Risk prediction

Kapustka, L.A., Williams, B.A., and Fairbrother, A. (1996). Evaluating risk predictions at population and community levels in pesticide registration. Environmental Toxicology Chemistry 15, 427-431. 

C. A., Murphy, D. J., Lotring, T., Damiano, A., and Harrell, F. L, Jr., The APACHE ID prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 100, 1619-1636(1991). 

In this new scenario much attention is being paid to the investigation of a series of markers of inflammation as reliable indicators of coronary risk. Their value is stressed by the observation that up to one third of events occurs in subjects without traditional risk factors. The C-reactive protein (CRP) seems to provide the strongest risk prediction for CHD in women (Albert 2000 Ridker 2001), although homocysteine, interleukin-6 (IL-6), and lipoprotein (a) [ Lp (a) ], among others, have each been independently associated with increased risk for CHD in women (for a review see Davison and Davis 2003 Rader 2000). 

According to Cardiovascular Risk Prediction Charts (BNF 2009), an elderly, non-diabetic, male smoker has a cardiovascular risk of 10-20% over the next 10 years. The use of statins to control hyperlipidaemia and reduce risk is recommended. 

The partition coefficient (K ) was used for the conversion of TBT concentrations in sediment and SPM to TBT concentrations in water. This Kp for TBT was calculated by multiplying the organic carbon partition coefficient (K ) with the measured fraction organic carbon (f ). Consequently, the K value has a strong impact on the final results of the risk prognosis. Generally, in risk predictions the lowest K value is applied to calculate concentrations in water (EC, 2003). This results in a worst-case water concentration, in accordance with the precautionary principle. However, with literature values for the K of TBT ranging from 3.0 - 6.2 (Lepper, 2002), it is more appropriate to base an assessment on local measured values. In this study the... 

LOE - Chemistry Predicted risk Predicted ISI LOE - Toxicology Observed ISI LOE - Ecology Presence of gastropods at all locations Likely ecological health status in relation to TBT Action... 

While substantial improvements have been made in risk stratification, current risk stratification algorithms still fail to identify prospectively the 20% of patients at risk for relapse. Likewise, these risk stratification methods imperfectly predict treatment toxic-ities, particularly less common or dose independent ones. The use of germline genetic data in risk stratification procedures has the potential to improve risk prediction and may ultimately allow for even more individualized treatment choices. 

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 ... 

Bonassi, S., and W.W. Au. 2002. Biomarkers in molecular epidemiology studies for health risk prediction. Mutat. Res. 511(l) 73-86. 

Robust epidemiologic evidence has identified an inverse relationship between HDL-cholesterol levels and CHD risk. Indeed, HDL-cholesterol is included in the Framingham CHD risk prediction scores (29). HDL-cholesterol protects against... 

A prognostic focus, with the intention of developing methods for effect and risk prediction... 

In single-species risk prediction for individual toxicants and toxicant mixtures, the effect is expressed as the proportion of an exposed population that is likely to be somehow affected by toxic action (quantal responses), or as a reduction in performance parameters such as growth, clutch size, and juvenile period (continuous responses). Both concentration addition- and response addition-based methods are commonly applied for both response types. Assemblage-level risk prediction has only been introduced more recently (e.g., De Zwart and Posthuma 2005) and is founded on similar principles while focusing on the fraction of species that are likely affected by mixture exposure. 

In order to calculate cardiovascular risk for a primary prevention patient such as Mr HA, use a validted risk calculator. These are JBS CVD Risk Predictor Charts (Heart, 2005, 91 1-52) BNF Extra (contains JBS CVD risk prediction programme. Available at http //www.bnf.org/bnf/extra/current/450024.htm) QRISK (Available at http //www.qrisk.org/). 

Figure 5.12 The principle of tiering in risk assessment simple questions can be answered by simple methods that yield conservative answers, and more complex questions require more sophisticated methods, more data, and more accurate risk predictions. PEC = Predicted Environmental Concentration, PNEC = Predicted No Effect Concentration, HI = Hazard Index, CA = Concentration Addition, RA = Response Addition, TEF = Toxicity Equivalency Factor, RPF = Relative Potency Factor, MOA = Mode of Action, PBPK = Physiologically Based Pharmacokinetic, BRN = Biochemical Reaction Network.

There is a clear need for mixture risk assessment, since most environments are characterized by mixture exposure situations. This means that risk assessments should pay specific attention to all aspects of mixture exposures and effects in order to make accurate risk predictions. 

The risk assessment for mixtures shows much similarity with that for single substances, but there also are some important differences. In order to make accurate risk predictions, risk assessment should pay specific attention to all aspects of mixture exposures and effects. The establishment of a safe dose or concentration level for mixtures is useful only for common mixtures with more or less constant concentration ratios between the mixture components and for mixtures of which the effect is strongly associated with one of the components. For mixtures of unknown or unique composition, determination of a safe concentration level (or a dose-response relationship) is inefficient, because the effect data cannot be reused to assess the risks of other mixtures. One alternative is to test the toxicity of the mixture of concern in the laboratory or the field to determine the adverse effects and subsequently determine the acceptability of these effects. Another option is to analyze the mixture composition and apply an algorithm that relates the concentrations of the individual mixture components to a mixture risk or effect level, which can subsequently be evaluated in terms of acceptability. 

Calibration is assessed by comparing the observed with the predicted proportions of events in groups defined by the risk prediction or score. 

In conclusion, the ABCD and ABCD scores are reliable tools to predict the early risk of stroke after TIA. However, although they are sensitive and easily calculable using clinical information readily available at the time of assessment, they have a high false-positive rate and were deliberately designed to include only clinical data so that they could be used for initial triage. Further information may, therefore, be required to refine risk prediction. 

The role of perfusion imaging in short-term risk prediction after TIA and minor stroke is uncertain (Latchaw et al. 2003). 

Several published models have addressed risk prediction in the medium and long term following TIA but many have methodological flaws (Hankey and Warlow 1994). Three models that have been proposed for use in the targeting of longer-term secondary prevention for patients after initial TIA are reviewed below (Table 17.1). 

Risk prediction in specific conditions following stroke or TIA using modeling is helpful in targeting secondary preventive treatments that might themselves be associated with benefit and harm. Models can provide data on the risk for a specific patient, which can then more reliably inform the risk - benefit ratio for that individual and guide decision making about treatment. Risk models have been developed in symptomatic carotid disease (Ch. 27) and atrial fibrillation (Ch. 14) and these will be discussed below. 

Risk prediction in carotid stenosis is more fully discussed in Ch. 27. 

Increased stroke risk predicted by compromised cerebral blood flow reactivity. Journal of Neurosurgery 79 483-489 Youkey JR, Clagett CP, Jaffin JH et al. (1984). Focal motor seizures complicating carotid endarterectomy. Archives of Surgery 119 1080-1084... 

---