RISK DATA GENERATION

Box 6.1 Successful risk data generation and safety communication schemes... 

General societal risk data. It is important to consider the context of societal risk data. Some particularly important factors include whether or not the risk is voluntary, and whether persons exposed to the risk derive any benefit from the activity that generates the risk. Covello, Sandman, and Slovic, Slovic, and Wilson and CroLich provide examples of general societal risk data and discuss risk comparison and perception. 

Thus it was clear that drug metabolism can differ between ethnic groups, and data generated in one population cannot be directly extrapolated to another population. When such differences exist, one ethnic group may be at an increased risk of therapeutic failure or toxicity because of differences in drag metabolism. 

The methodology for conducting aquatic model ecosystem studies was well established by the late 1990s. However, the use of the data in risk assessments raised a number of uncertainties regarding their interpretation and implementation [32]. Four of the uncertainties that were identified were the extent to which aquatic model ecosystem data generated in one location could be applied to another situation, the potential influence of mixtures of chemicals or stressors, whether the timing (season) of application would influence the outcome of the study, and whether differences in ecosystem properties (e.g., trophic status) might influence the results. 

Failure rate data generated from collecting information on equipment failure experience at a facility are referred to as facility-specific or field failure rate data. Facility-specific data contain failure rates specific to equipment (e.g., a certain valve or pump in use at a facility by manufacturer, make, model, and serial number) and are cataloged accordingly. The collection of facility-specific data from internal operations for use in a risk analysis is desirable because such data reflect the practices, environmental factors, and other reliability influences specific to the equipment under study. The ideal situation is to have valid historical data from identical equipment, in the identical application, functioning under the identical operating and maintenance conditions. Where these are not available, but data on similar equipment are, then they may be used with appropriate judgment. 

It is possible to conduct animal studies in an infinite number of ways. Although individually designed studies are often scientifically sound, and in many cases serve a particular purpose very well, they pose problems in a regulatory context. Free movement of chemicals between countries is based on the mutual acceptance of the risk evaluation made by each country and this, in turn, relies on the mutual acceptance of the data generated when testing the chemicals. Experience has shown this acceptance to be extremely difticult, if chemicals have been tested by different methods. 

Determine whether there are more cost-effective alternatives to additional data generation and risk assessment refinements. What-if analyses can be used to examine the savings in risk management that might result from additional data generation. Techniques that may be suitable for this include Bayesian Monte Carlo and expected value of information (EVOI) analysis (Dakins et al. 1996). 

Therefore, the data generated from them have to be viewed with caution. This is particularly the case if the data are being used as part of a risk assessment. Such in vitro data may underestimate the toxicity in vivo. 

The second answer with respect to the earlier acquisition of risk data must come from the laboratory. Not from animal studies, which in this field are of very restricted value, but from biochemical and particularly hematological work. When during the 1990s various groups began to examine in detail the effects of the third-generation contraceptives on processes related to the clotting system, they identified a series of properties that could very well explain an increased incidence of thrombosis. 

In the case of atrazine, the most widely used triazine, refinement of the conservative Tier I estimate for all populations results in at least a 200-fold reduction in exposure and risk. In addition, the total exposure and risk for atrazine using a Tier III approach is approximately 138000-fold to 450000-fold less than the chronic NOAEL obtained from a rodent study. A 600-fold to 1600-fold difference exists between the NOAEL and maximum theoretical exposure and risk when tolerance values are utilized. Further refinement of the estimate using Tier IV methodology would result in even lower exposure and risk, since this Tier III analysis used data generated from structured field trials (maximum label rate and minimum preharvest interval). 

The U.S. EPA and many other organizations point out that, when information about potential risks is incomplete, basing decisions to avoid unnecessary health risks is potentially the best option.8 When a good set of scientific data is available on a material, then the Precautionary Principle is not appropriate. Scientific data generated in the EU Risk Assessments or under risk assessment programs, such as REACH, that deem materials safe for continued use should effectively rule out the use of Precautionary Principle. 

Some believe that risk assessment is not complex or difficult while others believe that it is complicated and hard to understand. Irrespective of a person s exposure to the risk assessment process, risk assessment procedures are complicated and complex beyond the scope of individual person s input. Simply put, risk assessment is both a science and an art. Risk assessment has properties of science, because the process totally depends on data generated by good scientific practice. The difficulty is, while scientists can possibly wait until final conclusions are reached using agreed scientific methods, the society and the risk manager cannot wait that long. The science and art aspect of risk assessment is illustrated in the table below. 

The registration phase as such does not necessarily lead to reduced risks, and the quality of the data generated is not guaranteed. The ECHA is merely required to make a completeness check within three weeks (Article 20), after which the chemical in question can be used. 

Much of the data generated has been as a result of advances in trace analysis in different environments, linked to a lack of understanding between hazard and risk (probability of intrinsic hazard causing an effect). 

---