceptability or unacceptability, 3) the type of decision that must be made, and 4) the consequences of the decision. 

The ADI is generally viewed by risk assessors as a soft estimate, whose bounds of uncertainty can span an order of magnitude. That is, within reasonable limits, while exposures somewhat higher than the ADI are associated with increased probability of adverse effects, that probability is not a certainty. Similarly, while the ADI is seen as a level at which the probability of adverse effects is low, the absence of all risk to all people cannot be assured at this level. 

To provide the most realistic representation of risk, all forms of uncertainty arc considered. Rather than assuming the existence of some representative condition prior to the accident scenario, a study models the full range of conditions and other uncertainties that can affect the scenario. Results include uncertainties in the frequency and consequences of each scenario. The upper uncertainty bound shown for the QRA risk estimates is a measure of the analysts confidence in the results. There is a 95 percent chance that the risk is less than the upper bound. 

The primitiveness of methodology echoes the lack of a clear theory of carcinogenesis. However, the gaps in knowledge and the uncertainties in methods do not constitute sufficient justification for abandoning efforts to provide the public with plausible upper bounds for cancer risks due to environmental chemical exposures. For a large number of such exposures, these estimates will necessarily be based on animal data. When quantified human exposure data are available and are related to cancer risk, these data can be useful either as a basis for extrapolation or as a standard for assessing the plausibility of risk estimates based on animal data alone."... 

A common risk evaluation and presentation method is simply to multiply the frequency of each event by consequence of each event and then sum these products for all situations considered in the analysis. In insurance terms, this is the expected loss per year. The results of an uncertainty analysis, if performed, can be presented as a range defined by upper and lower confidence bounds that contain the best estimates. If the total risk represented by the best estimate or by the range estimate is... 

According to the WHO (WHO/IPCS 1994, 1999 WHO 1996, 2000), it should be noted that cmde expression of risk in terms of excess incidence or numbers of cancers per unit of the population at doses or concentrations much less than those on which the estimates are based may be inappropriate, owing to the uncertainties of the quantitative extrapolation over several orders of magnitude. Estimated risks are therefore considered to represent only the plausible upper bounds and vary depending upon the assumptions on which they are based. 

An initial step in addressing such situations should be the performance of an analysis of the sensitivity of a risk assessment model to changes in the variable. If the model proves relatively insensitive to conservative bounds to the variable, then further consideration of uncertainty for this variable may be unnecessary and a point estimate may suffice. 

The risk interpretation of biomonitoring results will tend to have additional uncertainties. That is because, in addition to the standard uncertainties encountered in risk assessment, there is the uncertainty of extrapolating from a blood or urinary concentration to an external dose. There will be variability both in the timing between sample draw and most recent exposure and in the relationship between blood concentration and dose. Those kinds of variability are compounded by uncertainty in the ability of a PK calculation or model to convert biomarker to dose accurately. For example, reliance on urinary biomarker results expressed per gram of urinary creatinine leads to an uncertain calculation of total chemical excretion per day because of the considerable variability in creatinine clearance per day. That complicates an otherwise simple approach to estimating dose. Furthermore, the conversion requires knowledge of fractional excretion via various pathways, which may not be present for a large sample of humans. The uncertainties created by these factors can be bounded via sensitivity and Monte... 

UCL takes into account measurement uncertainty in the study used to estimate the dose-response relationship, such as the statistical uncertainty in the number of tumors at each administered dose, but it does not take into account other uncertainties, such as the relevance of animal data to humans. It is important to emphasize that UCL gives an indication of how well the model fits the data at the high doses where data are available, but it does not indicate how well the model reflects the true response at low doses. The reason for this is that the bounding procedure used is highly conservative. Use of UCL has become a routine practice in dose-response assessments for chemicals that cause stochastic effects even though a best estimate (MLE) also is available (Crump, 1996 Crump et al., 1976). Occasionally, EPA will use MLE of the dose-response relationship obtained from the model if human epidemiologic data, rather than animal data, are used to estimate risks at low doses. MLEs have been used nearly universally in estimating stochastic responses due to radiation exposure. 

Some risk assessors and risk managers followed these initial efforts by beginning to seek analyses that are more sophisticated. Instead of just an upper bound worst-case overestimate, they wanted more information on the actual range of variation of the exposure, and some measure of the uncertainty associated with the estimate. 

In addition to uncertainty in exposure, dose, and cancer estimations, the possibility of multiple simultaneous exposures to different chemicals needs to be considered. Risks from simultaneous exposure to more than one carcinogenic substance are typically estimated by assuming that the individual risks are additive. This process assumes that intakes of individual substances are relatively small, that there are no synergistic or antagonistic chemical interactions, and that all chemicals produce the same toxic effect (US EPA 1989). However, because health risks from exposures to chemical mixtures are generally based on a combination of upper-bound risks calculated for individual compounds, these risk assessments tend to be overly conservative (Gaylor and Chen 1996 Hwang and Chen 1999 Kodell and Chen 1994). 

Before leaving the topic of systematic error bounds, two points should be made. First, as Is perhaps obvious, the probabilistic meaning of false positives and false negatives Is necessarily altered. These "errors or risks are now Inequalities ["no greater than..."], and their validity rests greatly on that of the systematic error bounds, just as in the case of uncertainty intervals for high level signals. Second, estimation of non-Poisson random error and systematic error empirically, by comparison and replication is not an easy task. One can show that at least 15 and 47 replicates, respectively, are necessary just to detect systematic and excess random error components equivalent to the (Poisson) standard deviation [(12), p 25f (13)1. 

The US EPA has subsequently published a comprehensive toxicological review of bromate (US EPA, 2001). Studies with rats based on low-dose linear extrapolation, using the time-to-tumour analysis, and using the Monte Carlo analysis to sum the cancer potency estimates for kidney renal tubule tumoms, mesotheliomas, and thyroid follicular cell tumours, gave an upper-bound cancer potency estimate for bromate ion of 0.70 per mg/kg day. This potency estimate corresponds to a drinking water unit risk of 2 x 10 per pg/L, assuming a daily water consumption of 2 litres/day for a 70-kg adult. Lifetime cancer risks of 10 , 10 , and 10 are associated with bromate concentrations of 5, 0.5, and 0.05 pg/L, respectively. A major source of uncertainty in these estimates is from the interspecies extrapolation of risk from rats to humans. 

With regard to carcinogenicity, a weight-of-the-evidence evaluation suggests that dioxin and related compounds (CDDs, CDFs, and dioxin-like PCBs) are likely to present a cancer hazard to humans [157]. While major uncertainties remain, efforts of this reassessment to bring more data into the evaluation of cancer potency have resulted in a risk-specific dose estimate (1 x 10 risk or one additional cancer in one million exposed) of approximately 0.01 pg TEQ/kg body weight/day. This risk-specific dose estimate represents a plausible upper bound on risk based on the evaluation of animal and human data. "True" risks are not likely to exceed this value, may be less, and may even be zero for some members of the population. 

---