Sunday Full-Day Workshops

1. Approaches to assessing environmental justice: perspectives from the scientific, regulatory and regulated communities

Speaker: Unni Blake

Description: The environmental justice (EJ) movement arose from community concerns surrounding how people of color and/or low-socioeconomic status have borne the disproportionate impacts of environmental hazards, contributing to disease and health disparities. Risk assessors, risk modelers, and regulatory analysts are tasked with addressing these concerns and finding solutions to address environmental injustice. This workshop explores how the regulators, the scientific community, and the regulated community navigate the complex EJ landscape. 

2. Tutorial on Influence Diagrams: From Basic Shapes to Software Solutions

Speaker: Cameron MacKenzie

Description: Influence diagrams, also known as Bayesian belief networks, are extremely useful tools in risk analysis. They can incorporate several uncertain factors, combine data and expert opinion, and facilitate Bayesian analysis to update probabilities. However, many risk professionals have little-to-no knowledge about influence diagrams or know how to create an influence diagram to calculate the probability of a risk or the consequences that may arise from the risk. This workshop will teach attendees about influence diagrams by beginning with the basic shapes in influence diagrams and finishing by showing them how to use Netica software to solve influence diagrams.
This workshop will introduce the four basic node types for influence diagrams: uncertainty or chance nodes, decision nodes, deterministic nodes, and value nodes. Arcs or arrows can have different meanings depending on the type of nodes that they connect. Influence diagrams link different factors together via conditional probabilities. Attendees at the workshop will practice creating influence diagrams for different risks in engineering and infrastructure, security and defense, and ecology and climate. These examples will demonstrate risk problems with no decisions, with one decision, with sequential decisions, and for single and multiple criteria.
Although influence diagrams represent an easier method to visualize a problem than with a mathematical model, influence diagrams can be very difficult to solve without computer software. Attendees will learn how to use Netica software (which can be downloaded for free for limited-size problems) to solve influence diagrams and identify the optimal risk management alternative for complex, uncertain risks.

3. Bayesian benchmark dose (BMD) analysis for toxicological and epidemiological data using the BBMD Platform

Speaker: Kan Shao

Description: This full-day workshop will begin with an introduction on the benchmark dose modeling in a Bayesian framework and then provide participants with hands-on experience of using the Bayesian Benchmark Dose modeling (BBMD) system to perform dose-response assessment using toxicological and epidemiological data. The workshop will cover a number of important topics in Bayesian BMD modeling, including using Markov Chain Monte Carlo (MCMC) algorithm to fit dose-response models, using appropriate statistics to evaluate goodness of fit, estimating the distributions of model parameters and quantities of interest (e.g., BMD), calculating model averaged BMD estimates to take model uncertainty into account, and employing the Monte Carlo simulation for probabilistic low-dose extrapolation, etc. More importantly, the workshop will extensively explore the major functionalities of the BBMD system for dose-response assessment through case studies: (1) for toxicological data, BMD analysis of single and multiple datasets for dichotomous, continuous, and categorical data will be discussed and practiced; (2) for epidemiological data, BMD modeling with quantification for exposure uncertainty will be explored. In short, the workshop will provide participants with both theoretical and practical skills of using the BBMD system for dose-response assessment.