Measurement Methods for Dose-Response Analysis
In the field of risk assessment, estimating the quantitative relationship between a chemical dose and a response is a critical component. Two powerful tools that aid this process are the Benchmark Dose Software (BMDS) Suite and Categorical Regression (CatReg). Both are part of EPA's BMD tools and offer various models for benchmark dose modeling.
The BMDS Suite, which includes BMDS Online, BMDS Desktop, and pybmds, provides easy access to numerous mathematical models. In its latest interim release, the Suite has undergone enhancements, adding the NCTR nested dichotomous model for developmental toxicity data, a Rao-Scott transformation tool, and the ability to clone an analysis, as well as an online BMDS User Guide.
The Rao-Scott transformation tool is designed to improve analysis involving categorical data. It applies the Rao-Scott correction, which adjusts for complex survey design or clustering in the data, effectively transforming the test statistics to better approximate the chi-square distribution under these conditions. This correction helps to reduce bias and improve the accuracy of inferential statistics when dealing with categorical variables.
Categorical Regression (CatReg) complements the BMDS Suite by providing categorical regression analysis on toxicity data. While CatReg fits regression models to categorical outcomes to estimate benchmark doses, the Rao-Scott transformation tool within the BMDS Suite adjusts the test statistics derived from these models to properly account for the survey design effects or correlated data structures. This combination ensures robust dose-response modeling with appropriate statistical adjustments, leading to more reliable benchmark dose estimates.
It is important to note that while later versions of R limit some CatReg functionality, the Windows-based CatReg application requires R for Windows version 3.4 to 3.4.4 for its operation. For those seeking to learn more about BMDS and CatReg, additional resources include training videos, aids beyond the software documentation, and going beyond the user guide.
In summary, the BMDS Suite and Categorical Regression (CatReg) are valuable tools in the risk assessment process, providing accurate modeling and statistical inference for categorical dose-response data. The Rao-Scott transformation tool within the BMDS Suite enhances CatReg by providing adjusted statistical tests that respect the data's underlying sampling design or correlation, enabling more valid conclusions from categorical dose-response analyses within the BMDS framework.
The BMDS Suite, integrated with various mathematical models, provides access to improved models, such as the Rao-Scott transformation tool and the NCTR nested dichotomous model, thus aiding in the analysis of categorical data within the field of risk assessment. Alongside this, Categorical Regression (CatReg) offers categorical regression analysis on toxicity data, complimenting the BMDS Suite and ensuring robust dose-response modeling with appropriate statistical adjustments, making it a significant tool in the realm of science and technology.
