From "weight of evidence" to quantitative data integration using multicriteria decision analysis and Bayesian methods

• Published in: ALTEX, alternatives to animal experimentation Vol. 32; no. 1; pp. 3 - 8
• Main Authors: Linkov, Igor, Massey, Olivia, Keisler, Jeff, Rusyn, Ivan, Hartung, Thomas
• Format: Journal Article
• Language: English
• Published: Germany 2015
• Subjects: Bayes Theorem; Decision Support Techniques; Environmental Health - methods; Humans; Risk Assessment - methods; United States; United States Environmental Protection Agency; United States
• ISSN: 1868-596X; 1868-596X
• DOI: 10.14573/altex.1412231

"Weighing" available evidence in the process of decision-making is unavoidable, yet it is one step that routinely raises suspicions: what evidence should be used, how much does it weigh, and whose thumb may be tipping the scales? This commentary aims to evaluate the current state and future roles of various types of evidence for hazard assessment as it applies to environmental health. In its recent evaluation of the US Environmental Protection Agency's Integrated Risk Information System assessment process, the National Research Council committee singled out the term "weight of evidence" (WoE) for critique, deeming the process too vague and detractive to the practice of evaluating human health risks of chemicals. Moving the methodology away from qualitative, vague and controversial methods towards generalizable, quantitative and transparent methods for appropriately managing diverse lines of evidence is paramount for both regulatory and public acceptance of the hazard assessments. The choice of terminology notwithstanding, a number of recent Bayesian WoE-based methods, the emergence of multi criteria decision analysis for WoE applications, as well as the general principles behind the foundational concepts of WoE, show promise in how to move forward and regain trust in the data integration step of the assessments. We offer our thoughts on the current state of WoE as a whole and while we acknowledge that many WoE applications have been largely qualitative and subjective in nature, we see this as an opportunity to turn WoE towards a quantitative direction that includes Bayesian and multi criteria decision analysis.