Recorded at AIHce EXP 2021
What Does a Full Shift Sample Miss?
Your full shift TDI samples have all been <8-hr OEL but occupational asthma still occurred in your work force. Why? The exposure control you installed is not delivering the results you expected. Why? A battery plant worker's blood lead level suddenly increases significantly; but he works in a laboratory. Why? Cases covered in this session will examine how our current exposure assessment methods may be presenting an incomplete picture that can lead to poor risk characterizations, exposure assessments and control outcomes.
Presenter/Author: Tom Morris, MS, CIH, CSP, Morris Innovative IH & Safety Solns, Cincinnati, OH
Evidence of Absence: Bayesian Way to Reveal True Zeros Among Occupational Exposures
Industrial hygiene measurements often include non-detects. The current paradigm for quantitative risk assessment using such measurements is based on the assumption that the data arise from lognormal distribution. This implies that non-detects are believed to correspond to low levels of exposure that are greater than zero (i.e. every samples situation is "exposed"). While this might be reasonable in many situations (e.g. continuous exposure regimen at low exposure levels, a proportion of which are too low to be detected), it might not in others (e.g. intermittent exposures regimen). A more flexible view is to consider that there may be some true zeros among non-detects. The statistical model of such distribution is the zero-inflated lognormal. We used available theoretical developments to estimate this model using Bayesian approach, created an easy-to-use online calculator, and evaluate the impact on risk assessment of allowing some non-detects to be from unexposed situations.
Co-Authors I. Burstyn / Drexel University / Philadelphia Pennsylvania U.S
Presenter/Author: Jerome Lavoue, University of Montreal, Montreal, Quebec, Canada