Bayesian Decision Analysis (BDA) Part 1: Flipping the Profession –“Normalizing” Statistical Analysis
This post is the fifth in a series on insights into exposure assessment as presented at AIHce 2016 in Baltimore, Md. References to specific products or services do not constitute endorsement by AIHA or The Synergist.
“Core to our profession is the ability to make correct decisions about exposures to ensure they are properly controlled,” according to AIHA member John Mulhausen, PhD, CIH, CSP, the director of Corporate Safety and Industrial Hygiene for 3M Company. “Whether they’re chemical exposures or noise exposures or heat stress exposures or biological agent exposures, that’s what we do.”
That “what we do” relies heavily on an IH/OH’s professional judgment, which is often shaped by years of experience. It’s faster and easier than applying traditional methods of statistical analysis and is generally accepted as a valid decision-making tool. But judgment alone isn’t a science—and it isn’t always correct. One group, AIHA’s Bayesian and Professional Judgment Users Group, is working to help IH/OH professionals improve their judgment and strengthen decision making through a statistical analysis technique known as Bayesian Decision Analysis (BDA).
“We’re hoping to flip the profession,” said Mulhausen. “If you look at the literature, what you see is that we aren’t always as accurate as we need to be. So all of this effort is around getting the profession to understand that there is a need to make a change—and then bringing the tools to the profession to help us all make the change.”
The flip is gaining momentum. Mulhausen helps to lead AIHA’s Bayesian and Professional Judgment Users Group, a network for all professionals who are actively utilizing Bayesian tools and professional judgment within their professional practice. The group shares best practices, challenges, and opportunities for strengthening exposure judgments using IH Bayesian tools. Members are encouraged to share their experiences with BDA.
As IH/OH professionals in the field begin to adopt BDA, proponents of the technique are also working at the university level to arm professors with the tools necessary to teach BDA, hoping to spawn a generation of industrial hygienists who regard BDA as a natural and essential component of exposure assessment protocol.
Mulhausen argues that better judgments lead to safer workers and more efficient use of resources in the workplace.
“We are trying to normalize the practice of statistical analysis of data ... because it has some real advantages over some of the other tools. It’s easier to use, it’s easier to understand the outcome, and it’s much easier to communicate [the results],” he said.