Risk predictions, validation

However, despite all the problems identified with the earlier simple indices and the question of whether the entire premise for pHs calculations and subsequent scale and corrosion risk prediction is valid, or even particularly relevant to modern day cooling systems, is to miss a vital point. The fact is that Langelier and Ryznar have been with us for more than 50 years and, despite their obvious limitations, there are no other easy to use or easy to understand replacements currently available ... 

The robustness of a quality standard can be increased considerably by accounting for the bioavailability of the substance in an approach that is more accurate than use of added risk. The validation and implementation of international standards for such approaches is desirable. Bioavailable concentrations are influenced by the physical and chemical properties of a substance, soil characteristics, weather, and land use. Further efforts are necessary to gain understanding of the main regulating processes of availability and bioavailability to improve the prediction of risks for substances in soil and to derive more accurate SQS. 

Glance LG, Lustik SJ, Hannan EL, Osier TM, Mukamel DB, Qian F, et al The Surgical Mortality Probability Model derivation and validation of a simple risk prediction rule for noncardiac surgery. Ann Surg 2012 255 696-702. 

The importance of construct validity. It is possible that a direct relationship between a predictor (such as the measure of accident proneness) and a criterion (such as the number of at-risk behaviors or recordable injuries) can be found (predictive validity) without supporting the underlying principle(s) or theory. This would indicate the absence of construct validity. Suppose, for example, an individual could figure out how to answer the survey questions in order to receive a favorable score. Then, construct validity would be questionable, even if criterion validity were high. 

CIBA GEIGY Corporation is presently using models as an aid to data interpretation for risk assessment. Our general philosophy is to use the model as an aid to risk assessment and not as a predictive tool to eliminate definitive studies. Hopefully, environmental fate models will be useful as a predictive tool as they become validated. 

Eq.(l) is not valid, which means that the linear-no threshold theory fails, grossly over-predicting the cancer risk at low doses. 

The Perbellini PBPK model for -hexane is the only validated model for this chemical identified in the literature. The Fisher model was intended for risk assessment to predict which of 19 volatile organic chemicals may be present in milk at a high enough level after workplace exposure to raise health concerns for a nursing infant. 

Genetic applications of molecular technologies have different concerns than routine clinical laboratory testing because results are used in a predictive manner (e.g., the probability that a gene is present and, if present, whether it will be expressed and to what extent of severity). Therefore, validation procedures may differ, depending on whether the test will be used in high-risk situations or in the general population. 

In conclusion, the authors of the cited studies all agree that further research into environmental risk assessment of hospital effluents, incorporating different types of substances used in care and diagnostic activities, as well as cleaning operations (pharmaceuticals, detergents, disinfectants, heavy metals, macropollutants), is vital. Moreover, further studies need to be focussed on evaluating the risk posed by pollutant mixtures, and work is needed to validate the predictive models proposed thus far [19, 49], to evaluate chronic toxicity due to PhCs and then-mixtures and to provide experimental data pertaining to specific case studies. 

The OECD Database on Chemical Risk Assessment Models includes information on models (computerized or capable of being computerized) that are used by OECD Member governments and industry to predict health or environmental effects (e.g., QSARs), exposure potential, and possible risks. The methods described have not been evaluated or validated by OECD. 

Preliminary predictions of absorption of a substance can be made from its physico-chemical properties if no other information is available. Also elaborate computer programs are available that make predictions about, e.g., dermal penetration or metabolic pathways. However, these systems have often not been extensively validated against appropriate experimental data and it is not always certain if the results obtained in such models reflect the situation in vivo. On this basis, modeled data should only be used for risk assessment purposes where it is supported by other kinds of evidence. 

For risk assessment of chemicals based on a DNEL as described above, the output of the exposure assessment (Chapter 7) is usually an estimate of dose or concentration. Exposure data can either be measured or predicted. Measured exposure data are preferred if they are valid, i.e., an actual exposure estimate. However, in most cases, measured data are not available and therefore model-generated data must be used for the risk characterization, i.e., a predicted exposure estimate. 

---