Model Uncertainty and Risk Estimation for Experimental Studies of Quantal Responses

Bailer, A. John; Noble, Robert; Wheeler, Marcia A.

doi:10.1111/j.1539-6924.2005.00590.x

Cited by 65 publications

(80 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These criteria are defined as −2 log(L) + K, and −2 log(L) represents the maximum value of the −2 log likelihood for a particular model, K = 2P for the AIC, K = P log(n) for the BIC, and K = 3P for the KIC, here n represents the total sample size (i.e., the number of observations not the number of dose groups), and P represents the number of parameters in the model. These criteria were used in MA for risk assessment by Kang et al (2000), Bailer et al (2005a) Bailer, Wheeler, Dankovick, Noble, and Bena (2005b), and Moon, Kim, Chen, and Kodell (2005). Given one of these criteria the weights are formed using the following formula:…”

Section: Model Averagingmentioning

confidence: 99%

“…Model averaging, for risk assessment, was first proposed by Kang, Kodell, and Chen (2000) for continuous microbial dose-response data, and further considered by Bailer, Noble, and Wheeler (2005a) for dichotomous dose-response data. These papers suggested a model averaged benchmark dose (MA-BMD) formed by taking a weighted average of model specific BMDs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Model Averaging Software for Dichotomous Dose Response Risk Estimation

Wheeler¹,

Bailer²

2008

J. Stat. Soft.

Self Cite

View full text Add to dashboard Cite

Model averaging has been shown to be a useful method for incorporating model uncertainty in quantitative risk estimation. In certain circumstances this technique is computationally complex, requiring sophisticated software to carry out the computation. We introduce software that implements model averaging for risk assessment based upon dichotomous dose-response data. This software, which we call Model Averaging for Dichotomous Response Benchmark Dose (MADr-BMD), fits the quantal response models, which are also used in the US Environmental Protection Agency benchmark dose software suite, and generates a model-averaged dose response model to generate benchmark dose and benchmark dose lower bound estimates. The software fulfills a need for risk assessors, allowing them to go beyond one single model in their risk assessments based on quantal data by focusing on a set of models that describes the experimental data.

show abstract

Section: Model Averagingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Model Averaging Software for Dichotomous Dose Response Risk Estimation

Wheeler¹,

Bailer²

2008

J. Stat. Soft.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The option to perform some form of averaging of BMD (or BMDL) values instead of using the result from a single model has also been discussed to some extent (Fitzgerald et al, 2004;Bailer et al, 2005). For example, Bailer et al (2005) suggests the use of Bayesian model averaging.…”

Section: Selection Of Dose-response Model and Bmdlmentioning

confidence: 99%

“…For example, Bailer et al (2005) suggests the use of Bayesian model averaging. According to this approach a collection of models (e.g.…”

Section: Selection Of Dose-response Model and Bmdlmentioning

confidence: 99%

The current state of knowledge on the use of the benchmark dose concept in risk assessment

Sand

Victorin

Filipsson

2007

J of Applied Toxicology

101

View full text Add to dashboard Cite

This review deals with the current state of knowledge on the use of the benchmark dose (BMD) concept in health risk assessment of chemicals. The BMD method is an alternative to the traditional no-observed-adverse-effect level (NOAEL) and has been presented as a methodological improvement in the field of risk assessment. The BMD method has mostly been employed in the USA but is presently given higher attention also in Europe. The review presents a number of arguments in favor of the BMD, relative to the NOAEL. In addition, it gives a detailed overview of the several procedures that have been suggested and applied for BMD analysis, for quantal as well as continuous data. For quantal data the BMD is generally defined as corresponding to an additional or extra risk of 5% or 10%. For continuous endpoints it is suggested that the BMD is defined as corresponding to a percentage change in response relative to background or relative to the dynamic range of response. Under such definitions, a 5% or 10% change can be considered as default. Besides how to define the BMD and its lower bound, the BMDL, the question of how to select the dose-response model to be used in the BMD and BMDL determination is highlighted. Issues of study design and comparison of dose-response curves and BMDs are also covered.

show abstract

“…This technique, which estimates risk from a weighted average of all models considered, was used by Kang et al (2000) to estimate risk across different microbiological dose response models. Bailer et al (2005a), Bailer et al (2005b) applied a similar technique, Bayesian model averaging (BMA), to animal toxicity studies where no a priori model was assumed. Both averaging methods calculate excess risk from a weighted average of risk estimated from each model considered.…”

Section: Introductionmentioning

confidence: 99%

Comparing model averaging with other model selection strategies for benchmark dose estimation

Wheeler

Bailer

2008

Environ Ecol Stat

Self Cite

View full text Add to dashboard Cite

Model averaging (MA) has been proposed as a method of accommodating model uncertainty when estimating risk. Although the use of MA is inherently appealing, little is known about its performance using general modeling conditions. We investigate the use of MA for estimating excess risk using a Monte Carlo simulation. Dichotomous response data are simulated under various assumed underlying doseresponse curves, and nine dose-response models (from the USEPA Benchmark dose model suite) are fit to obtain both model specific and MA risk estimates. The benchmark dose estimates (BMDs) from the MA method, as well as estimates from other commonly selected models, e.g., best fitting model or the model resulting in the smallest BMD, are compared to the true benchmark dose value to better understand both bias and coverage behavior in the estimation procedure. The MA method has a small bias when estimating the BMD that is similar to the bias of BMD estimates derived from the assumed model. Further, when a broader range of models are included in the family of models considered in the MA process, the lower bound estimate provided coverage close to the nominal level, which is superior to the other strategies considered. This approach provides an alternative method for risk managers to estimate risk while incorporating model uncertainty.

show abstract

Model Uncertainty and Risk Estimation for Experimental Studies of Quantal Responses

Cited by 65 publications

References 15 publications

Model Averaging Software for Dichotomous Dose Response Risk Estimation

Model Averaging Software for Dichotomous Dose Response Risk Estimation

The current state of knowledge on the use of the benchmark dose concept in risk assessment

Comparing model averaging with other model selection strategies for benchmark dose estimation

Contact Info

Product

Resources

About